Treat Customers With a Lot of Traffic as Leading Indicators
In a B2B startup, you’ll have a few customers who send a lot more traffic than others.
For example, a startup I know works in the field of incident management: customers create incidents using their platform. Now, some customers send incidents at a high rate, like multiple in a minute. The startup put in rate limits.
But that’s the wrong solution. The right way of looking at such customers is as leading indicators of scalability. As the startup progresses, there will be more and more customers who will have such needs. So when a customer sends high traffic and your system is unable to cope, treat it as an early indication that you’ve hit a scalability bottleneck. And getting an early indication is good, because you can work on it before it becomes a crisis.
Instead of thinking, “One customer should not be allowed to take up disproportionate resources”, think: “A customer can use as much as they want. They just have to pay for it.” For example, I can store 20 terabytes in Google Drive — I just have to pay for it. From a business point of view, it’s also good to have customers who are demanding and willing to pay more.
One startup I know put off scalability for years. They recently got a big prospect who wanted to send high traffic, 10x more than any current customer, and asked if the startup can support it. The founder had to say no. The customer left.
Scalability is not something that can be done on demand like creating a Trello account. It needs time, and you can’t foresee how much it will be. So you can’t play the game of “How long can I put off?” It’s like playing the game of seeing how fast you can drive without getting into an accident.
When you get a customer who wants more, take them on, and don’t put in rate limits 1. Of course, be honest 2 with the customer, and tell them this is more traffic than you’ve handled but you’re eager to work on it 3.
Except as an emergency measure. It’s good to have these knobs on the server-side to use in an outage. Think of them as seatbelts — if your seatbelts are saving you regularly, you’re driving wrong.
I have seen people say yes dishonestly, only to not keep up the promises made after signing the customer on, resulting in disputes, threats of legal action, and a bad word of mouth.
Don’t give them a timeline, even if they pressure you, because you don’t know the amount of work needed, or competing priorities, a priori. Maybe offer a deal where they don’t pay you until you hit an agreed-upon scaling threshold.