Acquisition-led fragmentation
Buy-and-build works commercially and fragments technically. Every completed deal adds a stack the group did not design, integrations built under deadline, data structured for someone else's business, and owners who may already have left.
How the gaps form
- Each acquisition is integrated just enough. A ledger feed into group finance, a customer sync into the group platform, an identity link. The rest stays as it was: working, undocumented, unowned.
- Temporary becomes permanent. The spreadsheet emailed monthly to group finance was a Day-1 workaround. Three years later it is the finance feed.
- Knowledge leaves with people. The one engineer who understood the warehouse system retired; the integration still runs; nobody knows what depends on it.
- The estate is documented in fragments. An architecture diagram from the deal, a CMDB nobody updates, an integration platform that sees its own flows and nothing beyond them, a monitoring tool per company.
Why conventional approaches leave visibility gaps
Integration platforms see the flows built on them and not the file transfers, the direct database links, the emailed spreadsheets or the integrations built on the other four platforms the group inherited.
Monitoring tells you a system is up. It does not tell you the feed between two systems is delivering 8% of the expected records, or that a field changed type upstream.
CMDBs and architecture repositories hold what someone typed in, on the day they typed it, with no way to tell what was verified from what was guessed and no reading from the live estate to keep it true.
The multi-year transformation will fix it, eventually. Meanwhile the group keeps acquiring, and the board keeps asking whether the numbers can be trusted.
What is at stake
Regulatory returns built from feeds nobody watches. Month-end close that depends on a legacy system with no owner. Customer data crossing entities without anyone able to say where it goes. A Day-1 that slips because the cutover depended on an integration nobody had mapped. None of these is a technology failure in the usual sense; all of them are failures of visibility and control across the seams between systems.