Jacques Mattheij

Technology, Coding and Business

What do you want, a cookie?

It started a couple of years ago, became fashionable last year and it’s become just about unavoidable now: the plague of the ‘customer satisfaction survey’. I can’t order a ball-point pen online without being mailed a link to a questionnaire to give feedback. Either from the company directly or - worse - from some third party they outsourced the survey process to. These outsourcing companies tend to outsource everything else as well, which is one of the reasons they need the survey in the first place: it’s the only way they’re going to find out what happened between you and their various subcontractors, who can’t be trusted to accurately report their dealings and so the company relies on you, the customer to ‘close the loop’ and do their QA work for them.

What did I think of the process? Am I satisfied with the product? The company? The delivery?

And if these were optional it would be one thing but if you fail to respond be prepared to be hounded by follow ups. What’s wrong with simply doing your job and maybe dropping a little card in the box with the product with a url or a toll free number on it I can hit up in case I’m not satisfied. Satisfaction should be the default, 95%+ of the deliveries and the products sold through online stores should work without a hitch. And in those cases where there is an issue with the delivery or the product you can bet your bottom dollar that the customers will complain.

So effectively the stores that do these ‘customer satisfaction surveys’ are asking for a pat on the head, a virtual cookie if you wish, for just doing their job. No doubt the results of these meaningless surveys will be used internally to prove to higher ups how great a job they’re doing. But you don’t measure customer satisfaction by stalking your customers relentlessly until they say ‘yes, ok, I’m satisfied, now please leave me alone’. This goes double for those cases where the customer satisfaction survey arrives before the actual product does.

You measure customer satisfaction in one way only: By measuring repeat sales to the same customer (this is harder for companies that sell pricey products that one normally does not buy on a regular basis, in which case surveys could be done but only to a statistically significant sample of the customers, not blindly to everybody since the results won’t be any different). And if those repeat sales don’t happen then you have many tools at your disposal to figure out where things go wrong, the most obvious ones are to try the order process yourself and to test the products you sell before you stock them. You know, like those brick and mortar stores used to do, to build up knowledge about your inventory. Another pretty direct way to measure customer satisfaction is to measure the rate at which product is returned, because it has become almost standard that products can be returned ‘no questions asked’ within a certain amount of time either because the stores compete on this metric or because (such as in many European countries) it is written into law. Who wants to deal with the support department anyway if a product return is just as easy or even easier?

Really, the things that are pretty much guaranteed turn-offs for me with respect to e-commerce are being required to make an account for a purchase (I don’t need an account with you, I just want to buy your product) and to be hounded with surveys post purchase. Roger Binns (whose feedback I graciously incorporated in version 2.0 of this post) relates an interesting story where the online survey for a camera mysteriously refused to register bad scores. That’s one way to improve on your metrics I guess. What’s worse than a survey? A survey that doesn’t work once you decide to tell them company what it should fix to retain your patronage.

From now on I’ll send companies that engage in the survey practice a cookie, accompanied by a small survey about the cookie and the delivery…