Customer Loyalty Mathematics

Dr. Mark Klein is a physicist who is using his mathematical skills to help companies tune their marketing programs. He has developed some mathematical models which, according to Dr. Klein, will improve the effectiveness of marketing campaigns. The reason I have chosen to mention this is that it is rare to find someone who is specifically mentioning the use of mathematics to understand customer loyalty. What appears most often in the market is different companies and consultants saying they do customer loyalty surveys better than others. In most cases the difference between the various groups (companies or consultants) is negligible.

The good news is that Dr. Klein mentions some mathematical techniques that he uses. The bad news is he doesn't say enough for someone like me to understand those techniques and be able to validate them.

He has a "Field Guide to Mathematical Marketing" available from his web site (55 pages). The book, which can be downloaded from his website at no cost, does an excellent job of making the case for using mathematics to better understand your customers. For that reason alone, the book is worth the time to read, especially if you think that mathematics has no place in analyzing customer loyalty.

Unfortunately, there are no mathematics in the document. The document does contain vary good descriptions of the statistical applications that appear to be one of the key tools he uses (factor analysis, factorial designs and fractional factorial designs). I am surprised he doesn't mention some of the other experimental statistical designs such as Latin Squares and Graeco-Latin Squares as long as he is using experimental designs. In any case, as I learned in school, it is always a good idea to know what is inside a black box (for those not used to the term, a black box is something that performs but you don't know how or why) before you buy it and thus until I know more about his application of the various methodologies noted, I am limited in the amount of support I can give. One of my concerns is what assumptions are required.

Some interesting observations made by Dr. Klein:
1. He believes that customer loyalty to be a more predictive measure of customer purchasing than customer satisfaction.
2. He also believes that customer satisfaction is a backwards-looking measure whereas customer loyalty, as noted above is more predictive of the future behavior of your customers.
3. He believes that most people in marketing (and probably in customer service/support) avoid using mathematics even though it would be beneficial.
4. He is a strong proponent of selling to existing customers and believes that the more information that you have about you're customer, the better job you can do in developing marketing programs to encourage them to buy.

Finally, I reviewed the software program on his site that allows a demonstration. Here is where I am a little concerned about the "black box." I am reminded of the phrase from my programming days that said "garbage in, garbage out" was corrected to say "garbage in, gospel out." The demonstration is excellent without being too complex for the non-mathematical observer, but it is not clear what is happening and why. Of course, that may be the idea, to sell the concept without giving away the company secrets.

The bottom line is that I am not recommending Dr. Klein's mathematical modeling because I do not have sufficient information to evaluate how he is using the techniques. I am, however, pointing out that there are a few people with enough education, such as Dr. Klein, to use the power of mathematics to help companies improve their understanding of their customers and customer loyalty.

The Real Impact of Customer Complaints

John Goodman, past president of TARP and currently president of e-satisfy has given some factual data behind some of the statistics that are bandied around in the customer service business. I was glad to see his statistics and understand where they came from. While this data is relatively old, it contains data that I see in the blogs on a regular basis. (This might suggest that many people are satisfied with using old data and just using the phrase "research has shown." The information shown below was published by John Goodman in Competitive advantage, June, 1999.

1. 50% of consumers will complain to a front line person and in the B2B environment the number is 75%.
2. Only 1 to 5% will escalate the complaint to a local manager or corporate HQ. For large ticket items 50% go to the front line and 5 to 10% escalate to local management or corporate HQ. Interestingly, if the company has an 800 number the percentages double. The most important statistic is that only 1 out of 100 to 500 complaints will ever be seen by a senior executive.
3. If the complaint results in out-of-pocket loss the complaint rate is 50 to 75%. Otherwise, the complaint rate for mistreatment, quality and incompetence evoke only a 5 to 30% complaint rates.
4. Tom Peters used to suggest that for every bad experience 10 people will be told and for every good experience only 2 people will be told. From this he noted that bad news travels faster than good news. John Goodman's data suggests that the ratio is really 2 to 1. The original study for Coca-Cola in 1981 showed that a median of 5persons heard about a good experience and 10 heard about a bad experience. A later study for Coca-Cola found that resolution of a problem on the first contact improved loyalty 10% higher than resolution via multiple contacts. A similar study for a domestic auto manufacturer found the numbers to be 8 and 16 respectively.
5. The general rule of thumb I have heard for many years is that it costs five times as much to acquire a new customer as it does to retain an existing customer. Once again, John Goodman gives us the history. They studied new auto customers and found that there was a cost of $375 in advertising for each auto sold. Since the company had a 50% base of loyal customers, the actual advertising cost to acquire a new customer was actually $750 per auto sold. The goodwill expense to retain a customer was $150. Hence the ratio of cost to acquire a new customer versus the cost to retain a customer was 5 to 1.
6. One of the final statistics is one that seems to vary from person to person. According to a study first published in 1988, customers who complain and are satisfied are 8% more loyal than if they had no problem at all. From my studies, I have found similar statistics.
7. I have heard the statistic that 68% of all people who defect leave because of a poor employee attitude. The statistic that John Goodman offers, that is based on a study by TARP suggests that 20% are caused by employee actions, 40% by corporate products and processes and 40% are caused by customer mistakes or incorrect expectations.

The bottom line is that some of the more often quoted customer complaint statistics seem to be constant and do not change with time. At least now we know where some of these "rule of thumb" statistics were derived. The statistics noted above were derived from studies while John Goodman was president of TARP.