How many times has a customer told you they would love to buy your product if only it did this tiny little thing that you don’t have on your roadmap? The natural instinct is to make a deal with the customer for a long-term comittment if you deliver the missing functionality. The client gets what he wants, you get a new customer and everyone is happy, right?
The problem with this is that customisation is one of the most disruptive things that can happen to a lean startup with a small team trying to achive great things in record time. Engineers have to stop what they are doing, get their heads around a totally new request and see how they can pull it off.
Even pricing custom work can be tricky as the pricing has to include development and testing time as well as all the complexity of validating the specification and the opportunity cost of having your scarce engineering resources working on something non-strategic. Things can get worse if there are interdependencies between the customer’s infrastrucutre and your system as any changes on the customer’s side leads to more work for your team. Most startups I have worked for dramatically underestimate the cost of these customisations and usually end up loosing money on them.
To my mind, the best approach is to set up a separate team that focusses soley on customisation if this can enable substantial sales. This way you can more easily understand the impacts and costs of customisation while leaving your core technology teams focussed on making your product even more awesome.