UK Businesses Are Wasting Up To £8.9bn a Year on Software Nobody's Using
We modelled the UK's subscription creep problem end-to-end, from the size of the SaaS market down to what it costs per employee. The result? £7.6bn a year, gone on unused seats, forgotten trials, and renewals nobody gets round to reviewing.
The Subscription Creep Report
Every business has at least one software subscription nobody quite remembers signing up for. A tool that a former employee set up two years ago, a free trial that turned into a monthly charge, or a licence for someone who left the company last spring. On its own, none of that looks like much. But add it up across an entire business, and it turns into a genuinely large number.
That's what we set out to measure with our Subscription Creep Report. We pulled together the strongest independent research already out there on SaaS waste, from the likes of Zylo, Gartner and Vertice, and combined it with official UK business data from the ONS. The result? Our own modelled estimate of what UK businesses are losing to subscription creep every year, broken down by company size, region and sector, plus a closer look at what's happening inside marcomms specifically, since that's the stack we know best.
We've also built a free calculator, so you can see roughly what subscription creep might be costing your own business.
Subscription Waste Calculator
Curious what subscription creep looks like for a business your size? Pick your company size band below, or drop in your actual monthly software spend if you know it, and we'll estimate how much you could be losing every year. It's important to note that it won't be an exact figure and is just an estimate, but it should be close enough to make you want to go and check.
See what subscription creep could be costing your business, based on our national model.
UK businesses are on track to waste up to £8.9bn this year on software that nobody's using. Our best estimate? £7.6bn.
Nobody signs off on a budget line called 'wasted software spend'. It happens gradually, a tool bought for a project that wrapped up months ago, a discount trial that rolled into a full-price plan, or three different logins for tools that all do roughly the same job. Multiply that across the UK's 1.4 million employing businesses, and you get a number too big to file away as a rounding error.
A few things stood out as we built the model. The waste that's growing fastest isn't the old, familiar kind; it's individually expensed AI tools, up 108% year-on-year in spend and largely invisible to whoever's meant to be keeping track. Bigger businesses lose the most money in absolute terms, but that's not down to carelessness. Even with dedicated procurement teams, £434,949 a year still slips through the gaps at the average large UK business. And by Vertice's numbers, the problem is getting worse, not better, combined waste has climbed from 62% to 65% of all software licences in the space of a year.
We split the £7.6bn three different ways. Firstly, by the size of the business, then by where in the UK it's based, and finally by the industry it operates in. None of these are official statistics, as nobody publishes UK software waste broken down this precisely, so each one comes from applying our national model against real ONS business data. Here's what falls out.
Where the £7.6bn Goes
We estimated average spend across company sizes by mapping total spend against ONS employment data. While this is a modelled assumption rather than raw data, it reflects a well-known reality in procurement, that larger organisations run larger software stacks and inherently rack up more waste.
Large UK businesses (250 or more employees) top the list at £434,949 wasted per business, per year, more than any other size band, even with dedicated IT and finance teams keeping an eye on things.
Medium-sized businesses lose £31,351 each, typically because they've grown fast enough to build a sprawling stack without yet building the process to manage it.
While small businesses lose a comparatively modest £6,333 a year, and micro businesses £1,164, though at that end of the scale, it's less about the size of the number and more about nobody having the time to notice it at all.
London Carries the Largest Modelled Share
Split across the UK's 12 regions in proportion to each region's share of UK businesses, London carries by far the largest modelled share at around £1.5bn a year, more than the South East and East of England combined. That's simply a function of London having the largest concentration of UK businesses, not proof that London businesses are worse at managing their software.
At the other end of the scale, Wales, Northern Ireland and the North East each carry a modelled share under £300m, still a meaningful number for any regional business page, just a smaller slice of a very large pie.
Click a column heading to re-sort the table.
| Region | Share of UK Businesses | Modelled Annual Waste |
|---|---|---|
| London | 19.7% | |
| South East | 14.8% | |
| East of England | 9.9% | |
| North West | 9.7% | |
| South West | 8.6% | |
| West Midlands | 7.9% | |
| Yorkshire & The Humber | 7.0% | |
| East Midlands | 6.6% | |
| Scotland | 6.3% | |
| Wales | 3.8% | |
| Northern Ireland | 2.9% | |
| North East | 2.7% |
Professional Services Top the List
Split across 14 broad industry sectors, again in proportion to each sector's share of UK businesses, our model shows that professional and technical services come out on top at roughly £1.16bn a year, with retail and wholesale (£1.1bn) and construction (£1.07bn) close behind. Sectors like these tend to run bigger, more complex software stacks in the first place, so it's a fair bet they carry a genuinely larger share of the waste than their business count alone suggests.
At the other end, sectors like education and finance likely run tighter procurement than the numbers show, so we'd treat their share as a floor rather than the full picture.
Click a column heading to re-sort the table.
| Sector | Share of UK Businesses | Modelled Annual Waste |
|---|---|---|
| Professional, scientific & technical | 15.3% | |
| Wholesale & retail | 14.5% | |
| Construction | 14.1% | |
| Business administration & support | 8.1% | |
| Arts, entertainment & other services | 7.0% | |
| Information & communication | 6.9% | |
| Accommodation & food services | 6.5% | |
| Production (incl. manufacturing) | 5.4% | |
| Agriculture, forestry & fishing | 5.2% | |
| Property | 4.5% | |
| Health | 4.3% | |
| Transport & storage | 4.2% | |
| Finance & insurance | 2.1% | |
| Education | 1.8% |
The Marketing Stack Has Its Own Problem
As a marketing agency, this hits closest to home. Nobody has published a UK-specific figure for marketing-department software waste, so rather than force one out of numbers that don't quite fit, we've led with the strongest stats we could independently verify, plus one narrow, clearly-labelled estimate for the fastest-growing part of the problem.
The average organisation runs about 103 marketing-related SaaS apps. Only 31% of marketing leaders say their stack is well integrated, and 64% admit they struggle to even track everything they've bought, which probably says more about how subscription creep happens than any other stat in this report. You can't manage what you can't see. Even the tools that are being tracked aren't being used properly; marketers use as little as 42% of their martech stack's capabilities according to Gartner, while The CMO Survey puts the share of purchased tools that are genuinely used at just 56.4%.
The fastest-growing corner of that waste right now is what we're calling 'zombie tools'. This includes individually expensed AI subscriptions, copywriting assistants, image generators, and transcription tools, for example, bought on a personal card for $20 to $40 a month and cancelled by nobody once someone's finished with them.
Our conservative, illustrative estimate puts the cost of these alone at £63.7m a year across UK marketing departments (range: £34.0m–£112.1m), and that's before counting anything else in the stack. AI-native software spend overall is up 108% year-on-year, so if anything, this figure will look small within twelve months.
The "Zombie AI" Floor
UK marketing departments could be wasting at least £63.7 million a year on forgotten AI subscriptions alone, copywriting assistants, image generators, transcription tools, bought on a personal card and rarely cancelled once someone stops using them. That's before counting the rest of the martech stack.
Four Ways to Stop Subscription Creep
The good news? None of this actually requires new software to fix. It just needs someone to actually look at and monitor it regularly. Here's Harry's advice for getting on top of it, and it's refreshingly low-tech:
Want a Second Opinion on Your Own Stack?
If you'd like help working out where your marketing and martech spend is actually going, and what's worth keeping, get in touch.
Get in Touch →Methodology & Sources
We took a top-down estimate of the total UK SaaS and software subscription market, $18.5bn in 2025 (IMARC Group), converted it to sterling at the GBP/USD rate on 6 August 2026 (1.3452), and applied a waste rate drawn from three independent sources: Zylo (36%), Gartner via Productiv (55%) and Vertice (65%). That produces a range of £4.95bn to £8.94bn, with £7.56bn as our published central figure.
We then used ONS Business Population Estimates 2025 to translate that total into per-business and per-employee figures, and to provide a breakdown across micro, small, medium and large UK companies. The size-band split assumes waste is distributed roughly in line with headcount, since no public dataset currently breaks UK SaaS waste down by company size.
The regional and sector splits use the same logic against a different ONS release that publishes business counts by region and industry, allocating our central estimate in proportion to each region's or sector's share of UK businesses. These are the least rigorous parts of the model, genuinely useful for a pitch angle, not a precise regional or sectoral claim, and we've flagged that throughout.
For press & journalists: for interviews with Harry Nisbet, additional cuts of the data, or the full underlying data sheet, please get in touch via our contact page.