Thursday, August 11, 2011

Net Promoter Score - The Search for the Magic Pill

Finding tools or strategies that will help a company grow is of interest to just about every business owner and operator. We're all on the lookout for the newest approach, for that proverbial magic pill that will give us a leg up on the competition.

Fred Reichheld, in his book The Ultimate Question, believes that he has uncovered that very magic pill. In his view, the answer to a single question - How likely are you to recommend us to a friend or colleague? - is the only thing that business operators need to know from their customers. The result, marketed as the Net Promoter Score (NPS), has received a lot of attention in recent months.

ENTERPRISE RENTAL COUPONS

The premise of NPS is simplicity itself. Responses to the likelihood to recommend question are solicited on a 0-10 scale, with 0 meaning the least likely to recommend and 10 meaning the most likely to recommend. Responses are then grouped in the following manner:

· Customers with responses of 9-10 are categorized as Promoters. · Customers with responses of 7-8 are categorized as Neutral or Passive. · Customers with responses of 0-6 are categorized as Detractors.

The theory is that "Promoters" are satisfied and loyal customers who will keep buying from a company, and are most likely to suggest that friends and acquaintances do the same. "Passives" are somewhat satisfied but generally unenthusiastic customers who aren't particularly motivated to offer a referral, either positive or negative. "Detractors" are dissatisfied customers, quite possibly trapped in a bad relationship, probably seeking alternatives and assumed to be unafraid, perhaps even eager, to share their experience with others.

Reichheld has devised a simple mathematical formula to summarize scores - he takes the percentage of customers who are Promoters, and subtracts the percentage of customers who are Detractors. (Note that Neutral customers are assigned a value of zero and left out of the equation.) The result is the Net Promoter Score, which the author claims is the only metric a company needs to predict growth.

Among the attractive aspects of NPS is that it's simple, easy to understand and can be disseminated across an organization with relative ease. Common sense tells us that if a high percentage of customers indicate an unhesitating propensity to recommend a company or product, that company's sales force enjoys a built-in extension of its efforts. And, in today's complex business world, being able to rally the troops around a single metric is appealing to any executive.

But, many critics have argued, is NPS really the only metric a company needs to predict growth? Further, given the means by which NPS is gathered and calculated, is it even an accurate metric? And whether it is or not, predicting growth is one thing, but does NPS hold any promise for driving growth?

Additionally, while a handful of B2B companies have recently adopted the NPS concept, available evidence suggests that the majority of companies who have embraced it are in the consumer products arena. Given the lack of any sort of appreciable history to date, on what basis can the assumption be made that NPS is likely to work in the far more complex business-to-business marketplace? After all, word of mouth referrals may have much more relevance in the rental car or financial services businesses, than in a highly specialized, technology driven industry whose products are far less likely to be discussed among family and friends over a cup of coffee.

And then there's the biggest question of all.

Is NPS the most accurate way to predict customer behaviour? Interestingly, Mr. Reichheld himself acknowledges that it is not.

All of which leads to the following discussion.

NPS - Do The Numbers Really Deliver?

First off, let us acknowledge at the outset that just as Mr. Reichheld is in the business of selling customer satisfaction surveys, so are we. The primary differences between us are that we conduct surveys in a very different manner than Mr. Reichheld is used to (or even familiar with, I would wager), and we only conduct surveys. We leave books and consultative promotion to others.

We set out over 18 years ago to develop a B2B customer satisfaction survey that would outperform the conventional survey methods of the day - paper, telephone and (for a spell) web. In the subsequent years, we have developed a survey process that is widely recognized as second to none in terms of response rate, candid replies, and highly actionable data. Today that process, known around the world as the InfoQuest Business Process Review, has been used to conduct over 82,000 surveys in 59 countries and 21 languages. That is pointed out merely to help establish that we've been around long enough to have learned a few things.

The foundation of NPS is the claim that it is the only metric a company needs to predict growth. We are struck by the carefully chosen language employed there, noting that the emphasis is on "predict", as opposed to "drive" or "cause". While there is a strong temptation to join the raging debate, and doubt, over whether NPS is causal in its effects on revenue change, or merely correlated, we believe there are other issues that warrant scrutiny. Let's take a look at them one at a time.

Does NPS Explore Customer Needs in Adequate Depth? InfoQuest surveys are comprised of anywhere from 36-60 questions. In our methodology, a series of satisfaction questions are posed to each participant, the responses to which establish basic performance benchmarks in 8-12 key customer touch points. An additional series of statements, to which respondents express levels of agreement, operate as drill-downs to provide insight into what can or should be to done to improve customer satisfaction in those areas.

The final question in every survey is always, "On an overall basis, how satisfied are you with our company?", which we have found is the single most reliable metric for reflecting customer satisfaction, and which we have been able to quantify as the single most accurate predictor of future revenue behaviour. We'll touch on that shortly.

The reason for posing the overall satisfaction question last is that doing so produces the most considered - and consequently, most accurate - response to that critical question. By first walking a participant through the various touch points that comprise the overall business relationship, we enable responses to the overall satisfaction question which take into account the complexities of the entire relationship. Confirming the need for that approach, tests in which the overall satisfaction question has been posed first, instead of last, have produced a 25%-30% increase in scores.

The resulting question is, can a single metric, posed on a stand-alone basis, accurately reflect customer sentiments on any topic? When considering whether one might recommend a brand of toothpaste, a fast food chain, or a particular airline, perhaps it can. But in the B2B environment, where relationships are driven by a far greater number of factors, it is absolutely vital to identify and measure satisfaction with each of the elements that comprise the overall business relationship. Performance in such areas as quotation procedures, delivery schedules, sales performance and design input, to name just a few, significantly impact and contribute to overall customer satisfaction and loyalty. In other words, a customer's willingness to recommend a company or product is not a freestanding outcome; rather it is a product of many different factors.

Mr. Reichheld asserts that NPS is the only metric needed to predict growth, but he falls far short of proving, or even claiming, that it can help achieve growth. The reason is clear. When relying on a single metric, the unavoidable trade-off is to sacrifice the ability to produce change. You may learn that a high percentage of customers would be unlikely to recommend your product or company, but what good is that knowledge if you can't do anything with it? Being told what customers think, but not why they think it, falls into the same general realm as a doctor telling you you're sick, but failing to provide a specific diagnosis or a recommended treatment.

In our view, the clear problem with expecting a single question, or even a short series of questions, to produce an accurate business-to-business response is that the approach lacks depth, frame of reference, or any sort of experiential relevance.

"Likely" to Recommend vs. "Willing" to Recommend. NPS utilizes the question, "How likely are you to recommend us to a friend or colleague?" The implied message is, to what degree can we expect you to go out and overtly suggest our company or product to others?

An individual's "Likelihood to Recommend" is influenced by many factors, including that person's overall predisposition to recommending anything, the potential audience to whom such a recommendation might be made, and even the desire of the person making the recommendation to appear "in the know". While many consumer-based product or service discussions can and do occur in casual settings - back yard conversations about a new movie or popular new restaurant may be commonplace - the same cannot be said for discussions about suppliers of control valves or thermoplastic compounds.

Though the difference is subtle, our approach has always been to pose the statement, "I would recommend your company to a friend or associate", utilizing a simple, non-subjective four-point response scale ranging from "Fully Agree" to "Fully Disagree". For the B2B respondent, the implied question is not will you go out and actively recommend us, but given the opportunity, do you have adequate faith in our performance to stake your personal credibility on recommending us? Given the difference between a routine consumer purchase and a multi-million dollar manufacturing contract, "willingness" to recommend is a much more appropriate measure than "likelihood" to recommend.

Is the NPS Metric Really Accurate? As stated, NPS groups respondents into three broad categories based on responses to a single question, which is answered on a 0-10 scale. Among the questions that must be considered when trying to understand NPS are, where did that scale come from, and is it accurate?

Numeric scales inherently tend to produce patterned and predictable results. Specifically, responses tend to be influenced by deeply ingrained experience with school grading systems wherein any mark below 60-70 was generally considered failing and/or unacceptable. Anecdotal conversations with hundreds of companies over the past ten years have consistently revealed that the majority of 5 point numeric scales seem to consistently generate average responses of just under 4 (or about 75% on the academic scale), while 10 point scales consistently generate average responses of right around 7. Given that history, if a numeric scale has to be used at all (and doing so would not be our choice), we will at least agree that any score of 6 or less (aka "Detractors") should probably be viewed as a failing mark.

But, from a revenue perspective, should angry detractors who rate the company a "0" on a scale of 0-10 be weighted the same as uninspired customers who give it a "6"? While both may arguably be viewed as failing marks, there is failure, and then there is monumental failure. We've all seen the studies that have consistently suggested that a dissatisfied customer will, on average, tell 8-10 other people about their negative experience. Yet is it logical to assume that there is but a single, universal degree of customer dissatisfaction, and that everyone who falls under that definition will behave in the same manner?

In an attempt to answer that question, let's take a look at some numbers.

First, to avoid subjective interpretation and predictable outcomes attendant to alpha and numeric scales, InfoQuest surveys utilize a simple, unambiguous and multi-cultural scale, which is as follows:

Satisfaction Questions Drill-Down Statements

Totally Satisfied Fully Agree Somewhat Satisfied Partially Agree Insufficient Information to Evaluate Insufficient Information to Evaluate Somewhat Dissatisfied Partially Disagree Totally Dissatisfied Fully Disagree

In preparing this discussion, we reviewed the results of our most recent 40,000 sets of B2B customer satisfaction survey responses. We pulled the data for the single statement, "I would recommend your company to a friend or associate". The aggregate responses were as follows:

No. Respondents % Respondents Fully Agree 22,225 58% Partially Agree 12,364 32% No Response 2,313 - Partially Disagree 2,538 7% Fully Disagree 1,098 3% Totals 40,538 100%

On a per company basis, looking at the percentage of respondents who Fully Agreed, we found the following:

High Company Score 100%

Low Company Score 3%

Median All Companies 63%

Looking at those numbers, we have concerns about the factual basis on which NPS performance claims are being made. Specifically, in his book, Mr. Reichheld claims that the average NPS score is under 10%, which bears little resemblance to our own numbers.

Admittedly, a direct comparison can only be made with a litany of caveats. We acknowledge that the numbers shown above utilize a different metric, a data collection method that avoids the inaccuracies of telephone and paper surveys, entailed a slightly re-worded question, and was built into a full survey, not presented as a stand-alone. We also have to take into account the fact that The Ultimate Question was just published in 2005, that it was based primarily on B2C data at the time of publication, and (in our view) was lacking in empirical evidence and hard facts.

Is there, buried in that mix, an adequate explanation for the tremendous difference in outcomes? Or is the NPS metric grossly understated as a way to create a sense of need for an otherwise unsupported and perhaps unsupportable literary marketing ploy?

See if the next section suggests any answers to those questions.

NPS - Right or Wrong, it is Needed? The following question was found on 2 October, 2006, on a blog* authored by Fred Reichheld:

"Can a one-question survey predict growth as accurately as a long survey?"

This was Mr. Reichheld's response to that question:

"If you can convince a customer to spend time answering dozens of questions, you can predict that customer's behaviour more accurately than you can with one question. The problem is, most customers in this busy world won't give you that much time - witness typical survey response rates from 2% to 20% - and you couldn't afford the surveying and data processing expense if they did.

In B2B the problem is even thornier, because the senior execs who drive purchase decisions are the least likely to tolerate lengthy surveys"

* http://netpromoter.typepad.com/fred_reichheld/2006/07/questions_about.html Online at 10/2/2006

Our response? The average InfoQuest response rate over the last 15 years stands at 74% in North America, 70% in Europe, and 72.4% globally. That response rate, built entirely on B2B survey activity, is based on (typically) top revenue accounts, and the senior level decision makers within those accounts. It is also based on the delivery of from 36-60 (or more) questions and statements in each survey.

Which leads to the unavoidable question - if a survey entailing dozens of questions will predict customer behaviour more accurately than can be predicted with one question, why would anyone settle for just one?

Quantifying the Outcome - An Alternative

Years ago we developed a statistical model that identified and quantified the correlation between customer satisfaction and revenues. Not referral potential, but actual cash expenditures. Based on the analysis of over 20,000 worldwide customer responses gathered over a three year or greater period of time, and after comparing those results to account based revenue histories over the same period of time, some staggering conclusions were reached. Specifically, over time -

A Totally Satisfied Customer contributes 2.6 times the revenue to a company that a Somewhat Satisfied Customer contributes.

Totally Satisfied Customer contributes 14 times the revenue that a Somewhat Dissatisfied Customer contributes.

A Totally Dissatisfied Customer decreases revenue at a rate equal to 1.8 times what a Totally Satisfied Customer contributes to a business. That finding was based on not only loss of existing account revenue, but on the additional impact brought about by negative referrals.

To put that into more user-friendly terms, the chart below shows the relative percentage of revenue contribution, over time, for varying levels of satisfaction. Assuming each of your customers had one dollar to spend on your particular product or service, the chart shows how much of that dollar you can anticipate receiving.

Survey Index - Normalized to Simplified Result One Hundred Percent Ratios

Totally Satisfied 100% 1.0

Somewhat Satisfied 38% .4

Somewhat Dissatisfied 7% .1

Totally Dissatisfied -180% (2.0)

Look at those numbers again. A Totally Dissatisfied Customer decreases revenue at a rate equal to twice what a Totally Satisfied customer contributes. In other words, you can have twice as many satisfied customers as dissatisfied customers and still be losing ground.

The culmination of those findings is presented in The Revenue Index, which has been a standard element in the deliverables of InfoQuest since 1996.

NPS theorizes that a detractor effectively negates the impact of a promoter, and that everyone else, classified as passives, represents no impact at all. The Revenue Index, however, has clearly established that totally Dissatisfied customers have far more impact than merely negating Totally Satisfied customers, and that

the rest of your customers, call them the passives, still generate financial impact, albeit limited in nature.

The fundamental difference, of course, is that NPS posits how your customers may discuss your company or product to others. The Revenue Index identifies and quantifies the percentage of existing available revenues you can expect to receive from those customers over time. The difference is undefined theory versus actual cash behaviour.

A final key consideration is that Willingness to Recommend and Overall Satisfaction do not directly correlate to each other, which is a major difference relative to the B2B versus B2C marketplaces. We previously showed an analysis that revealed how 58% of 40,000 respondents Fully Agreed that they would recommend the company being discussed. However, only 39% of those same respondents indicated that, on an overall basis, they were totally satisfied with the same company.

Why? Because as previously discussed, there are many factors that affect overall satisfaction. In a highly technical manufacturing or service business, it is not uncommon for companies to be valued for their engineering skills, but to be found lacking from an administrative standpoint. Customers may very well be willing to recommend the engineering acumen, but still be open to making a change if a better overall package of service becomes available.

Summary and Conclusions Any tool that will help a company increase customer satisfaction and loyalty is going to be looked upon favourably. If it can establish an initial benchmark, help the company monitor progress over time, and provide an actionable means of better providing for customer needs, it is going to catch the attention of a lot of business operators. Just remember the old adage - if it sounds too good to be true, it probably is.

Satisfied and loyal customers are the product of a corporate commitment to excellence, plain and simple. To effectively respond to customer needs and desires requires top-down support, bottom-up commitment, a current, candid and detailed view of customer opinions, and valid metrics. Customer satisfaction is just like any other enterprise or activity; if you can't measure it, you can't manage it.

With very few exceptions, building a satisfied customer base is the product of a company taking a good hard look at itself through the eyes of its customers, and then going out and systematically addressing, and fixing, what it sees. Success is predicated on understanding each of the many dynamics that comprise and contribute to the customer relationship. It is an outcome that is driven, not pulled.

The attraction to NPS is its perceived simplicity; propelled by the claim that it is one metric that tells you everything you need to know. The problem is, there is no single metric that can live up to that claim, and that includes our own. That's not to suggest that NPS represents anything inimical to the health and well being of any enterprise; merely that it is less than it is being sold as. Might it be useful as one of many tools for monitoring customer sentiments and behaviour? Yes. Is it powerful or accurate enough to be used as the only tool? Absolutely not.

Mr. Reichheld would have us believe that if you concentrate on building a high score to a single metric, everything else will follow. While that makes for attractive theory, the reality is, it just doesn't work that way. Tracking change in an organization is one thing. Driving change is another matter entirely. To bridge the gap between collecting information and actually putting it into play, a customer survey needs to entail several key components.

1) It needs to explore all of the dynamics and touch points that comprise and contribute to the customer relationship. No single question, posed in a vacuum, will produce an accurate set of responses. Customer loyalty and recommendation behaviour are products of satisfaction with the total customer relationship. They are not, and cannot be effectively dealt with, as a free-standing outcome.

2) Information produced by a survey must be actionable, as opposed to merely interesting. Survey data and summary metrics need to be clear, non-subjective and unambiguous. Scales and data collection methods that tend to influence replies or generate predictable outcomes must be avoided. Metrics need to be based on fact, not supposition and theory. Arbitrarily calculated combinations or groupings of responses - such as assuming that all detractors are created equal - merely produces informational clutter while obscuring the true opinions of individual customer respondents.

3) It needs to utilize metrics that are quantifiable. Among our objections to NPS is the total lack of support data attached to its claims. Rather than, improve this metric and growth rates will increase, focus on tools like the Revenue Index, which provides a reliable measure for predicting and tracking the impact of specific actions and outcomes on revenue contribution.

4) It needs to ultimately provide a clear sense of direction. They key to any survey is not in learning simply what customers think. You need to learn why they think it, and how to most efficiently and effectively change their opinions for the better.

Net Promoter Score - The Search for the Magic Pill

NPS has been heralded by its creator as the proverbial magic pill. It is our belief that no such pill exists. Developing satisfied customers takes dedication, commitment and work, plain and simple. Tracking an arbitrarily derived number is not going to get the job done.

Howard Ploman President InfoQuest International

www.infoquestcrm.co.uk

ENTERPRISE RENTAL COUPONS

0 comments:

Post a Comment