English councils spent £2.8bn on temporary accommodation last year - up 25% in twelve months. The councils that bend that curve won't build their way out first. They'll join up the data they already hold.
£5 million a day.
That is roughly what London's boroughs alone now spend keeping homeless households in temporary accommodation. Nationally, the bill has just passed £2.8 billion a year, up around 25% in twelve months, and more than double what it was five years ago.
The part that should keep a Section 151 awake is that almost none of that spend was ever chosen.
No cabinet voted for it. No business case requested it. No transformation programme set it as a target. It is the cost of arriving late; the price a council pays when a family it could have helped in March turns up in crisis in September.
For most councils, temporary accommodation has quietly become the single largest avoidable line in the revenue budget. And the instinct, understandably, is to treat it as a housing-supply problem: not enough homes, rents too high, the market broken.
All true. All largely outside a council's control on the timescale that matters.
But there is a second story underneath the first, and it is one councils can act on this year.
Temporary accommodation is also a data problem.
By the time a household presents as homeless, the council has usually been watching the warning signs for months, across systems that never spoke to each other. A missed council tax payment here. A housing benefit change there. A school flagging a family in difficulty. A debt referral. Each signal sitting in a separate database, owned by a separate team, none of them joined up into a single picture of a household heading for the cliff edge.
The information needed to intervene early almost always exists. It just isn't assembled in time to be useful.
Consider what changes when it is.
Maidstone Borough Council brought its disconnected datasets together into a single early-warning view of residents at risk of losing their homes. In its first year, the council has publicly reported preventing around 100 households from becoming homeless, a 40% fall in the local homelessness rate, and roughly £2.5 million in societal savings. The model identified imminent homelessness accurately in well over 80% of flagged cases.
Read those numbers as a Section 151 would.
That isn't a housing intervention. It's a finance intervention that happens to run through housing. Every household kept out of a nightly-paid room is four-figure weekly spend that never lands on the budget and a family that never has to be uprooted.
This is the heart of what I'd call digital impact, as opposed to digital delivery.
Delivery is "we procured a new case management system." Impact is "we moved homelessness prevention upstream and took avoidable cost out of the temporary accommodation line." One is an output.
The other is an outcome the audit committee can defend.
Now let me be careful and fair here, because this is where the conversation usually goes wrong.
The large platform vendors and systems integrators build genuinely impressive things, and several councils I respect run serious data and AI estates with them to real effect. If your ambition is a large, enterprise-grade analytics platform designed and operated end to end, those firms are excellent at exactly that.
But notice that's a different question from the one a homelessness budget actually asks.
Most councils staring at a runaway temporary accommodation line don't need a bigger build first.
They need to know, quickly, and before committing serious capital, which of the signals they already hold are the ones worth joining up, pre-assessed against their own data and their own risk appetite, so the first intervention lands in months rather than years.
Different shape, different product, different moment in the journey.
One answers "build me the platform." The other answers "tell me, before I spend a penny, which three data sources will move this line the most." Both are valuable.
They are not the same thing and getting the order wrong is how councils spend two years and a capital programme to arrive where prevention could have started this autumn.
The councils that will look smart in eighteen months are doing three unglamorous things now.
First, they treat the temporary accommodation budget as a prevention scoreboard, not just a housing cost. They ask what upstream signal each placement represents.
Second, they join up the data they already own before buying anything new. The cheapest predictive model in local government is the one built from records you already hold.
Third, they instrument for outcomes from day one. Not "we built a dashboard," but "we moved X households out of the crisis pipeline and released £Y back to the General Fund."
So here's a ninety-second self-audit for any leadership team reading this.
Could you say, today, how much of last year's temporary accommodation spend was on households your own systems had already flagged as at-risk months earlier? And which three datasets, joined up, would give you the earliest warning?
If the honest answer is "we don't currently know," that isn't a failure.
It's the most valuable gap on your books because it's also the cheapest and fastest one to close.
If you're a Section 151 or a Chief Executive reading this, the front half of this conversation sits with your Director of Housing and your Head of Digital, and the back half, whether the saving is real and bankable, sits with you. Tag them in the comments and I'll share the prevention-mapping question set we use in council workshops to trace a temporary accommodation line back to the upstream signals that drive it.
Genuinely curious, and there's no wrong answer: in your council, what's the one dataset you already hold that would give you the earliest warning a household is heading for crisis and what's stopping it from being joined up today?
────────────────────────────────────────
#LocalGovernment #DigitalImpact #Homelessness #TemporaryAccommodation #PreventionFirst
