The cost of your first lost customer

Here (behind a paywall, I am afraid) is research that proves something innovation theorists have been saying for ages: the cost of losing one of the first customers to something new is much, much higher than the cost of losing one later.

This is a study that examines the adoption dynamics of internet banking. The result is extremely interesting. When you take word of mouth effects into account, the first customer defection is at least 75% more expensive than the last. An early-adopter who defects is worth more than $800. And a laggard (part of the last 16% who ever try something new), around about $210.

The theoretical underpinning of this is well understood. Those first customers are the ones who are typically opinion leaders. They are the ones that influence others to try whatever-it-is. Their views have a disproportionate effect on everyone else.

Presumably, if you believe this, you'd want to make sure your new product introductions are perfect before you'd ever let anyone try them in the wild.

With this in mind, then, I am interested that so many banks have decided to have their innovation functions create public "labs" websites, where they actively seek to bring customers on board with new things early. By letting them try out things they are working on, I assume, such bankers believe they create the impression they are much more innovative.

This is a concept copied from the Web 2.0 community (and, before that, from packaged software), where a select group of customers are allowed to "beta test" things. The idea is that a company can get the last defects out of its products and have an early taste test of consumer opinion.

The fact of the matter, however, is that the economics here are just plain wrong. If you test things with all these self-selecting early customers, the costs of a mistake make the whole thing far too expensive for any reasonable manager to contemplate.

With such economics, the only reasonable basis for having a "labs" site (or an open beta, as some organisations call it) is as a shadow launch vehicle for new things which are completely ready to go to prime time. A recruitment vehicle for early adopters.

With that in mind, I had a look at some of the "labs" sites I know of at banks. A few are playing with Twitter and other social media tools. One has a couple of trial calculators that do some interesting things. There's a money manager. And some interesting mobile banking demonstrations.

Now, these are not products. They are not fully formed new service initiatives. The institutions involved are playing.

Who would ever have thought it would be the innovators with the most expensive toys on the block?

Comments

[Graham Hill]

James

Your readers can view the working paper of the journal article you refer to - What is the Value of a Lost Customer? - at http://tinyurl.com/a54mon for free.

We must not forget that it is an entirely theoretical academic paper, not an empirically tested one. I would be very careful about developing any theory of innovation on such a paper until it has been empirically tested on a large dataset of real live data.

I am also curious about your suggestion that early adopters are also social influencers. I haven't seen anything in social networks research (e.g. Newman, Watts, Barabasi, etc) to suggest that that is necessarily so, rather that they are separate and unrelated behaviours. Even practical social networks analysts like Jeremiah Owyang of Forrester talk about different influence levels of early adopters, with the clear suggestion that the influential early adopters are few and far between.

Involving users in innovation is always a gamble, but I believe the general thinking is that without their active involvement, the probability of the resultng innovation being fit for purpose let alone best in an emerging class is greatly reduced. This is certainly von Hippel's findings in his work on lead-user innovation at MIT. Not involving users is in fact a far bigger gamble, as the 80% failure rate of internally-developed new products on market introduction shows.

Graham Hill
Customer-driven Innovator

[Colin Henderson]

My non validated thought on value of early adopters is this.

They try but fail to encourage the mass to adopt. However they influence others, such as MSM media, bloggers, social media, to spread awareness. This assembly of encouragement eventually presses the growth curve over the infamous "chasm" [in theory].

So yes I agree early adopters as a group are essential for a growth strategy.

[James Gardner]

@Graham: Your comments surprise me. The bass diffusion model on which the paper is based is extremely well validated. As is the underlying data from which the authors derive the parameters they use for the model.

I have no doubt at all about its veracity, and indeed, in my own innovation group, we've done curve fitting to get our own parameters for p and q on our own internet banking data... they're close enough to what the authors found.

I am also surprised by your comments on early adopters. It is well established - by EM Rogers no less - that they are social influencers, and indeed the most important group in a diffusion process. I quote from diffusion of innovation, 5th Edition:

"This adopter category more than any other has the highest degree of opinion leadership in most systems. Potential adopters look to early adopters for advice and information about an innovation. The early adopter category is considered by many to be the "person to check with" before adopting a new idea" p 253.

I agree with involve users in development, by the way. But not in testing. Customer co-creation is powerful, but failing to pay for testing, in my view, is expensive. That, I think, was the point I was trying to make, however unclearly. :-)

J.

[James Gardner]

@Colin: See my previous comment on this subject. But I think we are agreed on this. If you want to get your message out, you can either do spray-and-pray, or at least try to get to the influnencers who will make a difference.

[Nick Marsh]

Hi James, great find, and thanks Graham for providing the link.

I wonder if there is something in the scale of changes presented to users that determines their defection rate? Google is the best known advocate of (what they call) 'granular bucket testing' small changes to interface designs with random selections of users to evaluate changes to the user experience.

Clearly their strategy is very successful, and they are engaging users in testing, unpaid (and unbeknownst!) Of course a freely provided search service is a very different proposition to online banking, which in my personal experience brings up much stronger emotions and concerns around safety and security of my money, and is tied to all sorts of real world experiences. I don't want my bank to 'mess around' too much - who would? Perhaps this explains why there's so little innovation of the UE of online banking? (Is this another post though...)

It's worth pointing out that google also has its 'labs' for their new service offerings and this article - http://www.lukew.com/ff/entry.asp?143 - suggests they mix up their usability work with qual usability testing too. I guess a balanced diet might be the best way forward...