Do you really need super charged analytics to gain insight? You say the word “insight” and people instantly think you need to engage in super-statistics such as multi-nomial logit modeling with recursive Bayesian based Gibbs sampling to find it. In reality, some of the best insights, like some of the best food, are relatively simple.
Let’s take the example of scores. In the customer experience measurement (CEM) world the mean or “average” score is so common it is often expected in reporting. Sure we do really wild stuff sometimes like subtracting promoters from detractors, but at the end of the day we are still dealing with means. However, there is meaning to be found in other measures of central tendency and dispersion as well.
One application for the underdog of descriptive statistics is the measure of standard deviation. For those who are a little foggy on your 200-level stats class from your undergrad days, the standard deviation is a simple metric of score dispersion or variability. The higher the number, the more variation in scores. The lower the number, the less variation within the scores.
Now let’s say you are managing a brand that demands high consistency in standard operating procedures. Most strong brands are very good at this and reap the rewards. A Coke is supposed to taste like a Coke no matter whether you are in California or Calcutta. In consistency we find certainty. People like certainty.
So why don’t we start looking at the customer experience the same way. The goal here is to not only do well, but to do consistently well. Now, those who do well in general are consistent, which is why they have higher scores. However, there is a case to be made to favor consistency over a great experience. If you are certain about what you are going to get, even if it is mediocre, at least you know what you are getting.
Below is a chart created from the data we collect in CEBenchmarks™. This is a small subsection of the data that looks at customers’ evaluations of each budget hotel’s check-in process.
We can see cases where certain hotels do fairly well consistently (Rodeway Inn, Budget host), ones that do well with higher variability (Budgetel), ones that are evaluated low with relatively lower variability (Travelodge) and one with low scores and high variability (Knights Inn). So if I am brand manager, what I want to be able to is get more scores to be both consistent (low variability) and high. In fact, I would argue (and we have empirical data to support it ) that is is better to be consistently mediocre than inconsistent but good. At least people know what to expect in the former case, in latter case…you just never know. Microtel seems to occupy that space in the economy segment of lodging delivering a simple but consistent experience. Any frequent flyer will tell you, they can handle known inconveniences much more readily than unforeseen ones. I know that airline inconsistencies drive me crazy.
To further illustrate the point I took the luxury hotel Ritz Carlton and also plotted it among the budget hotels. An unfair comparison I know, but you can see that they are both very well-rated and very consistent in their check-in scores. Like most high-end service providers, they are hitting it consistently out of the park. Their high scores are, by definition, driving down variability. I would also guess that those hotels that control variability vs. ratings probably are much more profitable and efficient. But that’s a topic for another post…
What do you think? Unintuitive as is, would you rather provide a consistently-simple experience than one that is perhaps fancier but uneven?
