Kim Warren on Strategy
Strategy insights and living business models
You can defy gravity - for a while.
The case of Klarna - the buy-now, pay-later provider - showed how to mess up an AI implementation. They assumed that an AI chatbot would eliminate the need for customer service agents, because (more here).
But there's a much more general insight from this case.
Mistake now, pay later
That insight is not just that we need to match service capacity with customer-driven need - but that if we don't do that
- the price we will pay will show up some time later
- ... and that price could be huge
- ... and recovery could take a long, long time
A story ... we have a large, growing customer-base, driving rising revenue. We start with enough support staff to handle their enquiries, but fail to increase the team fast enough.
Somehow we defy gravity, and continue to grow revenue for more than 3 months after the time when we could have avoided the later down-turn.
Figure 1: How inadequate service damages future revenues.

What's the story?
- the rising need for customer-support passes our capacity in late July, and service quality starts falling
- ... but it takes a while before customers become annoyed enough for customer-churn to start rising from early August
- ... but potential new customers only learn about this some weeks later, when our reputation suffers and the new-customer win-rate starts falling from the start of September
- ... but that customer win-rate is still higher than the churn of existing customers, so revenue continues to grow, though more slowly
- ... until customer losses exceed the win-rate and total customer numbers start to fall, taking revenue down with it - in October!
But when did we actually make the error? It takes some weeks to hire and train new support staff, so to avoid the demand/capacity mismatch, we actually needed to hire faster from June!
Why is this happening?
Having failed to hire enough support staff, there comes a point some weeks later when those staff cannot handle well the incoming enquiries from customers.
Service quality starts falling - calls go unanswered, customer problems do not get fixed ... Problem is, customers put up with this, maybe for quite some time. But their annoyance grows, until - some time later - some get so annoyed that they leave.
Meanwhile, prospective customers know nothing of this. They are still being attracted by our great product and value. So we keep winning still more customers whose service demand we cannot deal with. ... so service quality gets still worse - still more customers get annoyed and leave.
Eventually, word of our bad customer service gets out - our reputation is wrecked. And now we are both losing customers and failing to win new ones.
Only then do we see the revenue-hit caused by our long-ago mistake.
Figure 2: Working model of service quality, customer annoyance, reputation and revenue.

Get this model
This model is included in the set of marketing models available here and can be adapted to your situation.
And yes, every element in this model is measurable in your real world.