Again and again in the press, even now in the mainstream press, there are reports of hyper-growth (unicorn) companies that suck at customer/partner support. They are so bad at it that their clients start petitions to get any kind of response from the companies. Here is the thing, there is a systematic inherent reason why hyper-growth startups suck at support, and no matter how much support staff hiring they do will fix it.

What are hyper-growth startups?

Hyper-growth startups have the aim to grow exponentially. This means the more they grow the more they grow, the bigger they get, the bigger they get. And if possible also faster and faster. Measured by whatever metric they seem fit. Best limitless, but in a limited world at least until market saturation, then they jump into adjacent neighbouring markets and grow further.

A Hyper-Growth Startup. Blue might be customers, revenue, …

They try to achieve this by different means, a mixture of “going viral”, by customer referring customers, wom, aggressive scale-up sales, throwing VC money into the bottomless depths of Google and Facebook/Meta online ads, Growth-Hacking, Spam, building a market. And with enough money, creativity, and no shame, for a time it might even work.

People have issues. Always.

Customer have issues (red). More customers, more issues.

Customers (sometimes also called partners), humans, create problems. Always. They want things, they do things, things happen. So they need support to fix whatever they want to achieve with you. They need support. The more customers, the more issues. The more issues, the more support they need. The need for support grows with the growth of the company.

Support to the rescue/doom.

So now the companies hire support staff to respond to and solve their clients’ problems. Thing is, hiring people (i.e.:”customer support agents“) only grows – at best – linear.

Yellow is customer support. At best grows linear.

Additional the more people you hire, the bigger the customer support team gets, the less efficient they become (internal complexity, friction, turnover). And don’t forget, that the overall organisation growth puts the organisation as a whole under more and more pressure. At best just burn rate wise.

Your support sucks. If not yet, it will.

So, whatever the hyper-growth companies does. Even if the had the best support in the past. No matter how much money they invest into their support team, no matter how fast they hire, their exponential business growth will always lead to shitty support. And the more focus there is on a huge customer support team, the more strained (not only money-wise) the overall organisation becomes.

Less support.

1) Slow down, stop to grow exponentially.
2) Less support and acceptance that you support sucks.
3) They should just support themselves.

Yeah, about 1) not gonna happen. They sold a hyper-growth company to the VCs, now they need to deliver. So even though 1) “Less overall growth” is the most straight forward solution, it is also the most unlikely to voluntarily happen.

So 2 and 3 are the way to go. At first, stop hiring additional support as long as you are head of the curve.

Make the fact that you have a nearly-not-existing support part of the cost of doing business with you. Make it a known fact. Communicate clearly. Yeah, your lack of support is now a limiting factor of how many and most importantly what kind of customers you get (delayed growth), but instead of being a constant struggle for the users and the organisation, it’s a known constant that the users and the organisation can adapt to, and will, quite quickly.

Shift the resources to self-support

Put the money and internal organisation effort into a lot of clever, highly maintained, un-sucking resources that enable users to support themselves. And most importantly, these resources must get better and better over time. We are talking bottom up forums, top down managed wikis, self-support wizards and tools, and processes to ongoing identify the biggest pain points and solve them one after the other. Not on the support layer, but the root cause. And on the support side, if the users manage to reach it, establish self-serving defaults that solve the easy cases good enough damn quickly.

You support will still suck, in some cases, but in less cases than in the other old school support scenario.

The future

A few short months ago I was at a pretty cool machine learning surf camp and during a late-night drunk discussion with the trainers, we discussed if ML could be used to solve the customer support problem. A learning layer between the customers and the support team, that with every interaction between customers and support makes the support better and better and better. The answer was: “Definitely, let’s make that a startup! It will be a unicorn.”