Duff Anderson

Who’s Directing Customer Opinion – Swarm Theory

To understand what a large group believes and how it’s beliefs may be changing, limit your analysis to what they say the most often in the same way. There is a simple analogy that demonstrates why this approach to analyzing pluralistic feedback works well to distill the most relevant message while eliminating the need to categorize every individual one.

Consider a sporting event where 30,000 fans attended. The goal is to understand what the outcome of the contest was, however, all you have available is a recording of what was said aloud by each individual fan.

One approach would be to listen to each fan, create a subjective list of concepts categories, and then place each concept in the category it belongs and then look at the size of the categories relative to each other. Although this approach may produce interesting results, it is resource intensive, time consuming, and you will have a hard time understanding what happened in the game.

However, analyze collectively only the chanting, the cheers and boos, when people speak in the same way most frequently, and you’ll quickly get an accurate picture of the outcome.

This approach leverages the idea of swarm theory. It explains why large groups act with directed purpose while individuals remain self directed. There is no central command pushing a particular outcome, the theory of swarm behavior relies heavily on the idea that monitoring relative change in the frequency of events determines future behavior.

Let’s return to the game; imagine a player is well set up to score, a crescendo of sound arises from the crowd, now everyone is paying attention. It is the change in frequency of audible comments, the hum that is alerting the collective attention, not a central leader saying look now!

iPerceptions is leveraging this idea by developing new tools for understanding open-ended customer feedback. Using concept quantification, concept concordance, and the iPerceptions perceptual framework, we are generating metrics like the iPerceptions Momentum Index, that examine the rate of change of several key factors such as average size and negative – positive presence of concepts. These metrics point you to important changes in customer opinion as they are occurring and potential changes as they are just about to.

Know satisfaction – measuring customer satisfaction is a process not an event

I am often challenged by decision makers who believe that periodic customer satisfaction studies and their results are not actionable, or that organizations themselves are not capable of repeatedly acting on the results.

In a sense these decision makers are putting the cart before the horse. These decision makers see customer satisfaction as a performance measure that should be gauged every once in a while. They see customer satisfaction as an event, as a checkup.

However, leveraging customer satisfaction results necessitates understanding that customer satisfaction is a process not an event. By continuously listening to customers and measuring their satisfaction based on a consistent framework and language, the ability to take action emerges. Organizations have the chance to get to know satisfaction. Continuous listening moves customer satisfaction from measurement to decision support. It helps you see trends, understand meaningful changes, and be confident when things are constant.

The danger of using customer satisfaction measurement as a checkup is demonstrated well in TV sitcoms. In my time, the classic Jack Tripper misinterpretation always leads to the wrong and comic decision because he acts based on a snippet of information.

Knowing what drives customer satisfaction, past, present, and looking forward is the most powerful decision support possible, but it becomes a reality only as a process not an event. So keep listening to know satisfaction.

Panel research is not the Voice of the Customer

I’ve run into a number of satisfaction studies lately that claim to speak for real customers. A little digging showed that these claims are based on studies that use panels of people similar to your actual customers. The Internet has allowed us to speak directly to customers in the context of real experiences. There is an emerging new breed of market research analytics that is democratizing the voice of the customer. Panels are no longer needed to speak for people, people can speak for themselves.

Our research shows that the way panels rate satisfaction is not the same as the way actual customers in real situations rate it. What is important to them and how satisfied they are is very different. The difference is easy to understand; For example, when I'm in my underwear at 1 AM in the morning, on my laptop, trying to get something done, I'm not experiencing things in the same way as when I am tasked to do something and then evaluate it. I’m not creating fictitious situations and then evaluating them. I can provide real and immediate feedback on my actual experience based by my actual needs.

Panel and focus groups have their place. My point is that the voice of the customer is different; it is about actual customer satisfaction. I’m not saying it's better than panels, I’m saying it's different. It’s a new type of information that is more closely aligned with what customers actually experience and how well their expectations are satisfied.

With a panel, people are tasked to behave in a certain way and then asked to rate their ability to act in that way. They have been qualified using demographics or psychographics, whether they are in market to purchase, or whether they make a certain income or not. They are often rewarded for participating. Voice of the customer respondents are qualified by their own unsolicited choice to visit the site for their own purpose and they are not directly rewarded for participating.

A good analogy for the difference between the information that comes from a panel study and the voice of your customer is the difference between a date and a blind date.

On a blind date you've been tasked to evaluate each other. The way that you function and the way that you respond to each other is completely driven by the fact that you’ve been tasked to either like or dislike each other. That's exactly what a panel does, it sends you on a task and it asks you for your opinion.

Let's look at a regular date; the mutually qualified situation, where you've met somebody and you've agreed to get together. In this case the exploration is more natural and closer to the truth. The outcome is not driven by the fact that you're solely there to evaluate each other. It is driven by your mutual interest in a potential relationship.

A real conversation between real people, where there is mutual interest, is how I see the methodology driving the new voice of the customer analytics. Customers have real needs, companies offer real solutions. The voice of the customer is driven by this real interest and a sincere desire to share and listen. A big step forward in democratizing the voice of the customer, a big step towards understanding customers needs.

Representative sample – is it really the voice of the customer?

In the late nineties online customer research was quickly dismissed as non-representative simply because the medium presented new challenges that didn’t conform to traditional research practices. However, within a few years the paradigm has shifted and innovative methodologies have captured the voice of the customer more clearly than ever before.

Below I have listed some of the reasons online research is allowing you to listen to your online customers like never before.

1. It facilitates inclusive random sampling; by easily and verifiably soliciting participation on a random basis, like 1 in every 100 visitors.

2. Online research can naturally select for the most influential customers who best represent your brand; certain forms of self-selecting participation attract the ‘connector’ customer who regularly speaks up, s/he has a greater impact on your image. This customer will talk to others customers and in other public forums, reputations don’t create themselves, see The Tipping Point.

3. Timeliness of information; online research provides the ability for customers to provide feedback within the context of real situations at a cost that is not prohibitive. It can also provide immediate access to this feedback.

4. Online research can provide continuous sampling at low cost; sampling on a continuous basis reduces environmental and seasonal biases.

5. It can capture large sample sizes easily and at little expense; large data sets normalize the data and allow for powerful statistical analysis.

6. Due to the interactive nature of an online collection interface you can screen out the noise inherent in the medium; simple procedures like checking your dependant variables against predictors and throwing out respondents who’s dependant result is more than 2 standard deviations of their independent ratings. These are the customers who are simply browsing the survey and not answering seriously. However, don’t let them know you’ve thrown them out, the feedback process is still a positive trust building exercise with these customers.

While I’m not saying it’s a census, and it’s not a panacea for all your customer research needs, online research is proving to be the most representative cost effective way to listen to the voice of your customer. Solid decision support for leaders who believe that truth is most inherent in democratic systems, so share, be a part of it.

Customer Feedback - ask for it openly and politely - it doesn’t hurt - it helps

How likely are you to agree to provide assistance to someone asking you for feedback as you are leaving the office at 5:00 pm, or for that matter how likely are you to offer assistance because you saw a memo requesting feedback posted in the lunch room? These situations are good analogies for why exit surveys and passive requests for customer feedback online don’t work and provide biased views at best.

The right way to get customer feedback is to be upfront and polite. Provide customers with a clear and easy way to opt out, and when they do don’t immediately ask them again. Whether customers participate or not this approach builds integrity for the brand. This approach tells all your customers that their opinion matters, whether they have the time to provide it or not.

I thought the days when IT specialists were directing business strategy went the way of the dinosaurs in 2001. However the extinction was not complete as the following email I received yesterday shows.

“The request for feedback as proposed is that the invitation is presented before the customer reaches our home page. Even though this would only be presented to a percentage of our customers, the risk of losing bookings with the diversion is potentially material.”

Sorry, but wrong, wrong, wrong and wrong.

1. Wrong because the invitation for valued feedback builds trust and loyalty with all visitors, not only participants.

2. Wrong because research shows that booking and conversion rates go up slightly while a proper request for feedback is running.

3. Wrong because there is no opportunity for abandonment because whether the visitor clicks ‘yes’ or ‘no’ to participate after their visit, the outcome is the homepage.

4. Wrong because the opportunity that is lost by not listening to your actual customers in the context of real situations is material.

Listen to your actual customers, it doesn’t hurt, it helps.

Continuous Listening and the Street Voice

Quantifying qualitative feedback from actual customers in actual situations on a continuous basis is opening the door to analyzing other voices. The patterns found in focused feeds, like the webValidator, provide the building blocks for the natural language algorithms required to analyze the less focused feeds like blogs and 3rd party web sites in a meaningful way. The interaction between actual customer feedback, let’s call it Voice of the Customer, and less focused but powerful feeds like blogs, let's call it the Street Voice, should provide a new level of understanding to continuous listening.

Where are issues first detected? Do they appear first in focused customer feedback and then get circulated loudly by the street voice, or do customer issues gain momentum only after the street has made them heard. It should be interesting to see, using real data, how the theories put forward by Malcolm Gladwell in the The Tipping Point apply. The potential implications related to managing brand perception are huge. This added intelligence should provide organizations with powerful decision support, helping them decide which and when customer issues need urgent attention.

Voice of the customer – spring, summer, fall, winter, and spring…

Being the first day of spring, the anxiously anticipated vernal equinox, especially for those of us residing north of 40, got me thinking. We've come to expect certain things at certain times of the year.  These cycles have become second nature to everyone and we certainly would doubt the news that the next three months were going to be colder than the three just pasted.

We have achieved this confidence not by accident but by observing and studying the temperature continuously over time.  In the same location a particular temperature is considered hot at one time and cold at another. And who doesn’t check the weather channel when planning an outdoor activity?  

The same basic principles apply when it comes to using the voice of the customer as decision support. It is absolutely necessary to continuously measure the voice of the customer to better understand which perceptions and attitudes are changing all the time, which cycle, and which remain constant.  By taking a continuous approach to listening we have a chance of moving measurement to understanding, surprise results to insights, and cycles to things we can plan for.

Is it significant? measuring customer experience

Decision makers (DM) like comparing numbers, however, some approaches to understanding statistical difference are useful as decision support while others simply speak to mathematics. Take the following example: Group 1 scores a satisfaction rating of 5.6 while Group 2 gets 6.1.

Obsequious understanding
          DM: Is the 0.5 difference statistically significant?
If it made it into a report then my answer is probably something like:
          ME: Yes, 19 times out of 20.
But what have we learned?

Prodigious understanding
The question that leads to real decision support is:
          DM: How significant is this difference compared to other differences?
          Me: Actually, it’s the most significant difference found.

Leveraging the power of pattern recognition and statistical classification is the key. Focus on finding the most important differences and clearest patterns, ensure that you are barking up the right tree.

When it comes to using the voice of the customer as decision support: It’s about consistently pointing to the moon, not landing on it...where to go, not how to do it.

White paper: Discovering the most distinct user experiences on a web site
Request form requires qualifying information

Adapt or perish

One of the most misguided notions regarding the theory of natural selection is that the strong out compete the weak and win. In actuality the theory suggests that the one who is best adapted to its environment survives, while those who are not, don’t.

To extend this analogy to long-term survival in business, the consumer ultimately defines the environment. Following this argument to its conclusion sheds light on the increased interest in the arena of “Voice of the Customer", “Buzz“ and "WOM" metrics.

The Internet has made the customers’ voice much more audible. Enhancements in analysis algorithms and increased computer processing speeds have also improved our ability to hear it. I like to think of this enhanced ability as an organization’s hearing aid, its improved ability to listen to its environment.

Listening in itself is not enough; the challenge is to adapt to the environment. But how can you be sure you are adapting to the right message? You can’t, but you can increase your chances by following some simple rules.

1. Base insights on far larger samples than others
2. Gather heterogeneous samples of real people in real situations
3. Pass no judgments but merely describe what people are saying using a framework and language that facilitates decision making

My focus for this blog is to explore the issues surrounding this phenomenon, nothing short of the democratization of the customers’ voice. I see clearly a day when businesses will readily accept the voice of the customer as a basic requirement for good decision-making. Just as the role of the secretary has changed and made way for ‘MS Office’, so will decision makers and influencers change to make way for the ‘voice of the customer’.