CX.AI -Experience the new generation of CX Insights

All Posts By

Lilla Szücs

WAR FOR FEEDBACK

War For Feedback

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: October 14, 2021 * 6 min read

Participation in the customer survey and touchpoint feedbacks has been rapidly declining for years. Not only the amount of feedback is suffering but also the quality. What’s the cause of this trend? How can we fight it? And who will win the war of feedback?

Every time I book a hotel, flight, or rental car, every time I buy a ticket for a game, every time I buy a product at some website – virtually anytime I do a business interaction, I get an ask for feedback.

It’s common sense that I don’t have time to answer all of them. Even worse, I am developing over time the habit of not even noticing those requests.

The driving trends behind this obvious picture is twofold

TREND 1 – DIGITALISATION: With digital survey tools from Qualtrics, Hotjar, to Survey Monkey, it’s a matter of a few mouse clicks and an investment of a view bugs to collect feedback and analyze it somehow.

TREND 2 – ROLE OF INSIGHTS: The first “big data” hype dates now over 20 years back. Since then, the amount of data has doubled every two years. But all this is mainly transactional data collecting along with the business processes.

What is not scaling that fast is customer feedback.

Based on this better data, many problems along the business process get optimized e.g. leveraging AI.

The miracle remaining stays the customer. What drives his behaviors and decisions? How does he perceive his experience?

Businesses release that each company typically has a major bottleneck: to better understand than everyone else the customer.

It seems that this is already common sense. The enlightenment that only better customer insights can only deliver this is trending now and in the future.

A BCG study from 2016 has asked CEO’s what the key areas for improvements are. The clear uncontested #1 was “customer insights”.

Get your FREE hardcopy of the “CX Insights Manifesto”

FREE for all client-side Insights professionals.
We ship your hardcopy to USA, CA, UK, GER, FR, IT, and ESP.

Early Movers Will Win The War

Humans are not just customers, and if they are, they are customers of hundreds of products. If customer insights remain the ultimate bottleneck, it’s clear that managing the access becomes mission-critical.

There are three frontiers you need to operate:

Enough feedback: Acquiring enough feedback from customers is the most intuitive frontier

Quality feedback: Its not enough to gather any feedback. You need to make sure that the quality is good enough to draw useful decisions from that. If you want to cut survey length, find the type of question that captures the most useful information.

Use feedback better: When feedback becomes scarce, you really need to make the most out of it. This field is the greatest sin and area for improvement of our time.

Too often, we survey customers and do virtually nothing with it. It’s not just wasteful but unethical as your customer trusted that the time they invested would be used for good.

Join the World's #1 "CX Analytics Masters" Course

The S.U.P.E.R. Framework

Here is my list of 20 tools and tactics you may want to work thru to win the war for feedback.

They center around five strategies. First, improve, focus or alter the source of feedback (S. like Source). Second, better utilized the data you collect (U. like Use). Third, provide value to your customer so they have an incentive to give feedback (P. for Provide Value). Fourth, improve the execution of every step of the process. Improvements do not sum but multiply up (E. for Execution). Fifth, use multiple or the proper channels to reach out (R. for Reach)

SOURCES

  1. Manage a customer panel: By gathering customers willing to give feedback, you can make sure you will get enough feedback when you need it. Indeed, there is a drawback in representativity that you need to trade-off. Managing own customer panels got today much simpler and less expensive than in the past with software solutions such as Survey Ninja

  2. Public Ratings: Depending on the category, there is often plenty of customer feedback already online available. There is Amazon for consumer products, Google Maps for local businesses like Car Dealers, Tripadvisor for Restaurants, G2 for Software, or Google Play for Apps. Ratings have severe shortcomings when it comes to representativity. But when you are not interested in descriptives but in what drivers the rating, this data can still be gold.

USAGE OF DATA

  1. Calibrate low sample scores with ML: CX data is gathered to compute a CX/VOC score. Too often, the sample size is so small that the score is not reported. Score Calibration takes past samples and trains a Machine Learning Model to predict the expected score. It uses two ideas:
    1. Certain information (e.g. the sum of weights of the sample) can indicate how representative the sample is
    2. If you measure CX across regions or segments, the score of other splits can serve as a predictor.
    Overall, Machine Learning can typically calibrate scores and reduce the required sample size by half. On top of this, computing, confidence bands can help socializing those results.

  2. Collect only as much as needed: For some segments, you may have plenty, and for others, not enough. It is wise to become spares with send-outs for those segments where you get more feedback than needed. Because any survey request is paid by the good-will of your clients, use it only when needed. You may need it in the future.

  3. Utilize feedback better: Why do you collect customer feedback? …to compute a score only? No, you also want to know how to improve it. Making the most out of the qualitative feedback is a must. Its also an obligation to your customer to read, understand and act on their feedback. First, you should utilize tech that categorizes verbatimes like a human. Second, you need to run Driver-AI to understand how relevant those categories are to explain the total rating. This article gives all the details about the methodology.

  4. Fixing inner loop: The inner loop is the process of referring the customer verbatim feedback to the frontline colleagues. This process needs to be designed with care. The reason for this is that the most often mentioned topics are seldom the most important ones. When you refer to feedback to someone in order to read it, he will learn the wrong things. This is because people believe the most often mentioned topics are -therefore- the most important ones. This article has all details.

PROVIDE VALUE

  1. Valuable Feedback: There are three main motivations to take part in a survey. First is to make an impact and to improve the service with feedback. Second is to feel heard. Third is to help the provider. Respondents need to feel that their feedback is heard. Ideally they want to see that the provider took action or that the feedback is truly helpful. In B2B context, where the vendor plays an essential role for the customer, response rates are often high. If you consistently give feedback on what has been done with the feedback, some companies achieve responses rates of higher than 80%. Even when this is an astronomic number for your context, providing feedback to those who gave feedback is one way to nurture the good-will of your customers.

  2. Report: One incentive you could give before the interview is the promise to come back and report what you did with all the responses. Another trick to attract respondents is to frame a survey as a self-assessment (e.g. XYZ maturity assessment) and then send an automated report after completion. Customers will then take part because to benchmark themselves and get some judgment. Measuring CX will be just a side effect.

  3. Incentives: If needed -of cause- you can think of additional incentives to participate. When choosing, always think about things that do not relate to the loyalty as a customer. If you promise $10 you will attract people who need some money or are notoriously frugal. Is this a strong bias for your sample? No? Then this is the way to go.

  4. Psychologic value: The brain runs on fun. Everything boring or cumbersome has a hard time. Think twice about how you can make your survey entertaining. Anyways, show your appreciation and gratefulness. Using an active listening technology can even achieve that customers feel better heard and understood. At CX.AI, we developed a survey plugin that actively probes after text feedback. Although (or because) it is frank about being a bot, respondents open up. Like a robot cuddle toy where people know it’s a robot, they still develop relationships with them.

EXECUTION

  1. Optimize Email Delivery Rate: Most survey invites today are still send out via email. Everyone who has spent time optimizing email outreach to prospective customers knows that there is a whole science behind this. Achieving that enough of your emails will hit the inbox of your audience is not straight forward. Your email account needs to be set up right, the time and the format is relevant, and there are certain words and characters as well as the numbers of links and pictures will make your email die in spam or other filters. In short: not just blindly use a tool and send out; hire deliverability experts to fix this.

  2. Optimize Email Open Rates: Once our email hit the inbox does not mean it will be opened. Main drivers are the subject line but also the sender name and the first line on the text body. Subject lines should be concise (1-5 words) and spark interest. A tool like NEUROFLASH can help you optimize your subject line using AI.

  3. Personalize: Any Email without greeting me with my name must be spam or phishing. That’s a table stake. Anything else that you know about your customer can be used that the email feels more custom and therefore get noticed.

  4. Be Most Efficient: If you want to measure the likelihood to recommend in a survey, why not put the rating question right in the email? Clicking onto it leads to a website where more questions may be asked. Use the most intuitive and efficient way of feedback: ask open-ended questions and let customer use their own words to describe it: either per text, audio or video feedback.

  5. Active Listening: If it gets harder to convince customers to give feedback, it becomes even more critical to get the maximal amount of information, Active Listening is a real-time technology that can understand the topics raised by a respondent and then ask a more focused probing question. It is called active listening as it feels for the respondent that someone is listening (= trying to understand) and is interested in his view (=because he asks deeper questions). The consequence is much richer feedback that has been shown to increase predictive power by 50 to 100%.

  6. Shorter surveys: Every now and then, a stakeholder comes and wants to add a question. Fast forward some years, and you end up with a lengthy questionnaire that just tortures customers.  To me, a CX survey should have a rating and an open-end (text, audio, or video), ideally an active listening probe and making sure the context of the respondent is known (segment, product history, key demographics)

  7. Mindful surveying: Every request for feedback takes a toll on customer good-will. That’s why just ask as many, and as frequently you are sure you can act on and draw actions from it. For instance, a sophisticated structure is to have a yearly (or low level continuing) general survey, but then doing spot-light studies to understand specific areas in much more detail. E.g. if the general survey revealed that ‘uncomplicated complaint handling’ is a key, then a spotlight should explore what actually ‘uncomplicated complaint handling’ could mean in real life.

REACH

  1. Multiple channels: Most feedback is requested over email. Another 30 of companies still use the phone. But there are even more channels available like texting, social DMs. The best reach you achieve by trying it all together as each customer prefers different channels. Yes, each channel has its bias, but there are ways to debias the measure. If you want to win the war for feedback, you need to tackle this.

  2. Use frontline: A great way to achieve more feedback is to let the frontline ask for it. They do not need to collect the feedback but ask for their willingness to provide it. Who says no to a ask for 1-minute feedback?

  3. Predictive surveying: Using customer master data and eventually transactional data you can somehow predict how likely someone will follow a survey feedback invitation. Typically you can eliminate send-outs by 70% or more and still get the same sample size. Yes, they might be biased. But this bias again can be controlled by modeling. If you want no longer to spam your customers with survey invite request (from which 99% will not be followed), then this is worse exploring.

Winning the war for feedback requires investing in ‘Starwars’ tech, not in spears and catapults.  The S.U.P.E.R. framework gives you five areas to work on and dominate the ‘battlefield’ for human feedback:

Keep Yourself Updated

On the Latest Indepth Thought-Leadership Articles From Frank Buckler

Your Ultimate Masterplan

First, expand the source of feedback. Blending different sources of feedback will be the new normal. Debiasing measures have become a new science.

Second, better utilized the data you already have. As feedback becomes more valuable, it is crucial to utilize data better. I see HUGE potential here.

Third, provide value to your customer, so they have an incentive to give feedback. The way you plan your CX program tells a lot about how customer-centric you are.

Fourth, improve the execution of every step of the process – from email send-out to survey design. Improvements do not sum but multiply up.

Fifth, use the right or multiple channels to reach out and invest in predictive outreach.

If you want to dive into more cutting-edge CX thinking, the “CX Analytics Masters” Course is for you. It’s free for enterprise insights professionals. If you are looking to discuss some of the advanced technics mentioned above, with an expert, reach out at www.cx-ai.com

Now I have a question: Was this article helpful?  Please DM me directly with any comments or questions

Frank 

"CX Analytics Masters" Course

b2

P.S. Would you like to get the complete & interactive FREE CX Measurement Guidance for your business in 2022?

Simply subscribe on the free “CX ANALYTICS MASTERS” course below and enjoy the above-mentioned training guidance in its Class # 1.

“Solves key challenges in CX analytics”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI

How To Build And Maintain A CodeBook That Fits Your Needs?

How To Build And Maintain A CodeBook That Fits Your Needs?

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: October 9, 2021 * 7 min read

Building a codebook is an important step in the management of a data analysis project. Especially when you deal with unstructured qualitative data, you need to find the proper solution to categorize it, and use the reliable analytics within your organization.

Apparently, unstructured data appears chaotic at first glance, but with new forms of AI data analysis, it can be tamed to solve business problems.

First we look at how to set up a codebook correctly.

Get your FREE hardcopy of the “CX Insights Manifesto”

FREE for all client-side Insights professionals.
We ship your hardcopy to USA, CA, UK, GER, FR, IT, and ESP.

How To Set Up A CodeBook Correctly?

Be granular and think about actions – There are various types of coding. Manual coding and supervised coding require the human to build a codebook, and there are some do’s and don’ts around that. Of course, it depends on the use case, but in general, the recommendation needs to be granular and you have to think about the actions.Therefore, it’s not enough to have a code that is correct. It should also be useful and it becomes useful if you are more specific. It’s okay to categorize something as “poor quality”, but it would be good to know how to improve this quality. So, granularity is the king and growing a narrative translates into actionability.

Always have codes with the direction – It means that the code itself has the sentiment. Why? It’s because you can have codes like price and quality. But when people actually read it, they don’t know how to understand that. When you build categories, always construct them in a way that they have a direction like bad pricing, good pricing, bad quality, and good quality. This way, everyone who reads it, understands what it means when you say quality and price. So, it becomes an exercise, and later it enables you to detect granularities by having direction-based codes.

Be mutually exclusive – It means make sure when you choose a certain category, the other one can’t be the same. For instance, friendly service is not mutually exclusive with great service. You have to understand that one specific verbatim belongs to only one category. There is no doubt whether or not it belongs to. If you are not clear, write a description consisting of one or two sentences as it helps the coder to remember the real meaning. But if you can’t describe it, that’s probably not a good piece of code.

Cluster categories in category groups – You have to cluster categories in category groups to define a hierarchy as it enables you to have overview summaries. You can group the categories by classifying them into different clusters like the parent category and the child category. The important thing is to train the child of the specific category.

Label as customers speak – You have to be specific when you label a certain category. Typically, we tend to use general terms that are hard to understand. So, you have to be descriptive and use the language of the customer for the below reason:

  • First, you will better understand what it means and AI will understand it better too because the latest AI is not just the grouping of categories, it really looks at what’s the label of the group. What does it mean? Is it associated with the content of the verbatims? So, that’s why the label itself is important for supervised learning.

Use AND + OR when labeling – Sometimes, you have certain categories, and they can not be differentiated from one another. For example, brand, great brand, reputed brand are hard to differentiate from a customer’s point of view. It is important to use AND and OR the right way when you have certain categories. AND means the customer mentioned both categories, and OR means the customer mentioned one of them. It is necessary to give these details to the person who reads the label as well as the AI the correct guidance.

OTHERS category – What’s very much used are OTHERS category. The only use of it is to count how much we did not categorize specifically. But, it should only be used for manual coding because there are two reasons for that:

  • You risk to confuse the AI as it’s such a broad thing, and is trained to find something as a link between something to something specific.
  • When you do that, it may work somehow, But you will not be able to leverage the feature whether category or AI suggests to you the topics. It’s super important to understand that AI knows what it doesn’t know. It can only be managed if you do not use another category.

Non-informative categories – The non-informative categories are those that don’t belong to our codebook. Typically, customers may say:

“I gave it zero because I don’t recommend it in general.”

You need to pay attention to what the customer says because it helps you explain the outcomes. 

Minimum frequency per category – There is a question:

What’s the minimum frequency of a category?

It depends on the usage and it’s something you need to decide. You always build categories and open a new category if it’s mutually distinctive to something that is already mentioned in the other categories. At the end of the exercise, you can still decide to skip some of the categories.

Join the World's #1 "CX Analytics Masters" Course

How To Find New Categories?

Manual Coding – For manual coding, you have to read all comments. You build your codebook, and when new data arrives, you need to categorize everything from scratch and look for new categories. You also code everything that doesn’t fit to your codebook and visit the OTHERS category to find the new categories.

Unsupervised Categorization – It always finds new categories but changes the definition of the old ones. Therefore, it does not maintain consistency. So, if you can find a software that can fix your old ones, it may be able to find the new categories without changing the old ones.

Supervised Categorization – Supervised learning is trained by a domain expert. So, it needs humans to find the new categories, but there is a trick called smart sorting ( smart categorization). It sorts the verbatims in a way that you can be quite sure about them i-e., you will be sure what is the quality, and what is the price etc. So, supervised learning can help you detect new topics or categories very quickly and easily by getting rid of the OTHERS category.

Speed Training LM

SPEED-TRAINING: Reinvent Your CX Analytics and Win C-Suite by Storm

Crystal Clear CX-Insights with 4X Impact of Actions

60-minute Condensed Wisdom from The World's #1 CX Analytics Course

*Book your slot now, last few seats left*

How To Manage The Code Consistency?

Unsupervised learning – You know that unsupervised learning finds new categories by retraining and does not maintain consistency. You need to avoid retraining and find a tool that can fix category definitions. 

Manual Coding – How can you manage consistency in manual coding? 

  • Do not change coder – The biggest problem is changing the coder (human). It’s because the other person will code differently than the actual coder resulting in errors. During categorization, the coder goes into a learning process from the beginning to the end. So, another coder will have a different learning style and it creates problems. Therefore, you need to stick to one person for managing consistency. 

  • Do intensive hand-over – When you change the coder, you need to think of intensive hand-over. 

  • Gold Standard – The gold standard is to have a team of persons who code redundantly. Always have two persons categorizing the same stuff and they must have the ability to validate themselves.

  • Supervised AI – You can use supervised learning as your second order. If the person trains the AI, it will be the second order that can maintain the consistency of the code. 

  • Maintain Category Definitions – You need to explicitly write down what is the verbatim that belongs to the category, and what is the verbatim that does not belong to the category.

  • Small Codebook – You need to keep your codebook smaller than 100 codes if possible because of several reasons. The main reason is that the code receiver finds it hard to find the right code. So, the larger the code, the more likely you will be making mistakes in categorizing.

Supervised Categorization – The same thing applies for supervised learning because the training process involves manual coding. Focus on quality coding and make sure you train the machine right, so it will be consistently correct. In short,

  • Supervised learning is the same as manual coding.
  • Each time you train the AI, it might be a slightly different model.
  • Backward categorization with recalibration can cater both needs. You train the new AI and feed the past datasets to the new model. You will see that the frequencies and the categorization of the past verbatims are a little bit different. You can adjust and recalibrate the frequency scores. You can find the factor, so that the frequencies of the past verbatims remain the same, and you can simplify the recalibration factor for the new verbatims.
How To Deal With Multiple Languages?

If you are an international company, you have multi-language feedback. The question is how are you dealing with that? How are you categorizing it?

Native Coders – We need native people or native coders who can take a look and categorize. It’s because that is the best you can do as they are the best source to understand what the verbatim means.

  • Alignment – The question is how do you make sure that the categories are not really the same but are understood the same. There are some differences between languages. For instance, there are some words in German that don’t even exist in English, and the other way around. So, it’s impossible to have a hundred percent match of understanding between languages. 
  • Impractical for more than three languages – I would recommend you not to use this exercise of native coders if you have more than three languages. It is because those native coders really need to communicate and make sure that they understand all categories the same way. 

Translate first into core language – The alternative approach is to take all verbatims and translate them into one core language. You may think that you lose information.

Yes, you do. But you have one understanding of categories, and one native quarter categorizing everything. 

  • Use DEEPL.com – There is tech available that is much better than Google Translate. You can use DEEPL.com that supports twenty-four languages. 

So, if you have multiple languages (more than three), you can use this method as it gives better results than native codes. 

In a Nutshell

So far, we discussed that you need to build and maintain a codebook because it is important in managing a data analysis project. The following are the steps of building a codebook:

  • Be granular
  • Have a direction
  • Be mutually exclusive
  • Cluster categories
  • Label as customers speak
  • Use AND + OR when labeling

Further we discussed that you can find new categories by using the following categorization schemes:

  • Manual Coding – It reads all comments and avoids OTHERS category.
  • Supervised Categorization – It finds new categories but does not maintain consistency.
  • Unsupervised Categorization – It requires a smart categorization feature.

You can manage consistency in these methods by practicing some necessary steps. Further, you can deal with multiple languages by using two methods:

  • Native coders
  • Translate first into core language

"CX Analytics Masters" Course

b2

P.S. Would you like to get the complete & interactive FREE CX Measurement Guidance for your business in 2022?

Simply subscribe on the free “CX ANALYTICS MASTERS” course below and enjoy the above-mentioned training guidance in its Class # 1.

“Solves key challenges in CX analytics”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI

How To Validate Text Analytics System?

How To Validate Text Analytics System?

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: September 3, 2021 * 9 min read

A customer experience is a qualitative and emotion based experience. The companies are obsessed with turning this into a quantitative measure. Companies want to track a number whether it is a Net Promoter Score, Customer Effort or Customer Satisfaction score. Tracking a score like NPS can highlight the need to improve but the number alone can not provide the insight you need to make the improvements.

Many companies rely solely on this scoring system as they do not have time to do a thorough analysis of the feedback they receive. This is where the need for a text analytics system comes in that gathers the insights from thousands of open text customer comments.

Let’s first understand what text analytics is.

Get your FREE hardcopy of the “CX Insights Manifesto”

FREE for all client-side Insights professionals.
We ship your hardcopy to USA, CA, UK, GER, FR, IT, and ESP.

What is Text Analytics?

You can think of text analytics as the process of deriving meaning from text. It is a machine learning technique that allows you to extract specific information, or categorize survey responses by sentiment and topic. 

Companies use text analytics to:

  • Understand data such as emails, tweets, product reviews, and survey responses. 
  • Provide a superior customer experience.
  • Mine the voice of customer feedback and complaints.
  • Turn the unstructured thoughts of customers into more structured data.
  • Allow customers to provide feedback on their terms in their own voice, rather than simply selecting from a range of pre-set options.

Join the World's #1 "CX Analytics Masters" Course

How To Validate Categorization?

Now, let’s move towards validating our categorization as it is important to understand whether the categorization is correct.

The trick with Hitrates – The hitrates must be calculated the right way, and if you want to calculate whether your tech service category is correct, you can look at the hitrate. If you are categorizing none of your verbatim i-e., the verbatim belongs to none of your categories, then your hit rate is 98 or 99%, and that’s very high. 

Do you know why? It’s because you can be very sure that the likelihood that one of your codebooks is within one verbatim is very small. To have an accurate categorization, you need to look at the following grid.

Here,

  • True positive indicates the outcome where the model correctly predicts the positive class.
  • True Negative indicates the outcome where the model correctly predicts the negative class.
  • False positive indicates the outcome where the model incorrectly predicts the positive class.
  • False negative indicates the outcome where the model incorrectly predicts the negative class.

As evident from the above grid, false positive is the type one error, and false negative is the type two error. 

Alpha vs. Beta Failure – Alpha failure is also called False Positive, Type 1 error, or Producers’ risk. If the alpha failure is 5%, it means there is a 5% chance that a quantity has been determined defective when it actually is not. 

On the other hand, Beta failure is also called False Negative, Type 2 error, or Consumers’ risk. It is the risk that the decision will be made that the quantity is not defective when it really is. 

F1 score – It is the ultimate measure of consistency that takes both false positives and false negatives into account. It takes everything, weights it, measures its frequency, and comes up with the right measurements. So, F1 score is the gold standard score used in science to measure categorization quality.

But, F1 score only measures what you are doing is consistent or not. You are not sure if it’s correct. So, there is another term when we talk about validity, and that is Predictive Power.

Predictive Power – It is the measure of truth that helps you find the true categorization. The truth can be best found by determining whether or not it is useful to predict the outcomes. If you have something that is described through the category, and it has an impact in the world, we categorize it. It’s because we think it is important to drive outcomes. So, if this can predict outcomes because it was some kind of important, then it’s probably correct. 

In short, predictive power is the test to measure true categorization, and to predict and measure outcomes. So, the R2 of everything you do towards outcomes is the final measurement of whether or not your categorization is great.

Two years ago, we compared the different categorization schemes where we took lots of data and tried to compare unsupervised learning with manual coding and supervised learning. When we took it to the predictive power test, unsupervised learning achieved an R2  of 0.4. Then, we used open-source supervised learning and it was much better and much more predictive than unsupervised learning.

But it was not even close to manual coding. So we tried further and found a supervised learning approach, which we call your benchmark supervised learning approach that even exceeded the predictive power of manual coding. 

So, there is a big difference between different approaches and the field is evolving everyday, but it is important to test its power. The best is to validate its predictive power and you may ask why a machine can be better than humans. It might not always be better than a human but there are some advantages. First, you have seen that the training of supervised learning is augmented. So, the trainer itself becomes better by training because he gets feedback from the machine.

On the other hand, the sentiment of the machine is better than the sentiment of a human and when it comes to the tonality, this is what the machine can detect much better. It can find much better, and much more predictive, the tonality of the verbatim. 

In short, the supervised learning to categorize data is much better than manual coding due to the following reasons:

  • It leverages a knowledge database for sentiment codes.
  • It produces fine-grained scores instead of binary Yes/No predictions.
Speed Training LM

SPEED-TRAINING: Reinvent Your CX Analytics and Win C-Suite by Storm

Crystal Clear CX-Insights with 4X Impact of Actions

60-minute Condensed Wisdom from The World's #1 CX Analytics Course

*Book your slot now, last few seats left*

In a Nutshell

So far we discussed that text analytics is important as it can be used to improve customer experience. It can also be used to gather their feedback through which you can uncover a deeper insight. In order to validate your categorization, you need to have a concept of the following:

  • False Positives (Type one error)
  • False Negatives (Type two error)
  • F1 score
  • Predictive Power

Also, we compared different categorization schemes and concluded that automatic coding is much better than manual coding.

"CX Analytics Masters" Course

b2

P.S. Would you like to get the complete & interactive FREE CX Measurement Guidance for your business in 2022?

Simply subscribe on the free “CX ANALYTICS MASTERS” course below and enjoy the above-mentioned training guidance in its Class # 1.

“Solves key challenges in CX analytics”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI

The AI-for-Insights Maturity Model

The AI-for-Insights Maturity Model

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: 29.09.2021 * 6 min read

“What else can I do with AI” is what I have been hearing in professional insights groups recently. The number of solutions is exponentially growing, but AI has not yet affected the insights process as this might indicate.

We all have heard of text analytics or facial recognition with AI. More and more applications pop up, and it can feel crowded ….

…if you don’t have a compass, take mine.

Get your FREE hardcopy of the “CX Insights Manifesto”

FREE for all client-side Insights professionals.
We ship your hardcopy to USA, CA, UK, GER, FR, IT, and ESP.

AI Needs Us To Think Different

AI is like a magician buster. A magician does his tricks, and the outcomes are surprising. The audience doesn’t know how this could have happing. AI finds out by looking hundreds of times closely.

Whenever you have data about the input (e.g. text feedback) and data about the outcome (e.g. the themes the text fits into), you can let AI find out the missing formula that can predict outcomes with particular inputs.

The basic idea of AI is straightforward. This simple concept, however, is so different from what we are used to.

We are used to think “hypotheses”. We are used to think that we need to tell the computer what to do, how A may lead to B. We can’t imagine that algorithms can learn about complex behavioral mechanisms just without us.

Join the World's #1 "CX Analytics Masters" Course

AID Framework

Maturity comes in stages. First, kids learn to eat, run, and speak. Later, when entering teen age they can do everything an adult can do. They can be so eloquent you may believe they are adults. But cognitive maturity needs more.

In the same way maturity of AI application for insights evolve.

First, you see that AI is used to automate what a human can do in insights. Humans can categorize verbatims or tell you whether a person looks sad or happy. But a machine can be trained to do this at scale.

The second stage is to use AI for automating marketing, service, or sales activities – fulled by stage one. Again, a human can do all this too, but deploying a machine has costs, speed and quality implications.

The ultimate stage, however, is to use AI do discover insights about the link between input and output, between cause and effect. This links the information from stage one to inform stage two. Here is where the ultimate AI-insights-Loop closes.

#1 Automate insights

Most AI you know is automating what’s already here. But automation does not just mean lower costs and higher speed. As a result whole new research procedures evolve.

  • Text analytics: helps us to categorizes customer feedback. It is the process to quantifying qualitative text-based information. Today it can have the same quality as manual humanmade categorization even in more complex B2B settings. AI can even categorize the emotional side of text. It can read for instance anger, sadness or surprise. It can spot the tonality beyond the actual meaning of the words.

  • Association AI: Words are unconsciously associated with each other, like Milk and Cow. There is AI that extracts the recent associations that are implicit in public text data. This astonishingly is very close to what you can measure in Implicit Association Tests. As such, the tool can serve as a proxy. Mainly it is used where it is economically not feasible to apply IAT. Recent applications use this AI to find words and language that better resonate with your audience.

  • Facial recognition:  Based on the research of Paul Eckman there is a validated theory of how mimics can be interpreted as emotions. It includes 7 base emotions happiness, sadness, surprise, disgust, anger, fear, and contempt, plus 12 styles of joy. The machine now can even more reliably as humans read mimics and minor mimical cues from pictures and videos.

  • Voice analytics: Analyzing voice can be done today with a variety of cloud-based services. You can transcribe it into text, and you can read emotional cues in the voice. It is even possible to detect a COVID infection just by analyzing the coughing sound of a person.

  • Eye-tracking AI: Thousands of eye-tracking studies had been used to train AI models with the intention to predict eye-tracking outcomes without actual eye-tracking. It can undoubtedly be dependent on the context and the target group to where audiences look at first. But research shows that up to 90% of fixations are independent of this and actually hardwired. With this, you can analyze ads, websites, popups, or even newsletter emails before launching. It enables to faster iterate and improves the impact of assets.

  • Visual analytics: Similar base technology is used to understand the content of pictures and videos. The same way AI can be trained to detect emotions or to predict eye-tracking results, visual analytics platforms today detect by default the object a picture has, like a table, skirt, a man or clouds. Those machines can even be trained to see more abstract concepts like “stylish interior”, “pop culture”, or “a teens room”.

While all this is amazing already, the true power of AI comes with combining it with the other two stages of AI maturity.

Keep Yourself Updated

On the Latest Indepth Thought-Leadership Articles From Frank Buckler

#2 Insights into actions

There is no value in insights unless you do something useful with it. Here are some tech utilizing AI to improve marketing and sales actions

  • Chatbots: I don’t mean this cumbersome tree type question branching system that are mostly in use today. The basic use of AI in chatbots is to understand unstructured text and respond in a predefined manner. The more advanced chatbots use AI that actually mimic responses of actual humans. They are trained on hundreds of customer-agent chats and will respond to customer requests just like a typical agent. Sure this comes with limits, and it will not be able to chat on any off-topic comment.

  • Text synthesis: the GPT3 technology from Open-AI made it in the press in 2020 and led to a wave of text synthesizing services. AI today can write emails, poems and whole articles. In most cases, it still needs human review. But it has been proven to not just speed up the copywriting process but also help with ideas and inspirations. Leading applications that use GPT3 are now designed to not just produce well-written copy, but copy that works. This means emails that convert, slogans that people will remember and subject lines that make people open emails.

  • Picture and video synthesis:  Early versions of GPT4 now not only produce new texts but pictures and videos that are purely fictional. Besides the cost-saving effects (less photo and video shoots), the real power lies in the ability to teach the system to produce converting or convincing pictures.

  • Deepfake avatars videos: These are services that use a given video and change typically the face and let the person speak new arbitrary sentences. With this, you can produce very quickly explainer videos and even a video service interface, without the need to actually record a video.


Voice synthesizing: Computer-generated voice is actually a four decades-old discipline. So far it has always been rule-based and computer voice always was somehow recognizable as a computer. This changes now with AI as those flexible algorithms can inject those little imperfections that make a voice feel human.

#3 Discover insights

AI today can automate a variety of research processes, from text coding, facial reading to eye tracking. Then AI can help synthesize copy, pictures, videos, and one-on-one service interactions. 

But still, something fundamental is missing. 

Automating to read customer feedback still does not include 

Automating to read customer feedback still does not include understanding which of the customer feedback has to biggest impact when improving the matter. 

Also, to craft a good copy, I need to know what separates a “good” from a “not so good” copy. 

Any marketing and sales action relies on a simple assumption. The causal assumption is that an action will lead to a particular outcome.

AI now can help us to find those models and causal assumptions about the world that will be most impactful.

AI-powered operative learning loops

Imagine you run a weekly newsletter that drives traffic to your website and specific offers. AI can be used to optimize the conversion process in many stages. Image it increases open rates from 40 to 50%, click rates from 4 to 5%, and landing page conversion from 4 to 5%. This will result in a sales increase of 100%.

Based on enough examples, AI can not only predict which subject line, picture, and copy will convert better, it can also tell us why. 

Further, we can use AI to create subject lines, pictures, and copy at scale and use this to perform multivariate massive-scale experiments. Instead of sending all 10.000 recipients one or two versions, you can now send 500 different versions every week. 

Causal AI then can learn from the outcomes of these massive experimentations.

Causal-AI powered strategies

Besides the tactical optimization of marketing and sales process like the newsletter sent out, AI is used to understand winning strategies.

This is actually where Success Drivers has specialized in for more than 10 years. We helped T-Mobile to build a winning go-to-market strategy. We found the right product strategy for SONOS to foster growth. We enabled METLIFE to distill the winning DNA of successful advertising.

Now we are focusing with CX.AI on gaining strategic directives with AI from the customer feedback nearly every brand gathers today.

Its outcome is even used to enable proper organizational learning as feedback shouldn’t be just delivered to the frontline. Instead, it needs an AI-powered filter and sorting process to make customer-facing units draw the right conclusions.

GROW OR GO

AI is here to stay. Stage one of the maturity curve is slowly diffusing into the insights process for a few years already. New developments of AI solutions help to augment and automate customer interaction. 

All this makes the steps faster, cheaper, and more consistent. It is the ingredient to scale and to automate.

But a new quality level of insights is enabled by stage 3 of the maturity curve: applying Causal AI to understand the link between actions, context and results.

Technology for Causal AI is on the rise for more than 10 years. It is still waiting for its breakthrough simply because it is the third and final stage of AI maturity.

Research leaders who understood the enabling power of Causal AI are utilizing it already successfully today. 

If you want to dive into this field, the website of the pioneers, Success Drivers is a good start: www.success-drivers.com, and of course, its core produce CX.AI www.cx-ai.com 

Now I have a question: Was this article helpful? 

Please DM me directly with any comments or questions 

-Frank

"CX Standpoint" Newsletter

b2

P.S. Would you like to get the complete & interactive FREE CX Measurement Guidance for your business in 2022?

Simply subscribe on the free “CX ANALYTICS MASTERS” course below and enjoy the above-mentioned training guidance in its Class # 1.

“It’s short, sweet, and practical.”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI

How To Leverage The Power Of Sentiments?

How To Leverage The Power Of Sentiments?

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: September 26, 2021 * 5 min read

It would be best for an organization to analyze customer sentiment because it helps understand customers better and improves product experience. To analyze customer relationships better, you have to understand their feelings and the rationale behind their rating or sentiment.

Let’s understand first what sentiment is and what are its common types.

Get your FREE hardcopy of the “CX Insights Manifesto”

FREE for all client-side Insights professionals.
We ship your hardcopy to USA, CA, UK, GER, FR, IT, and ESP.

What Is A Sentiment?

By sentiment, people typically understand whether a comment, verbatim, or feedback is positive, negative, strongly positive, or strongly negative. But there are other understandings of it. It’s important to understand because if you build a codebook, it already has some sentiment as part of the code. For instance, the sentiment of a codebook indicates the good quality and bad quality of the code.

So, sentiment is an inherent part of a category definition. It has the power to measure the ROI (Return on Investment) of marketing campaigns and help organizations improve their customer service. It also gives businesses a sneak peek into their customers’ emotions so that they can be aware of the crisis. But you can understand the sentiment differently too. It depends on what kind of sentiment algorithms you use.

Sentiments can also be the tonality of a comment. Do you know what tonality is? You can think of it as the difference between a positive and negative experience. It can also be the difference between a lost customer or a lifelong relationship. You can understand tonality with the help of an example given below:

Suppose an XYZ customer complains about the missing feature in your product with a positive tonality and says:

“Your great product can be improved by using this (xxx) smart feature.”

You see that the customer gave a negative comment but it’s very positive. So, tonality is a type of sentiment that helps customers connect with your brand and encourages them to support your business – only if it is positive.

Join the World's #1 "CX Analytics Masters" Course

How To Classify Sentiments?

Apart from tonality, the other types of sentiments are as follows:

  • Emotional Coding – It is a type of sentiment that scores verbatims on the following seven universal emotions:
    • Anger
    • Fear
    • Disgust
    • Happiness
    • Sadness
    • Surprise
    • Contempt


There are some softwares and APIs you can use to categorize which emotions certain verbatims belong to, and what probably the emotions are that trigger those verbatims.

  • Associational Coding – It is the type of sentiment that scores your verbatims to what terms the language is associated with. It can be useful to uncover insights for branding. Artificial Intelligence can help in finding which other terms the verbatim is associated with. 

For instance, if a customer says straight to the point that:

“Your service is crap.”

It is a very masculine way to comment and is a dominant statement. Dominant AI can detect the dimensionality of hundreds of terms and determines what your verbatim belongs to. It’s useful when you use verbatim collected in the branding space so that you can look behind the words to understand the complete meaning.

AI-powered tools like Neuroflash mirror what people think and determine how customers perceive or claim a brand’s content.

Keep Yourself Updated

On the Latest Indepth Thought-Leadership Articles From Frank Buckler

In A Nutshell

So far we discussed that sentiment is an indicator that measures how customers feel about a certain product or service of an organization. It also helps brands discover the reason why customers leave some negative feedback. Sentiment can also measure the tonality of a comment. Further, we categorized the sentiments into the following types:

  • Emotional coding
  • Associational coding



"CX Analytics Masters" Course

b2

P.S. Would you like to get the complete & interactive FREE CX Measurement Guidance for your business in 2022?

Simply subscribe on the free “CX ANALYTICS MASTERS” course below and enjoy the above-mentioned training guidance in its Class # 1.

“Solves key challenges in CX analytics”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI

How To Avoid Benchmarking In CX?

How To Avoid Benchmarking In CX?

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: September 21, 2021 * 8 min read

While benchmarking can be a powerful tool for comparative analysis and understanding best practices, it can also lead to bad conclusions if the wrong information is compared.

But first, you need to understand what benchmarking is. It is a process to measure the performance of a company’s services, products, or processes against those of another business considered the best in the industry. 

Benchmarking is a simple and five-step process as shown by the following points:

  • Choosing a service, product, or internal department to benchmark
  • Determining organizations, you should benchmark against
  • Gathering information on their internal performance or metrics
  • Comparing the data to identify gaps in your company’s performance
  • Adopting the processes and policies in place within the best-in-class performers

Get your FREE hardcopy of the “CX Insights Manifesto”

FREE for all client-side Insights professionals.
We ship your hardcopy to USA, CA, UK, GER, FR, IT, and ESP.

Why Do Business Partners Ask For Benchmarking?

In the previous blogs, we discussed how to collect customer experience data and how to analyze it. Further, we saw how to quantify unstructured feedback and gain insight and use it in a dashboard interface to help the businesses draw the right conclusions. 

Sometimes, there comes the request to do some benchmarking. Why? It’s because most companies are used to it as business partners ask them to do it. So, if you present numbers to the marketing department of C-suite, they immediately think of:

  • Is it good?
  • Is it bad?
  • Do you have benchmarks?

Let’s talk about the perceived benefits of benchmarking, as it is an important topic to talk about.

  • We benchmark as it provides an easy answer to the question :
    • Are we doing good or bad?
    • Do we still have the potential to improve?
  • Benchmarking gives you relief when you are performing relatively well or even better than the best-in-class performers. You feel good when you do a great job and feel best when you hear that you are wonderful among all competitors in the industry.

  • Benchmarking gives you directions to meet competitive performance. It sets the goal that you achieve a competitive level or somehow exceeding it is the way to go.
Why is Benchmarking Dangerous?

In my opinion, benchmarking is very dangerous, and it provides the wrong incentives. You may be in a situation where you need to offer that. But what are the right alternatives? You may, over time, be able to offer alternatives instead of benchmarking. Why? Because it is not helpful for the business. So, benchmarking assumes the following:

  • You have the same type of customers
  • What’s important for your customers is important for competitive customers as well.

The best competitor is clearly pushing the limits.

Let’s see what the risks of benchmarking are.

  • False Signals Risk – It signals that everywhere there is a gap, you need to perform better. You need to understand what is important and what you should do to improve yourself. 
  • Wrong Benchmarking – It’s due to serving different customer segments. For instance, it is dangerous to compare apples with oranges because you will get the wrong signals. Both of these fruits belong to different classes, so we can not compare them.
  • Good vs. Bad Signals – When you get wrong signals (either good or bad) by serving different customer segments, the blame game starts and is not productive. It’s because, for several reasons, the benchmark is not actually the benchmark.

Join the World's #1 "CX Analytics Masters" Course

What Are The Alternatives To Benchmarking?

Let’s talk about the alternatives to benchmarking. Consider the example below that shows two key driver canvases.

These canvases belong to the same industry i-e.,

  • YOU – It denotes the customers.
  • LEADER – It denotes the market leader.

You can see all the different topics, and the green line is just a 10% mark. Below is the competent service (written in German) in the customers’ canvas and experience at the touchpoint (written in German) in the market leader’s canvas. The positive topics of the leader are much better than those of the customers, which means that the most important mentioned topics are some way better than for us. Interestingly, it is the case for 95% or more topics in which the leader is better.

Do you think this is the case only here? According to the research of Professor Byron Sharp from Ehrenberg-Bass Institute:

“The NPS and the frequencies of the items are better for the market leaders.” 

If you look at benchmarking, you will need to level up everything the customers have. They have just a handful, maybe four or so items, that are comparable to the leader. So, it’s questionable if this is a real benchmark.

You also see that the leader’s key drivers and key leakages are different from those of your customers. So, there is a strong indication that you have attracted a different breed of customers.

The above example shows that benchmarking doesn’t lead you to good information. If you want to improve your customer experience, look at your key driver canvas, and it will tell you how to become better. It may be useful when attracting other customers as you understand what the pain points or hidden drivers of the competition are. So, you can attract them with the right marketing. But, remember that the market leader canvas is for marketing and the other one is for customer experience management. 

So, benchmarking itself is not needed as it is not a descriptive exercise. It can not tell you what’s important and what’s not. It can only tell you where you are better and where you are worse.

My recommendation is:

  • Know what’s important as it’s enough.
  • Stop blame game with arbitrary targets.
  • Constantly challenge yourself and establish a “The sky’s the limit mindset.” It’s because with benchmarking, you only address or look at where you are bad in, but it would make much more sense to take certain criteria and excel by getting much better than the competition. You don’t have to look back. Instead, you have to look forward. 

Therefore, there is no reason for looking at the competition. Look at your customers because:

You have to serve your customers, not your competition. 

Here’s a nice quote:

“Look In The Mirror. That’s Your Competition.” 

So, do not benchmark as it sets the wrong incentives.

Keep Yourself Updated

On the Latest Indepth Thought-Leadership Articles From Frank Buckler

In a Nutshell

So far, we discussed benchmarking can benefit you by providing guidance. It can also direct you where to focus. It can only be helpful if it provides the right signals. The right signal tells you where you have the potential to improve and what is important to improve. So, if this can be met by benchmarking and circumstances can be justified, you can apply it. 

Further, we discussed the alternatives to benchmarking and concluded that it does not lead to improvements. It is a non-productive exercise that provides you with the wrong signals.

"CX Analytics Masters" Course

b2

P.S. Would you like to get the complete & interactive FREE CX Measurement Guidance for your business in 2021?

Simply subscribe on the free “CX ANALYTICS MASTERS” course below and enjoy the above-mentioned training guidance in its Class # 1.

“Solves key challenges in CX analytics”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI

How To Size The Fiscal Impact Of NPS?

How To Size The Fiscal Impact Of NPS?

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: September 20, 2021 * 7 min read

It would be best to use modeling to measure the fiscal impact as it greatly affects customer loyalty. Increase the amount of money each customer spends with your business. This way, you can generate positive word-of-mouth about your business.

Do you know how to increase the fiscal impact? Expecting a great product or service without excellent customer service is like expecting your beautiful garden flowers to flourish without giving attention to them. You can increase your fiscal impact by adopting the following two methods:

  • You need to be consistent in your service delivery
  • You need to encourage your customers to complain

You know that NPS measures the customers’ loyalty and how likely they will refer your products and services to others. It also helps to identify the promoters and detractors of your business. If NPS is so important, you should know how to measure it.

Get your FREE hardcopy of the “CX Insights Manifesto”

FREE for all client-side Insights professionals.
We ship your hardcopy to USA, CA, UK, GER, FR, IT, and ESP.

Why Do We Need Modeling To Measure NPS Impact?

In the earlier blog, we discussed the simulator where you can put in the fiscal impact of one person if you improve the NPS by one point. You can see the improvement in the customer value when you improve the NPS point. 

You cannot simply look at the correlation between NPS and the customer value or revenue per customer (you can have any measurement of customer value). For instance, let’s take the NPS ratings of last year of a particular set of customers and check:

  • Did the ratings evolve positively?
  • Did the customers with lower ratings buy less? Etc.

When you do this exercise, you will often see something like the diagram shown below.

You would see no correlation between NPS Rating and the impact. It’s against all theory, and you’d be thinking why the graph is like this. It’s because it is a spurious correlation and needs modeling.

In the example above, we can see that there are two different segments:

  • Top segment
  • Usual Customers

You will observe that the NPS significantly impacts outcomes if you use modeling to control the segments. The Usual Customers is the segment that does not have high expectations. On the other hand, the Top Segment is the one that expects high, and because of their expectations, they, on default, give low ratings. But when they are satisfied, they are likely to buy more because they have the means for that. So, they are the customers who have the bias to rate less, but if they are satisfied, they become pretty valuable. That’s how you can use modeling to control the customer segments.

The diagram below clears the understanding between the two customer segments.

Join the World's #1 "CX Analytics Masters" Course

What Is The Importance Of Fiscal Impact?

Are you aware of the importance of the fiscal impact? The fiscal impact is essential and super powerful because:

  • It helps you judge the ROI ( Return on Investment) of your actions.
  • It helps you compare the ROI of customer experience with new customer acquisition. 

There’s an old saying that:

“Keeping your existing customers is five times more effective than gaining new customers.” 

It may be a myth because it has not been proven in scientific studies. But still, it is a very fruitful exercise to have customer acquisition and experience as you need to understand the ROI.

You need to measure or assume a fiscal value of one NPS point – It’s a starting point to understand the ROI of your actions. It’s very useful to even start with an assumption. You can take your past experience or get experience from other experts and put it into the simulator (that’s an assumption). Afterwards, you can do the exercise of modeling.

How To Measure The Fiscal Impact?

It takes a one-off modeling study to measure the fiscal impact.

  1. First, you need to take a past customer experience survey. Take the data from the last term and amend it for every single customer to check:
  • How does the customer evolve?
  • Does he churn?
  • Does he buy more?
  • Does he buy an additional product? What’s the margin of it?
  • What’s the typical time the customer stays on this product? Etc.

    2. You need to calculate the value of the churn or the customer value. You can also use past NPS ratings along with customer master data and context information to predict how many products have been purchased and if the customer has churned. You can multiply your NPS rating with the value of your action to come up with an impact on the bottom line.

    3. Your past survey should ideally include different customer profile information. The more the data, the better you will have the picture of your customer. For instance, if you know where your customer lives, it will be the indicator of the potential confounder. Further, we’ll use modeling to predict the outcomes. We will use NPS and try to use all information it knows about the customer. If NPS is the signal information just really needed to predict the outcome, it can be assumed to have a true impact. So, modeling gives us an unstandardized impact value of the NPS rating onto outcomes.

    4. This value needs to go into the calculation that considers:
  • Product Lifetime
  • Margin
  • Customer Value

You need to transfer and recalculate the NPS rating towards the NPS score. 

This way, the impact value can give you great insights related to the impact of your actions.

Keep Yourself Updated

On the Latest Indepth Thought-Leadership Articles From Frank Buckler

In a Nutshell

So far, we discussed that we can use modeling to measure the fiscal impact, which is important because it judges the ROI of your actions and compares the ROI of customer experience with new customer acquisition. Further, we discussed, to measure the fiscal impact, we need to do the following:

  • Take a past CX survey.
  • Use NPS rating and other customer data to predict the value of the churn.
  • Modeling will give you an impact value of NPS.
  • You need to feed this value into the calculation.

Lastly, you need to find the P&L (profit and loss) impact of our NPS changes as it will guide your investment in the right direction.

"CX Analytics Masters" Course

b2

P.S. Would you like to get the complete & interactive FREE CX Measurement Guidance for your business in 2021?

Simply subscribe on the free “CX ANALYTICS MASTERS” course below and enjoy the above-mentioned training guidance in its Class # 1.

“Solves key challenges in CX analytics”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI

How To Get Buy-In By Using A CX Impact Simulator

How To Get Buy-In By Using A CX Impact Simulator?

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: September 19, 2021 * 7 min read

It’s easy to recommend the CX actions from insights. But how can business partners size the relevance if there is no means to evaluate the exact impact?

In this article, we will solve this problem and show you:

  • How to bring transparency in the future outcomes?
  • How to explain past changes in NPS scores?
  • How to learn from the past most effectively?

Get your FREE hardcopy of the “CX Insights Manifesto”

FREE for all client-side Insights professionals.
We ship your hardcopy to USA, CA, UK, GER, FR, IT, and ESP.

How To Simulate Impact of Actions?

We can’t comprehend and evaluate a large number of scenarios and outcomes on our own. So, here comes the need for another thing handy for the dashboards, which is a simulator.

A simulator is used to access the current or predict the future performance of a business process. It helps the practitioners discover new ways to improve their business processes through statistical, mathematical, and other analytical methods. If we consider the key driver canvas, we know it shows frequency and the impact that can be high or low. But what does high impact mean? The key driver canvas has no answer to this question, so we need a simulator to understand the meaning of impact.

The simulator helps business owners to investigate the performance of their business processes. You can understand the meaning of impact using a simulator if you perform the following actions:

  • Take a simulated change in frequency – You need to wait after simulating necessary changes in the frequency.
  • The result is the change in the CX score  – You need to multiply the unstandardized impact of the topic obtained from key driver analysis, and you will observe the change in the CX score.
  • Multiple changes can stack up – When you simulate several categories, you will need to stack up the inputs.

Simulation is beneficial for the business partners to understand the meaning of impact because of the following reasons:

  • It makes it obvious how huge your efforts are needed to get the NPS up by ten points.
  • It challenges you to be aspirational and to do step changes.

It is evidence-based feedback that is a motivator for businesses to set higher targets.

The above figure shows a simulator that looks just like a slider. When you change the important topics, the NPS changes.

This figure shows that you can also simulate the financial impact. You can put in the number of customers that have arrived and put in the currency. Further, you also need to assume or measure the impact of one NPS point on the bottom line. Then, you can see the impact of working on one topic towards your bottom line on the simulator.

Join the World's #1 "CX Analytics Masters" Course

What Are The Reasons For The Score Change?

There’s another feature very useful for the companies, and that’s the topic of explaining the reasons for the score change. But the companies face the below issue:

Issue – Typically, the score changes from one month or from one quarter to another. When it drops, the blame game starts. So, why did it drop, and who’s guilty of that?

Background the data has different interests. For instance, you worked on the competency of your staff. When you see the competency of your staff improved, you analyze it, shout out and say, “ This is why my project is working and showing correct results.” But this is not the way you should look at it because everyone will get different results and interpret them according to their understanding.

Solution – So, you will have dozens of different interpretations, but you need to have one evidence-based truce. You can do it by:

  • Evaluating the changes from one quarter to another, and
  • Simulating the frequency change of competency.

What does the simulator say if we improve by 2%? What is the impact if the NPS changes? To understand this, we need a simulator algorithm.


The simulator algorithm helps us evaluate the contribution of every single topic in the score change. We can add the different contributions of topics in a bridge diagram and pick up the most significant changes to the score to show them.

The diagram above shows the example of a dashboard that implements the simulator algorithm.

This is the bridge graph that shows the top contributions of the main drivers to the change in the NPS score. The main drivers are:

  • High-Quality Product
  • Reliable Performance
  • Great Sound Quality

The contribution of these drivers is essential, but some negative things happened, as shown by the “red” curves in the above diagram. These things have a negative impact on the NPS score change.

We can derive the following conclusions after analyzing the above bridge graph:

  • Some things go well even though the score drops.
  • Some things may get worse even though the score increases.

External reasons or sampling biases cause the change in the NPS score.

So, the impact can be positive or negative. You need not look at the outcomes and the NPS only, but you need to understand how you progressed while looking into the change in the NPS score.

Keep Yourself Updated

On the Latest Indepth Thought-Leadership Articles From Frank Buckler

In a Nutshell

We know that the key driver canvas is the core tool for communication. We discussed that we need a simulator because this makes the business partners feel the value of an action. It also challenges them to put in more effort to improve their business processes.

Further, we discussed that we need a simulation bridge to explain the past changes. So when the scores change, there’s a reason, and we saw the contribution of the top drivers in their change.

"CX Analytics Masters" Course

b2

P.S. Would you like to get the complete & interactive FREE CX Measurement Guidance for your business in 2021?

Simply subscribe on the free “CX ANALYTICS MASTERS” course below and enjoy the above-mentioned training guidance in its Class # 1.

“Solves key challenges in CX analytics”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI

The Hidden Key To Happiness

The Hidden Key To Happiness

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: September 15, 2021 * 6 min read

Spring 2021 did KANTAR, the largest market research company globally, reached out to 20,000 citizens in ten countries around the world – USA, Brazil, Canada, Netherlands, Italy, Portugal, France, China, India, Japan. The goal was to ask questions nobody with commercial interest would ever ask.

The question about happiness in life naturally emerged as a key question. Plenty of unstructured feedback about what concerns people were gathers. KANTAR now asked CX.AI to take this data and uncover hidden drivers of happiness. Here they are.

Get your FREE hardcopy of the “CX Insights Manifesto”

FREE for all client-side Insights professionals.
We ship your hardcopy to USA, CA, UK, GER, FR, IT, and ESP.

Truth hides behind facts

Facts do not equal Truth. Example?

People were asked about the things that need to change. They most frequently respond “more money/income/revenue”. This is a fact.

People were asked about their motto for life. Most frequently, they respond “be a human, help or respect one another”. This is a fact too

People were asked about their long-term goals. Most frequently, they respond, “stay healthy or get healthier”. Even this is a fact

But none of that is the truth. The truth is that none of those facts play an important role in explaining what makes people happy.

We all have an unconscious tendency to jump from fact to conclusion and from correlation to causation. Managers, as well as researchers, do it day in day out.

Facts are just data that needs analysis to find the link between what is happening and the IMPACT to happiness.

It’s not enough to ask closed or open questions. This only delivers data.

The same happens in Customer Experience Measurement. Companies ask for the likelihood to recommend and the open-ended question “why”. These programs just give data and facts but no impactful insights (=truth).

Join the World's #1 "CX Analytics Masters" Course

AI helps to uncover the hidden truth

Imagine you have data about an outcome you like to improve. Be it customer satisfaction, likelihood to recommend, likelihood to purchase a product concept, amazon rating, or in our case, the happiness with life.

Then you have more data about the person but at least an open question about possible reasons that may influence the outcome.

How useful and universal would it be if we could take those data, pour it into an AI that then helps find what best drives results.

Precisely this is the CX.AI methodology. It evolves in three steps

  1. A deep learning text analytics AI is trained by a human coder. This gives double predictive power over conventional text analytics systems. We deploy the Caplena.com platform.
  2. A causal machine learning software models nonlinear and indirect effects and considers driver interactions. It typically results in double explanation power to conventional key driver analysis. We deploy the Neusrel.com platform.
  3. Results are fed into an interactive dashboard that enables real-time simulations and predictions.

So now, let’s look at the results. Here is the ultimate question on the outcome we are looking for:

‘How happy are you feeling with your life in general currently?’

We discuss drivers in 3 areas: impact of goals, the impact of problems, and impact of the attitude to life.

The total model achieved an explanation power of .40 and more than doubled the performance of conventional key driver analysis with .18.

Before we dive into this, let’s look at an astonishing finding:

Woman are suffering

The average happiness rating of men and women is roughly the same. But this is just a FACT, not the TRUTH.

The causal machine learning is looking at all explaining information. It finds that being a woman leads to a .76 lower score (on a 10 point scale), if everything else is the same.

Women do obviously some things to keep up. But something about what it means to be a woman in our society is not measured by the survey data and affects their happiness.

Keep Yourself Updated

On the Latest Indepth Thought-Leadership Articles From Frank Buckler

Area #1 – Long-term and short-term goals

The survey asked, “What are some of your long-term goals? ….and you short term goals” and provided a multiple-choice list of 27 things to choose from.

Take a deep breath and look at the key drivers canvas that plots the topics on a graph with the frequency mentioned on the Y and the causal impact on the X-axis.

This are the people’s goals that make them happy and unhappy right now.

Around 50% have the long-term goal to travel the world. This goal does not fit in the actual pandemic and makes that 50% immensely suffer. It is the single most impactful reason for unhappiness today.

People with this long-term goal would answer on average nearly three scale points less (if everything else were the same).

The anticipation of a holiday in the short term provides relief instead.

To me, the findings reveal the so far underestimated importance of freedom to happiness and thus mental health. This is particularly interesting for countries that tend to lock in their citizens much more than others.

With some data matching to this study, we could easily measure the impact of this fact on higher mortality as it is known that lower happiness leads to mental and physical illness and death.

Again, respondents’ most often mentioned goal was to keep or improve health, but this goal seems to be a new platitude as an instant association that evolved in pandemic times.

#2 Things to improve

The second scheme is all around this question ‘Can you think of 3 things that would improve the quality of your life right now?’

As a result, earning money is common and important. The aspiration for more money is a predictor of unhappiness.

But gaining a better or new job is not as common but even more important. The aspiration of improving the work environment is an even larger predictor of unhappiness.

The intention to do more physical exercises as well as to buy a new car is an indicator of wellbeing as the impact of those indicators is somewhat positive.

This means people neither by a car nor go to the gym to fix a problem they have.

#3 General attitude in life

Lastly, we integrated information about the attitude in life in the model. It seemed plausible that not just the external challenges but the internal attitude is essential to gain happiness. The exact question posed was:

‘What would be your motto or one piece of advice you would want to pass onto the world?’

Most respondents mention mottos like:

  • Be a human, help or respect others
  • Never give up; great things take time
  • Live without worry
  • Enjoy every moment in life.

But none of those mottos have a significant impact on happiness.

Instead, what makes people happy is a deeply routed attitude and habit of GRATITUDE. ‘Belief in god’ is also impactful and naturally overlaps with gratitude, a core part of religious practices.

In contrast, what makes people unhappy are limiting, and distrustful believes like ‘Be careful who you trust’ and ‘make good use of your money / save money’

The Hidden Formula of a Happy Life

This simple formula of happiness has been extracted out of 156 potential explanatory factors. It’s a common theme across 20,000 people of 10 countries. It explains nearly 40 percent of why someone is happy in his life:

“Practice gratitude – day in day out. Try to trust more and forgive. Don’t care about money too much, let it go, invest in the now. Instead, choose the right job. Switch to something you are passionate about. Most importantly, try to find peace in substituting traveling-the-world with creative regional traveling plans.’

The exercise revealed another finding for the benefit of any business:

It’s not enough to look at facts or to correlate drivers and outcomes. It is even dangerous.

Instead, the study is a living proof, how powerful a proper analysis can be.

Imagine this methodology is run on your CX feedback, brand, or product testing data? How much more impactful could your business initiatives be, how much more successful could your new products become?

Simply by using better the data you already have.

Let’s talk.

-Frank

(DM me here)

"CX Analytics Masters" Course

b2

P.S. Would you like to get the complete & interactive FREE CX Measurement Guidance for your business in 2021?

Simply subscribe on the free “CX ANALYTICS MASTERS” course below and enjoy the above-mentioned training guidance in its Class # 1.

“Solves key challenges in CX analytics”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI

How To Derive The Right CX Initiatives Based On CX Analytics?

How To Derive The Right CX Initiatives Based On CX Analytics?

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: September 15, 2021 * 6 min read

You need to translate your topics into valuable actions so that you can recognize and articulate precisely what you want to achieve. When the data sources are consolidated, you can gather meaningful data insights.

Do you know what actionable insights are and how you can derive them into your organization? Insight is an understanding you have about the working of your business against how you thought it would work before you accessed the data-driven insight. It makes you rethink various business factors you may not have considered before.

You can derive insights about your customers or your business by structuring and transforming your data into information that can optimize your processes and understand your customers in a better way.

Get your FREE hardcopy of the “CX Insights Manifesto”

FREE for all client-side Insights professionals.
We ship your hardcopy to USA, CA, UK, GER, FR, IT, and ESP.

What is Dashboarding, Prediction And Prescription?

Let’s first understand dashboarding, prediction, and prescription before we discuss what actions to take with the help of key driver canvas.

Dashboarding – Organizations use dashboards to gain insights into the most significant aspects of their data. They gain real-time insights and competitive analysis through which they identify items that require urgent action.

Prediction – Prediction helps organizations make sense of potential outcomes or the future repercussions of a decision. It leverages historical figures and statistics and uses raw and up-to-date data to peer into a future scenario.

Prescription – Prescription also looks at future scenarios but uses a more technological approach. It takes a deeper look into the “What” and “Why” of a potential future outcome by utilizing artificial intelligence, mathematical algorithms, and machine learning.

Now, let’s look at the key driver canvas because it is the key tool to understand the main takeaways.

In the above key driver canvas, we have two dimensions:

  • The first one is the impact of the different topics that people are mentioning. It can be a positive impact or a negative impact.
  • The second one is the frequency that describes how often something is mentioned.

You can categorize these dimensions into three buckets or fields to arrive at a certain recommendation. 

  • The first field has topics that have high impact but low frequency. These topics should be approved. So, this field is called a hidden lever because it has a frequency that is not often mentioned. The other fields are also essential but have a high frequency.
  • The second field is termed as Maintain in the above canvas. The outcome of this field is to maintain where you are good at, but improve where you have room for improvement.
  • The third field is the red zone in the above canvas. It is the other room for improvement, and we call it linkages. It has topics that are often mentioned and have a high negative impact. So, there exists a tradeoff as a cup of persons are mentioning something that is heavily negative. However, they should be mentioning that thing that has more importance. 

In short, we need to follow the below steps to gain insights from a key driver canvas:

  • Identifying Impact Potential – We need to identify the impact of the potential i-e., dimension mentioned on the x-axis.
  • Hidden Levers –  We first address the hidden levers of the key driver canvas.
  • Key Leakages – We need to identify the key leakages in our red zone.

Maintain Strategy – We need to maintain our strategy where we are good at it.

Join the World's #1 "CX Analytics Masters" Course

How To Translate The Topics Into Possible Actions

When you go into discussion with any organization, you arrive at the question: What do we do now to improve our services? It is because customer service has a significant impact on an organization’s success. It’s a sign of bad customer service if the customer retention and loyalty levels are low. 

So, there are multiple ways you can adopt to improve your service, and you need to actually perform the action to translate the topics into possible actions. 

The actions that you need to perform are:

  • Jot down the actions that increase or decrease the frequencies.

  • Evaluate the actions based on:
    • Cost – Cost determines whether you can quickly implement the action. Easy here means you have all the necessary resources to carry out the action, and the outcome of the exercise is pretty sure.
    • Ease –  You can easily determine how to perform the action successfully if you determine the cost in the above step.
    • Impact on other topics – You need to determine the impact of cost on other topics to take proper actions to improve your service.
  • Integrate preferred action into the label of the dashboard. That’s an option, but it’s pretty interesting if you have excellent service and not enough people mention it. So, you need to do monthly training, and it’s advantageous to integrate your recommendations right into the dashboard for internal use. 

You can derive the recommendation from simple ease of impact metrics. For instance, you can categorize your actions as:

  • Easy
  • Not Easy

There are three situations associated with the above actions. They are:

  • The actions have low impact AND high costs.
  • The actions have a high impact OR low costs.
  • The actions have a high impact AND low costs.

The above figure demonstrates when the actions have low impact AND high costs, do not categorize them.

When the actions have high impact OR low costs, put them on the WAITLIST if it’s not easy to categorize them. However, if it sounds easy, you need to INVESTIGATE, as shown by the above figure.

There are things that have a high impact AND low costs. If they are not easy, it still really makes sense to INVESTIGATE them. On the other hand, if something has a high impact, low cost, and is easy, that’s your PRIORITY.So, that’s the easiest way to get improvements, and we can integrate this into the dashboard as well.

Keep Yourself Updated

On the Latest Indepth Thought-Leadership Articles From Frank Buckler

In a Nutshell

So far, we have discussed the terms dashboarding, prediction, and prescription related to the key driver canvas. The driver canvas is the primary tool to understand the impact of different topics and their related frequencies. We categorized the dimensions into three fields, namely:

  • Hidden
  • Maintain
  • Linkages

Further, we discussed the necessary actions we need to perform for the translation of our topics. We categorized the actions as Easy and Not Easy and associated three situations with them that are:

  • Low Impact & High Costs
  • High Impact OR Low Costs
  • High Impact & Low Costs

So, we can improve our services and customer satisfaction by deriving valuable insights from our data.

"CX Analytics Masters" Course

b2

P.S. Would you like to get the complete & interactive FREE CX Measurement Guidance for your business in 2021?

Simply subscribe on the free “CX ANALYTICS MASTERS” course below and enjoy the above-mentioned training guidance in its Class # 1.

“Solves key challenges in CX analytics”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI