David Maister - Professional Business, Professional Life
David’s ResourcesAbout David
NEW! Browse my materials by topic of interest:StrategyManagingClient RelationsCareersGeneral

Passion, People and Principles

What if Your Customers Can’t Be Trusted?

post # 398 — July 5, 2007 — a Client Relations post

I was participating in a discussion with a group of executives, when an insurance company CEO pointed out that, according to their market research, about 30 percent of their customers acknowledged that they would “cheat” on an insurance claim. (He didn’t elaborate on the precise details: it wasn’t that kind of meeting.)

But assume his research is correct: what is the appropriate company response? In most cases, as we know, companies will then get very suspicious of ALL of their customer claims (they can’t know which 30 percent are untrustworthy) and you end up with bureaucracy and MUTUAL distrust, which quickly spirals down.

Insurance companies get a bad rap (Hurricane Katrina, Mike Moore’s new film – Sicko) for too often denying claims. But the fault is not just on one side, is it?

Put yourself in the shoes of being an insurance company exceutive. Is there a middle ground between over-trusting a customer base which will exploit your goodwill 30 percent of the time, and acting defensively all the time and coming across to everyone as non-responsive?

There’s clearly a difference bewteen what you wold do as an individual, on-on-one, when you can take it case-by-case. But what do you do if you’re a corporation, trying to work across the country or internationally with hundreds of thousands of customers. What policies do you put in place, and how do you train your front-line people?


Carl Singer said:

AN answer (not THE answer) in two words is “data mining”

Given the specific of “hundreds of thousands of customers” you should have a wealth of statistical information about your customers and their behavior, and should be able to employ data mining and other statistical techniques to detect fraud and other patterns.

Off hand, these tools could (1) help evaluate trustworthiness of customers based on claim pattern, demographics, etc. (perhaps a slippery slope.) (2) help evaluate validity of claims based on the claims themselves.

[FYI - I built / co-taught the data mining course at IBM's Advanced Business Institute -- I did intro & framework & sample applications, others did the heavy lifting re: data mining tools & technology.]

I realize that this “technical” response has little to do with relationships and trust-building — but unless someone (an agent, perhaps) has a 1-on-1 relationship that is moot. Trust building may become a mass marketing exercise both positive (do the right thing) and negative (Johnny lied and Johnny went to jail.)

posted on July 5, 2007

ashutosh wakankar said:

On the face, it looked like a really thought provoking question but the more i thought about it, i am a unsure if i get what you are driving towards. How is the situation different for a CITIBANK loaning money to millions of people worldwide? Given a choice many would not want to return it- so the banks have the mechanisms/procedures in place. Billions of people visit departmental stores and many would walk away with stuff without paying if that possibility existed- so there are security measures! Everything will add to the complexity of doing business and resources required but that is the inherent requirement of the business operation and companies as well as customers will play along with the rules of the game…how would the middle ground between trusting and verifying be different for an insurance company from a retail chain or a bank or any other business?

Like in every situation in life (with a very few exceptions), the principle would be to operate from trust-but verify!

posted on July 5, 2007

Christian ter Maat said:

To me the real question is; Why customers cheat on insurance claims?< ?xml:namespace prefix =" o" ns =" "urn:schemas-microsoft-com:office:office"" />

Answer 1; some customers are thieves and the insurance companies should identify those individuals through e.g. data mining or other verifications and then grey list followed by black listing them.

Answer 2; at least in < ?xml:namespace prefix =" st1" ns =" "urn:schemas-microsoft-com:office:smarttags"" />Europe; customers feel that although paying their insurance fees, their claims are often denied or not fully compensated. Insurance companies’ first demands are: to fill in long reports (along with police reports), you have to have witnesses, no receipts no valid claim, we only pay if… ABC (conditional service), takes many weeks etc.

If you sign up for insurance as a client, you are treated like a King; if you claim like a Crook. On top, insurance companies report huge profits and show off with their pyramid buildings and other Enron like corporate behavior.

Insurance companies need to address here the issue of reputation management.


posted on July 5, 2007

David Kirk said:

I’m not sure David M’s point has been addressed yet.

  1. Data mining, while a potentially tremendously useful tool in many situations, and definitely an aid in detecting fraud and managing the risks is not perfect, and will usually only pick up the significant or repeated issues easily. Even if a claim is tagged as questionable by the system, there are costs involved in establishing whether the claim is in fact fraudulent. Oftentimes the cost comes close to any potential benefit. One response by companies using data mining is that as soon as a claim is tagged as “questionable” the dogs of war are let loose and the customer is treated as a criminal. David M’s problem hasn’t been solved.
  2. Insurance is different from many other industries, including other financial services industries such as banking. I can’t claim that I actually repaid my home loan when I didn’t. There is certainly credit risk in lending, but it is controllable doesn’t hinge on trusting the customer to repay the loan. Non-repayment is obvious and has repercussions.
  3. Claiming for damage to a car, where the claim includes several scratches and bumps and dents that existed prior to the claim is not the same as non-repayment of a loan. Claiming for lost sunglasses when the claim event was “slipped out of fashion” is another.
  4. There is a chicken-egg argument here. Do insurers treat customers like crooks because they have learnt from bitter experience that some (30%?) are? Or do customers feel nothing at ripping off an insurer because they were treated like crooks in the first place? More on this later.

The Ring of Gyges

I love the story of the ring of gyges; I won’t take up even more space by repeating it here. The crux is that a lot of the time, for a lot of people, doing the right thing relates more to the downside of being caught rather than any higher calling.

Assumed to be crooks and staff training

Stealing from a clothing retailer is possibly closer to the core problem here. Why do shops tag their clothes and have detectors at the door? To prevent theft. One might argue that by implication everybody is untrustworthy. Staff are usually trained not to say “you stole that shirt” but rather “may I ring that shirt up for you, sir?”

Sometimes it is enough for staff to appreciate that not all the customers are crooks, and to keep that in the forefront of their minds. Communications along the lines of “I appreciate that this is a hassle for you, ma’am and you are already upset at the loss of [item of value]. However, we do have problems with accidentally or fraudulenty incorrect claims which means we have to charge premiums. The more we can be sure that we are only paying correct claims like yours (sic) the better our premiums can be. We wouldn’t want to reward those who make you pay higher premiums.

But these are specific points, and as David M suggested, one-on-one is easier than a company-wide standard. I suggest the following:

  • Ensure a culture of customers being important and being the lifeblood of the company. Externally directed adverts focussing on this image can have a surprisingly positive impact on the way staff perceive the company and the culture and the way they treat staff.
  • Maintain rigourous standards for detecting fraud and storing accurate data for data-mining (yes I do agree it can be useful!)
  • Have initial and ongoing training with a focus on putting staff in the shoes of the customer. Make this regular and consistent.
  • Make use of the claim-capture system to place hooks around “remember the customer may have been through a traumatic experience”, including a part of the standard wording to include “unfortunately, sir, this is a necessary part of the process because some people make fraudulent claims. It shouldn’t take long.” The honest customers will almost always understand that there is sense to the process. Customers are smart.
  • Make use of the recorded calls for training and case-studies. Again, focus on the staff getting the customers’ perspective.
  • Another tactic I use is for the person dealing directly with the customer to say “In order for me to process the claim for you I need to justify to my supervisor that we have followed the correct process”. Again, focus on the rationale for the checks and balances, make the personal interaction positive and if something has to be blamed make it a “system”. Generally easier to accept.

These are just some ideas – I agree with David M that the problem is not trivial.

posted on July 5, 2007

Marie said:

I agree with data mining ang black listing those who were found guilty. However, I think the company should also learn more about the reason why some did it. It could be a mistake on their part.

posted on July 6, 2007

Nancy said:

I agree with Christian. It could be a problem in the part of the company which the customers are just making sure that they get back what they lost. In most cases, those customers don’t even bother complaining.

posted on July 9, 2007