How to Succeed with a BIG Project – Start Small

The government has ordered an investigation into the issue, so we don't yet know why the EHRs have failed, or why they have been deemed a risk to patients' health. But I've been in enough procurement processes, both private and public, to make some educated guesses.

What usually happens is that whoever it is that needs something makes a list of all the features they can think of. Then some big old colossus shows up and says that they have all of that, and more. So obviously the old colossus is picked.

Then a lengthy process of replacing all the current systems and ad-hoc solutions begins. It ends in tears. Rinse, repeat. Note that this is just as common in the private sector as in the public. We only hear more about the public sector because they are, well... public.

I mean, look at Apple’s decade-long car project, reportedly costing nearly ten billion before being cancelled, or their Apple Vision Pro, a product that completely missed mainstream demand. Speaking about VR, Meta’s Reality Labs has burned tens of billions on its Metavers™ bet with very little but a running gag to show for it. Or why not the Swedish bank SEB’s spectacular pratfall a decade ago.

The list goes on.

For once, there is a rather obvious solution. Instead of trying to do everything, everywhere, all at once, companies, hospitals, schools etc should find a task that relies on two separate systems and replace that workflow with a new one. It could be something as simple as automatically entering purchases into the accounting system when a purchase order is approved, instead of manually copying the same information from procurement software into finance software.

The new system should be:

• Open source
• API-first
• Platform-independent (like the web)
• Built on open standards

That's it. Once that works, they should look for one more task that can be integrated, and open source that. Hopefully, other companies or hospitals, or what have you, will notice and start using the new system. After a while, they will start creating their own open source projects.

Instead of starting one big project, you finish a small one. You put down the first paving stone of a road that eventually will lead to a complete overhaul. And then you keep building. This might sound naive. Like an impossible dream. Surely we can't get the big IT consultants to do this?

To that I say, we have seen what the alternative is, and that's not working. On that note, stop asking for certified developers. I promise you that behind every big fiasco there's an army of certified developers. A certificate isn't a guarantee of quality. You will have to put in some work, and look at previous launches from the company you hire. You need to call those references and ask some tough questions. But it will be worth it in the end.

Let's give this a try. Two systems out. One workflow in. Open source, web-based, API-first.