We'll be discussing the United passenger ejection for years, the incident was that bad. Worse, it was unnecessary. Circumstances were stressful, but the whole mess could have been avoided, had United cared to evaluate and respond to flyer expectations.
Those of us concerned with customer experience (CX) – every manager, business analyst, and executive policy-maker should be – need to examine what went wrong in Chicago. Why did United's procedure for overfull situations fail? Why did no passenger voluntarily accept the offer of a flight voucher and rebooking to leave the plane?
Every frequent flier knows the answer. Vouchers aren't cash. Actually, they're a pain to use. So while $800 is a fair sum, a strings-attached $800 voucher isn't worth that much. The perceived value – the value that mattered – was far lower than the value of keeping a seat, hence no volunteers, hence failure.
CX practitioners should factor 'perceived value' into evaluations. This article will examine why and how, proposing a perceived value score (PVS) to complement customer satisfaction, Net Promoter, customer effort and sentiment measures.
Defining perceived value
'Perceived value' (PV) is qualitatively defined as "the worth that a product or service has in the mind of the consumer." Let's quantify it as 'cost' weighted by 'expectation'.
My definition of expectation is going to be a bit complicated, but the basic PV notion is simple: combine fact and attitude to determine whether or not people think they're getting – or will get – their money's worth. Evaluate PV pre-sale to understand intent, how likely you are to close the deal. Or do it post-sale as a satisfaction, loyalty, and advocacy predictor, and to suggest a customer's transaction-effort tolerance.
'Perceived value' post-sale is similar to satisfaction, which is often measured via a customer satisfaction (CSAT) survey: how did your stay at the Ritz rate on a scale of 1 to 5?
Say Richie Rich rates that Ritz a 5 overall, Superb! Nice going if the typical (median) rating for a comparably upscale room is a 4. But say Richie paid a third more than he's used to paying for this quality, $800 rather than $600. We have an expectation factor of 1.2 (rating) times .75 (cost), or .9.
So the 'perceived value' is $800 times .9, $720. But at $800, the room cost is about 11% higher than that $720 'perceived value' so Richie feels slightly taken.
Factor in customer effort and sentiment
Customer effort could be factored into expectation. What value did United's passengers ascribe to an $800 voucher, given both the inconvenience of switching to a later flight and the strings attached to a flight voucher?
I'd bet that United employs smart data scientists who could create a suitable model. After all, the exercise is very similar to price modeling, which also responds to very dynamic conditions that involve seat inventory, demand modeling, costs, and temporal factors. Maybe they would have found that a voucher, in conditions like those in Chicago, gets a .4 effort factor: An $800 voucher would have $320 'perceived value'.
What value did United's passengers ascribe to an $800 voucher, given both the inconvenience of switching to a later flight? I'd bet that United employs smart data scientists who could create a suitable model.
You could also factor in sentiment, which quantifies an customer's opinion – or the aggregate of customer attitudes – about a product or service. Sentiment is often expressed as a positive or negative valence coupled with intensity, although the state of the art is rapidly moving toward emotion AI, noting that emotion categories – angry, happy, sad, and so on – are often better customer behaviour predictors than are cruder positive/negative evaluations. (For more on all this, covering the range of customer and market insights factors, check out a conference I organise, the Sentiment Analysis Symposium in New York.)
I will add, finally, that for PV computations, we'd hope to be able to discern fine-grained sentiment, at a product or aspect level – see my How four AI startups are helping brands exploit customer reviews for a take on aspect-level sentiment – but that brand-level sentiment could also be useful. At the risk of oversimplifying: 'brand equity' is pretty much the 'perceived value' of a brand.
From measure to score
Each of the measures I've cited has a formalism behind it.
The Customer Effort Score (CES) was introduced by the Corporate Executive Board. The CES core question is "How much effort did you personally have to put forth to handle your request?" and the CEB claim the resulting score "is 1.8x more predictive of customer loyalty than customer satisfaction measures, plus it is two times more predictive than Net Promoter Score (NPS)."
And NPS, developed by jointly by Bain & Company and vendor Satmetrix? The Net Promoter Score quantifies "willingness of customers to recommend a company's products or services to others" (citing solution-provider Medallia's definition). NPS is often used as a customer satisfaction proxy. It dismisses satisfaction surveys as simplistic and instead asks about a possible behavior, "How likely are you to recommend...?" NPS is computed by subtracting the percentage of "detractors" from the percentage of "promoters," resulting in a score ranging in normalised value from -100 to +100, from 100% detractors to 100% promoters.
A Perceived Value Score, as sketched out here, will nicely complement Satisfaction, Net Promoter and Customer Effort initiatives.
Similarly, i-SCOOP proposes Net Easy as an aggregate, normalised CES value.
These index values facilitate comparisons, important because ability to compare – to relate and rank multiple values – conveys insights that single, isolated quantities can not. It's competitive position, changes over time, and correlation of values with events that's interesting.
So let's turn PV into a score, similarly scaled to a normalised -100 to +100 range.
Perceived value score, barriers, motivation and activation
The perceived value of a product or service is a cash quantity. For instance, the PV of an $800 United Airlines voucher might be $320 cash. That gives us a Perceived Value Ratio (PVR) -0.6, computed by ((Perceived_Value - Nominal_Value) / Nominal_Value) equals -480/800. That's a Perceived Value Score (PVS) -60%.
Consider another example, a $100 restaurant Groupon deal that cost $50. Here the PV is $90 given the expiration date, but since the cost was $50, we get a PVR of 0.8, an 80% PVS. Those Groupons are going to be popular, unlike the vouchers. If the aggregate PVS is above zero, you have positive motivators in place. If it's negative, barriers to action overcome motivating factors.
Those barriers could be price, customer effort, availability limits or delivery delays, quality perceptions, competition, or any number of other factors. The key point is that they demotivate the consumer. The joy is that the same techniques used to quantify sentiment – text analytics via natural language processing, applied to reviews, survey verbatims, social postings, contact-center audio and transcripts, and the like – can be applied to study barriers.
So model barriers as contributors in the Perceived Value = Cost * Expectation equation, and adjust the cost and improve expectation factors until your PVS goes positive, or at least better than the competition's.
If a product or service over-delivers for the price paid, the customer is happy. We have positive PVS, making more likely a positive customer experience. If the product or service under-delivers, the PVS and CX are negative. (One caveat, however: deficient perceived value will have little effect on purchase – just as low satisfaction will not affect loyalty – when the consumer has few choices for a must-have product or service. That's one reason Why Companies Like United Think Customer Experience Is Irrelevant, per an article I published recently.)
A Perceived Value Score will quantify over/under-delivery, using attitude-measurement text analysis techniques that can be extended to discover motivations, behavior root causes, activating factors, and barriers. A Perceived Value Score, as sketched out here, will nicely complement Satisfaction, Net Promoter and Customer Effort initiatives to anticipate, and not just react to, challenges both extraordinary and ordinary.
For more on review customer experience analyses, consumer insights, and related topics, check out the 2017 Sentiment Analysis Symposium, taking place June 27-28 in New York, tagline Emotion–Influence–Activation, focusing on the business impact of sentiment data and technologies.
About Seth Grimes
Seth Grimes is the leading industry analyst covering natural language processing (NLP), text analytics, and sentiment analysis technologies and their business applications. He founded Washington DC based Alta Plana Corporation, an information technology strategy consultancy, in 1997. Seth created and organizes the Sentiment Analysis Symposium. He consults, writes, and speaks on business intelligence, data management and analysis systems, text mining, visualization, and related topics.