Off-the-shelf software: when to buy it — and what almost-fitting costs once you've outgrown it.

The stack you already run

You don't have a software problem. You have a seams problem.

A typical firm of ten to twenty people runs an accounts package, a CRM, something for jobs or projects, an e-signing tool and a rota app — plus the spreadsheets that hold the whole thing together. Each tool is good at its job. None of them knows the others exist. So somebody re-keys the invoice details, exports the job list every Friday, and keeps the one spreadsheet only they understand.

That somebody is the real line on the software bill — and no vendor prices it.

Our position, first

For the standard jobs: buy the box. We mean it.

The people selling custom software usually open by telling you you're drowning in wasted subscriptions. The official evidence says otherwise. When the Bank of England and the business department surveyed 2,885 UK SMEs in 2023, just 2% felt they had over-invested — 22% thought they'd invested too little (Bank of England / DBT, published March 2024). And UK firms have already, sensibly, bought the standard tools: among digitally active SME employers, 80% run accountancy software (DBT small business survey 2024, 8,396 SME employers). Off-the-shelf is simply right when:

• The job is standard. Accounts, payroll, email, documents — these work the same in every firm, which is exactly what mass-market software is brilliant at. Nobody should ever build you a bespoke email system.
• The tool matches how you actually work. If your process fits the tool's process without contortions, you're renting decades of product development for pounds a month. Keep it.
• The seams are few. One or two tools that barely need to talk to each other cost you almost nothing in carrying work. The trouble compounds with the tool count, not the tools.

If that's your whole stack — genuinely — stop reading and keep your money. This page is for the firm where the seams have started to show.

The visible line, from published prices

What the stack itself costs — summed from the vendors' own price pages.

Here's a recognisable five-tool stack for a 10–20-person firm, priced from each vendor's own published UK price list, as of 14 July 2026, on the annually-billed basis (monthly billing costs more — the vendors' own badges put the annual-billing saving at up to 18–44%). Prices exclude VAT. The column that matters is the third one: which lines grow when you hire.

Every price is the vendor's own published rate, annually billed, archived in the sources block. Two lines are the same vendor's separate products (monday.com sells its CRM and Work Management separately). Xero's payroll add-on grows per employee (£1.50/person/month beyond the first) and is left out, as are all promotional prices. Prices held flat across the 3-year row — conservative: Xero's own update page shows Grow at £33 in July 2025, £37 today, and £39 announced for 1 September 2026, an 18% rise in about 14 months.

Doubling the team from 10 to 20 multiplies the stack by 1.93× — every extra person adds £49.25 a month (£591 a year) across the four per-seat lines before payroll add-ons. But keep the scale in view: at £6,400–£12,300 a year, the subscriptions alone are not the reason to go custom. The reason is the line the table can't show — the paid hours spent carrying information across the seams. That number doesn't exist in any survey; it exists in your firm, and our calculator counts it from your own numbers, free.

Figure 1. The visible line compounds with headcount — and with the vendors' pricing pages: Xero's Grow plan has gone £33 → £37 → £39 (announced for 1 September 2026) in around 14 months, per Xero's own update page. All prices archived in the sources block.

The fourth option

Replace the seams, not the box.

This is where we differ from every "custom vs off-the-shelf" page written by an agency: we don't want to rebuild your accounts package. Where a tool fits, you keep it — we wire it in, so data is entered once and lands everywhere it's needed. Custom software, on our retainer, is for the part no box covers: the workflow that makes your firm different, currently living in spreadsheets and re-typing. It's built as modules — the worst pain first, live in weeks — for a flat month: £2,500 (Keep), £3,950 (Build) or £6,500 (Partner), published in full. Flat means flat: it isn't per seat, so hiring your eleventh, fifteenth or twentieth person doesn't reprice your software.

The mechanism deserves naming plainly: AI-assisted development. One senior builder — fifteen-plus years of shipped systems across child protection, recruitment, sales and accountancy — uses AI to do the volume work, with every output checked deterministically. That's why bespoke modules can arrive at a subscription-shaped price. And this exact pattern has run in production: the most recent build ran two years on it and replaced five subscriptions with one platform that runs ~80 automated jobs a night — the case study shows it in full.

The comparison to run isn't subscriptions versus retainer — it's subscriptions plus the manual work around them versus the retainer. For most firms we talk to, the manual work alone costs more than the retainer; the working behind that claim is public, and the only figure that prices your firm is the one you count yourself.

Figure 2. The retainer's job is the dashed lines: the paid human effort at the seams. The boxes that fit stay — wired in, so information is entered once. A model diagram, labelled as such; your firm's version of it is what the free process audit produces.

Where this goes next

• Stop re-keying between systems → — the seam mechanism itself, and what module one looks like.
• What your manual work costs — free, no email → — because the fit tax has no benchmark; it has your numbers.
• The pricing, in full → — flat months, published, no per-seat line anywhere.