I’ve been thinking a lot about point solutions; the other side of the coin is custom software. Which is a trap.
Many SIs I talk to worry that their flexibility-first culture prevents standardization (and repeatability, and scale); they’re right to worry. Flexibility without guardrails is how you end up doing bespoke work over and over. And in a fixed-fee world, this is even more of a wasted opportunity because of the potential huge margins on repeatable work… if you can repeat it.
I know, people reading this are thinking, yeah obviously, Ali… you post about this all the time.
The new part: The same thing is true for software.
Clients often think that their process is so special and bespoke it cannot be changed, so the software has to be customized. In the past, that was the right thing to do. Now, it’s not; instead, finding the right point solution for that specific workflow or tool is the key. These point solutions exist and are multiplying by the day, and there is probably one that does your specific narrow thing really, really well.
You want a tool for this specific, narrow use case; the trap is building custom software to do it. You might think: it's so easy to build software today, we can iterate, we can build exactly what we need.
The obvious downside is the cost to maintain; this has been well-covered.
But in addition, custom software has risks that are parallel to what you see with bespoke services. A point solution has the best practices of hundreds of teams baked in, and it’s always improving with all of their learnings (we hope). Custom software is frozen - it only improves when you intentionally pause and improve it. Over time, your process stays the same, and everyone else’s improves.
Standardization is undervalued: standardization of deliverables means you’re delivering the same best practices to each client, standardization of software means you’re benefiting from all of those best practices yourself.