Rationalise, Don’t Just Consolidate: Why the Crisis Is No Excuse to Collapse a Platform Estate Blindly
Consolidation counts. Rationalisation decides.
The line item that hides the hard part
A board paper crosses the desk this autumn that could have been drafted in any of a dozen organisations. Under the heading of cost reduction, between a recruitment freeze and a ban on non-essential travel, sits a single line: platform consolidation — £8.4m annualised saving, delivered over eighteen months. No one at the table will interrogate it. It has the two properties a number needs to survive an October board meeting this year: it is large, and it points downward. It is approved in the time it takes to turn the page.
I have watched several of these lines approved this season, and the pattern that recurs is not encouraging. The saving is real on the slide and imaginary in the ledger, because the number was built by counting the boxes to be switched off, and no one costed the reason those boxes were switched on in the first place. Consolidation is being sold as an act of tidying. It is nothing of the sort. It is the most invasive surgery an estate can undergo, and the crisis has just handed a scalpel to everyone who ever wanted one.
The danger of a downturn is not that it forces us to consolidate. It is that it lets us consolidate for the wrong reason — cash — and then calls the result rationalisation.
Two words that are not synonyms
We use consolidation and rationalisation as though they named the same act. They do not, and the difference is the whole argument.
Consolidation is a statement about quantity: fourteen general ledgers become three, nine machine rooms become four, a rack of messaging brokers collapses into one. It is measured in boxes removed and licences surrendered. Rationalisation is a statement about fit: it asks which capabilities the business actually needs, which systems earn their keep against that need, and what the estate should look like once the question has been answered honestly. Consolidation counts. Rationalisation decides.
- Consolidation can be planned from a spreadsheet of assets. Rationalisation cannot be planned without understanding what the business does.
- Consolidation delivers its saving on the day the box is decommissioned. Rationalisation delivers its saving over years, as avoided complexity fails to accumulate.
- Consolidation is reversible only at great cost. Rationalisation, done well, removes the reason you would ever want to reverse it.
The crisis rewards the first and starves the second. A saving you can bank inside the financial year is worth more, to a treasury team watching its funding lines, than a structural gain that arrives slowly and cannot be pointed to in a monthly pack. So the estate gets consolidated on a timetable set by the cash position, and the understanding that should have governed it is treated as a luxury the moment does not permit.
Why the sprawl was rational when it happened
Here is the part the cost case leaves out. The duplication we are now so eager to remove was not, for the most part, the product of stupidity. It was the fossil record of decisions that each made sense at the time.
The three overlapping payment engines arrived with three acquisitions, and each was kept because migrating it would have risked a settlement flow no one could afford to break. The second data warehouse was built because the first could not answer a regulatory question fast enough, and by the time the first was fixed the second had its own constituency of users. The middleware zoo grew because every integration project chose the broker its vendor bundled, and no one was ever funded to go back and standardise. Sprawl is what accountability looks like when it is spread across a decade and no single hand ever held the pen for the whole.
“An estate is not untidy because people were careless. It is untidy because every mess was, at the moment it was made, someone’s careful answer to a real problem.”
This matters because the problems have not gone away. Switch off the second warehouse to bank the saving and the regulatory question it was built to answer returns, now with nowhere to live. The cost of the box was visible on the asset register. The cost of the capability it quietly provided was on no register at all — which is exactly why the consolidation case could ignore it, and exactly why the saving will erode the moment that capability is missed.
The objection, taken at its strongest
The best case against everything I have said is simply this: we do not have time for understanding. The funding is being cut now, the run-rate must fall now, and a programme that opens with eighteen months of careful capability mapping will be cancelled before it removes a single server. Better a crude cut that banks something than an elegant analysis that banks nothing. In a year like this one, the argument runs, the perfect rationalisation is the enemy of the affordable consolidation.
I take this seriously, because it is usually right about the constraint and wrong only about the conclusion. The time pressure is real. But the choice it poses — crude speed or careful paralysis — is false. The discipline that separates a rationalisation from a blind consolidation is not, in the main, a matter of time; it is a matter of sequence. It costs very little to ask, before decommissioning a system, three questions: what capability does this provide, who depends on it, and where does that dependency go when this is gone. A team that knows its estate can answer them in days. A team that cannot answer them in days has just discovered the real problem — and should be very frightened of switching anything off.
What the disciplined version looks like
So the practical position is not do not consolidate. The cash pressure is genuine and the estate genuinely is too large. The position is that the crisis is the best cover a rationalisation will ever have, and it is worth the small additional rigour to make the cut structural rather than cosmetic.
- Cut by capability, not by box. Group the estate by the business capability it serves before you group it by cost. Two systems that look redundant on the asset register often serve two different masters; two that look distinct often do the same job twice. Only the capability view tells them apart.
- Cost the dependency, not just the asset. For every candidate for decommissioning, the saving is the licence and the running cost minus the cost of rehoming whatever depended on it. A saving quoted without that subtraction is not conservative — it is fictional.
- Sequence by reversibility. Take the irreversible decisions last. Collapse what you understand first, and buy time on what you do not by virtualising and containing it rather than killing it outright — a consolidated footprint is a saving too, and it does not burn the bridge.
- Name an owner for the whole. The estate sprawled because no one held the pen for all of it. Consolidation run the same way simply produces a smaller, newer sprawl. The one durable output of a crisis rationalisation is a single point of accountability for the shape of the estate — and that survives the recovery.
The fourth point is the one that lasts. When the funding returns, and it will, the pressure to add will return with it, and an organisation that used the downturn only to remove boxes will spend the upturn putting them back. An organisation that used the downturn to settle who decides what the estate is for will have gained something the recovery cannot take away.
The saving that is worth banking
There is a version of platform consolidation that deserves its line on the board paper, and it is not the one that promises £8.4m by counting decommissioned servers. It is the one that can say which of those servers the business would not miss, prove it, and switch them off in an order that leaves the estate not merely smaller but more legible than it was. That version costs a little more thought at the front and delivers a saving that does not quietly reverse itself the following year, when a missed capability has to be rebuilt at a premium.
The crisis will consolidate a great many estates over the next eighteen months. Very few of them will be rationalised. The difference will not show in this year’s cost report, where every consolidation looks like a win. It will show in the recovery — when the organisations that cut with understanding find their estates ready to grow, and the organisations that merely cut find they have paid full price for a saving they have already begun to lose.