Bespoke software vs off the shelf: how to decide

The honest starting position, from a firm that builds bespoke software: most businesses should buy off the shelf for most things. Accounts, payroll, email, payments, CRM. These are solved problems with mature products, and no custom build will beat them on price or on the number of people who have already found the bugs.

Bespoke earns its place in a narrower band, and it is worth being precise about where that band is, because getting this decision wrong is expensive in both directions. Buying the wrong package costs you years of workarounds. Building what you could have bought costs you the build.

The question that actually decides it

Not "is our business unique". Every business feels unique, and it rarely survives contact with a product demo.

The useful question is: is the way we do this thing a competitive advantage, or just a habit?

If the process is how you win work, how you deliver faster than the firm down the road, or how you handle something your sector finds painful, then bending it to fit a package throws away the thing you are actually selling. That is the case for building.

If the process is just how it ended up being done, and a well regarded package does it a slightly different but perfectly sensible way, buy the package and change the habit. It will be cheaper, faster and better supported than anything we could build you.

Where off the shelf clearly wins

• Anything regulated and standardised. Payroll, VAT submissions, statutory accounts. The rules change, and you want somebody else's job to be keeping up with them.
• Anything commodity. Email, storage, video calls, e-signatures.
• Where the product is the market leader and the price is per user. For a small team, a good SaaS subscription is often a rounding error next to a build.
• When you need it next week. A build has a lead time. A subscription does not.
• When you genuinely do not know what you need yet. Buy something cheap, learn what irritates you, then decide. Those irritations become an excellent specification later.

Where off the shelf tends to fail

The failure is usually not the software. It is the gap between what the package assumes and what your business does, and it shows up in recognisable ways.

You are paying for a module you use ten per cent of. Common with industry suites. The one part you needed came bundled with nine parts you did not, and the price reflects all ten.

Your process has been quietly rewritten by the tool. People do things in an odd order because the software insists. Everybody has stopped noticing.

There is a spreadsheet next to the system. This is the clearest signal in existence. If your team maintains a spreadsheet to hold what the package cannot, the package does not fit. That gap is where the errors and the wasted hours live, and it is exactly the territory covered in our guide on outgrowing spreadsheets.

The systems do not talk to each other. Data gets typed into the quoting tool, then the accounts package, then the job sheet. Sometimes the fix here is not a new system at all, it is connecting the ones you have.

You cannot get your data out. Or you can, but only in a shape nobody can use. This one is worth checking before you buy anything, not after.

Support means a ticket queue. When something breaks in a way that stops work today, the difference between a supplier who answers and a form that promises a response within two business days is the whole difference.

The middle option people forget

It is not a straight choice between one package and a full custom platform. Between the two sits the arrangement that suits most SMEs best: keep the packages that work, and build the thing in the middle that nobody sells.

Keep Xero or Sage for the accounts. Keep the payroll. Then build the operational system your business actually runs on, the bit that is genuinely yours, and connect it to the rest so nothing gets typed in twice. That is a far smaller project than replacing everything, and it puts the money where the difference is.

Portals are a common version of this: customers, suppliers or referrers doing for themselves what currently arrives by phone and email. Somewhere House Somerset is an example, structured self referral replacing paper and email forms, with one record per referral. There is more on that pattern on our customer portals page.

A short test you can apply this afternoon

Answer these six honestly.

1. Does a well regarded package already do this, without you changing anything important? If yes, buy it.
2. Is your way of doing it something customers notice or pay for? If yes, that is an argument for building.
3. How many hours a week does the current arrangement cost in admin, re-keying and correcting errors? Cost those hours for a year.
4. How many people need to use it at once, and do they need to see different things? Multi-user with real permissions is where spreadsheets and single licence tools give up.
5. What does it need to connect to, and do those things have a usable API?
6. If the supplier disappeared, what would you be left holding? With bespoke done properly, the answer is the code and the data.

If questions one and two both point at buying, buy. If three and four are painful and one has no good answer, you are in bespoke territory. If it is mixed, you probably want the middle option.

What bespoke costs, so you can compare like for like

Subscriptions look cheaper because the cost arrives monthly and the build arrives at once. Compare them properly: take the per user subscription across the people who need it, over three to five years, and add the hours the workarounds cost you in that period.

Against that, most bespoke projects for SMEs fall between £15k and £60k. We publish the bands, what moves a project up them, and what to ask any supplier before you commit, on the bespoke software development page. If you are comparing quotes from several suppliers, our guide to how much bespoke software costs explains why two quotes for the same brief can differ so wildly, which is almost always about what has been left out rather than the rate.

Two things to insist on whichever way you go: you own the code, and you can export your data in a usable form, whenever you like. Those two conditions turn a supplier relationship into something you can leave, which is the best guarantee that you will not want to.

If you want a second opinion on which side of the line your problem sits, the free process audit is a 60 minute session with no obligation. We will tell you if the answer is to buy something, and if we know a package that fits, we will name it.