3 Factors that Drive Customer Satisfaction Rating

When it comes to customer satisfaction rating, more of everything isn’t the answer. From regression analysis of years worth of customer satisfaction rating and from related works done by others, we find that customer satisfaction is driven by 3 basic factors (from stated rating studies):

  1. Buying experience: How easy it is to evaluate choices and complete the buying process? Customers treat buying experience as part of the product experience. While rational thinking dictates that these costs are incurred once and should be treated as sunk by the customers, research(Journal of Management Information Systems Winter 2007-08) shows that these costs remain sticky and customers treat buying experience as part of the product experience.
  2. Delivering what is promised: Does the product quality and its realized benefits match what was promised and most importantly what the customer expected it to be? This is not about delighting the customers are delivering more that what is promised. A customer who walks into WalMart has one set of expectation and the one who walks into Nordstrom has another. For the segment you are targeting, the product benefits must match your positioning and messaging.
  3. Experience when things go wrong:  In the case when things go wrong, customers need support, how easy it is to get support and how they are taken care of. No customer believes things will never go wrong but the type of support they receive and how the problems are handled are what customers treat as relevant to their overall satisfaction rating. For example, a Corolla customer does not expect the dealership to send a loaner car and tow-truck for services, but a Lexus customer does.

Go head test this out today. Run a very simple survey of 4 questions to your customers, (use 1-10 scale)

  1. Please rate your overall satisfaction rating with our products and services.
  2. Please rate how satisfied you are with your buying experience (how easy it is to find what you need, evaluate options and complete the buying process)
  3. Please rate how satisfied you are with our product quality (meeting your expectations, delivers what was promised)
  4. Please rate your support experience (ease of getting help, timeliness, how you were treated)

Run a regression using (1) as dependent variable and the rest as independent variable and you will find out how relevant the 3 factors are to your own situation.

Caution: Regression analysis still only finds correlation. There are numerous lurking variables that were not fully studied. But research from other data sets make it more likely that these variables have causation relation to customer satisfaction.

One question to determine a startup’s success?

Let us stop looking for a single predictor of success, be it a startup or an established enterprise. I saw a TechCrunch interview of Mr. Peter Thiel, of Clarium Capital.  The TechCrunch post is titled, “Best Predictor of Startup Success Is Low CEO Pay”, and Mr.Thiel was quoted as saying

In practice we have found that if you only ask one question, ask that.  (CEO Pay)

This is the classic single question trap. There cannot be single metric that can be “the best predictor”.  You should ask

  1. Does correlation mean causation?
  2. Is there cause-fusion – i.e., do successful startups pay low CEO salary?
  3. What about all other lurking variables? What if there is another variable that drives CEO’s low pay and success? (Omitted Variable Bias)
  4. If low salary is good, is lower salary better? (reductio ad absurdum)
  5. What about other startups that have CEOs with low pay and still fail?
  6. A startup, hamstrung by lack of resources and low cash flow may pay low or no salary, is that still a predictor of success?
  7. What about the congruence between the startup’s strategy and the needs of the market it serves?

Let us not forget  Predictive Analytics slippery slope. If you want to ask a startup questions, here is my list (not claiming predictability):

  1. What jobs will your customers hire your products for?
  2. Who do they hire now, i.e., who do they have to fire first?
  3. What are their alternatives?
  4. How much will they pay for it?
  5. What budget will that come from and how big is that budget?
  6. Where do they post the job opening?
  7. Where do they look for candidates and can you go there without considerable costs?
  8. What is their hiring process?
  9. What will they find compelling about your product’s candidacy?
  10. Will the job exist two years from now?

Analytics In Fantasy Football

Does the first week performance of your quarterback predict the performance for the rest of the season? This is a question posed and answered by Nando Di Fino in a column in The Wall Street Journal. I do not own a Fantasy Football league but I am excited to see analytics and econometrics being applied to player picks and trades. This is a very well written article and has some data analysis behind it. But I question the broad implications made by the article.

  1. Are the metrics used by CBSSports the right one to measure player   calibre and predict their performance?
  2. Is there omitted variable bias here? Could there be another  underlying factor that defines the overall season performance?
  3. I know Nando did not mean  causation but i think the   statement “Will a poor performance by a starting wide receiver
    foreshadow a season-long letdown? ” could be interpreted as such. If the  first week performance is indeed predictive of overall performance,  then the respective R-square values are very low (square of correlation coefficients) to fully explain the full season  performance.

This reminds me of a quote attributed Einstein, “Everything countable does not count and everything that counts is not countable”.

Fact Based Decision Making

IBM acquires SPSS:

“The opportunity is to move from sense-and-respond decision-making to a predict-and-act model,” said Ambuj Goyal, a computer scientist who is the general manager of I.B.M.’s information management business.

The growing appeal of the sector, Mr. Davis said, reflects the increasing pressure by the senior management of corporations for “tighter, fact-based decision making, especially in this economy.”

Other independent analytics software makers may well become takeover targets, said Mr. Evelson of Forrester. Among the candidates, he said, are Accelrys, Applied Predictive Technologies, Genalytics, InforSense, KXEN and ThinkAnalytics.

The broad consolidation wave in business intelligence software, analysts say, will bring increasing price pressure on some segments of the industry as major companies seek to increase their share of the market. And the open-source programming language for data analysis, R, is another source of price pressure on software suppliers.