Whatever You Can Buy, Your Competitors Can Buy

Everyone in your industry, including the three firms you lost a tender to last year.

This is not a criticism of any vendor. It is how the arrangement has to work. A software business is only viable if the same product sells many times over, and the cost of building it is spread across every customer who buys it. That is what makes good standard software affordable, and it is why this series keeps telling people to buy the ordinary parts.

But it follows directly that no capability arriving in a box can ever be the reason a customer picks you. It arrived in everyone’s box. At best it stops you being worse than the competition, which is worth having and is not the same as being chosen.

So the useful version is: buy it wherever you can, and somebody should tell you plainly when that is the right answer. But nothing you can buy will ever be why somebody picks you.

It comes back as a feature everybody has, and the process by which that happens is not sinister. It is just good product management.

A vendor improves a product by noticing what customers ask for, generalising it, and shipping it to everybody. That is the correct thing for them to do and it is why the products get better. It is also fatal to anything that was supposed to distinguish you.

So the sequence runs: you explain the clever thing your business does, they build a version of it because you asked, and eighteen months later it is a checkbox in the release notes that your competitors have too. You paid for the development and then watched it become an industry standard.

Put plainly to somebody who has to decide: they will build what you asked for. Then they will ship it to everybody in your industry, including the people you are bidding against.

That is where this conversation usually goes, and it is the less interesting half of it.

The question about whether a model processed some data is answerable, contractual, and mostly a procurement matter. It is worth getting right and it is not where the exposure actually sits.

The exposure is the same one as the section above, arriving through a newer door. When a business lives inside somebody else’s product, the way it works becomes visible to that product, and what is visible to a product tends to become a feature of it. The route from your way of working to your competitor’s capability does not require anybody to behave badly.

So the sharper question is not whether a model saw it. It is whose product your way of working ends up inside.

The part that would be a disaster to hand over. Usually much less than people expect, and much more specific.

It is worth being concrete about how small this can be. A mortgage origination business had a landing page that was already doing its job and needed exactly one thing that could not be bought, so that one thing got built and nothing else did. The words on the pages went into a content system they edit themselves; the calculations stayed in code where they can be tested.

At the other end, a portfolio platform for advocates has a tier built from nothing for one person, and it is deliberately not turned into a template afterwards. It is not built so that it could be. That is the same principle enforced as a product decision, and it is what lets that tier be sold honestly.

Neither business built more than it had to. Both built the thing nobody could sell them.

Because of what happened the last time somebody had software built for them, and that objection deserves a chapter of its own rather than a paragraph.

It is the next one.