The quiet cost of low-code
Three years into a major low-code rollout, the bill arrived. It was not the licence.
The promise
Three years ago a Fortune 500 retailer I was advising went all-in on a low-code platform. The argument was sound: hundreds of internal apps, dozens of departments, and an engineering team that could not scale fast enough. Low-code was the answer.
The reality
It worked. For a while. Apps shipped. Departments built things they could not have built before. The vendor was happy. The CIO was happy.
Then the integrations started.
Then the customisations started.
Then the migrations started.
What the bill actually was
It was not the licence. The licence was reasonable. It was the lock-in tax: every integration was a vendor-specific connector, every customisation a vendor-specific extension, every report a vendor-specific query.
When the company decided to consolidate platforms — for reasons unrelated to low-code — the migration estimate was nine months and four million dollars. Not because the apps were complex. Because the apps were not portable.
What I tell teams considering low-code now
The question is not "will this save us engineering time today?" The question is "what is the exit cost when we want to leave?"
If the exit cost is small, build with the platform.
If the exit cost is large, consider whether you can solve the same problem with a thin layer over standard tooling.
Filed in Enterprise Architecture
Tags: Low-code · Enterprise · Migration