Comparison
Bespoke or off-the-shelf software: should you build or buy?
By Rare Agentics engineering teamLast updated First published 8 minute read
Buy off-the-shelf software when it fits how you work and the process is not what sets you apart: accounting, payroll, email and most commodity tools. Build bespoke when your process is your advantage, when the workarounds around a product already cost more than a build, or when per-seat pricing and lock-in grow faster than the business. Most growing businesses should do both: buy the core, and build the edge that makes them different.
Why is build or buy the wrong first question?
Build or buy sounds like one decision. In practice it is a dozen small ones, one for each part of the business that software touches.
Very few businesses should build everything, and very few should buy everything. The useful question is narrower: for this particular job, does a product exist that fits how we work, and will it keep fitting as we grow?
We build bespoke software for a living, so read this with that in mind. We have tried to be honest about when buying is the better answer, because for a lot of what a business does, it is.
How do bespoke and off-the-shelf software compare?
| Off-the-shelf software | Bespoke software | |
|---|---|---|
| Time to start | Days: sign up and configure | Weeks: a first product typically takes 1 to 8 weeks |
| Upfront cost | Low, often a free trial | A fixed build price, agreed before work starts |
| Ongoing cost | Subscription, usually per seat, rising with headcount and tier | Hosting, AI model usage and support, with no per-seat pricing |
| Fit to your process | You adapt to the product | The software adapts to you |
| Who decides the roadmap | The vendor, for all its customers at once | You |
| Ownership of the data | In the vendor's system, on the vendor's terms | In your own cloud accounts |
| Leaving | Export what the vendor allows, and retrain everyone | Nothing to hand back, if it was built in your accounts |
| Best for | Commodity jobs that work the same in every business | The process that makes you different, and the joins between products |
| Main risk | Lock-in, workarounds and price rises | Building something a product already does well, and owning its upkeep |
Neither column wins outright. The table is a list of trade-offs, and the right answer changes from one part of the business to the next.
When should you buy off-the-shelf software?
Buy when the job is the same in your business as in everyone else's, and a good product already does it.
Accounting, payroll, email, calendars, file storage, video calls and the basics of a sales pipeline are the obvious examples. These products are built by large teams, tested by thousands of customers and kept up to date with tax rules and security patches. You would spend a lot to build a worse version.
Buy, too, when you are still finding out what the process should be. If the way you sell or deliver changes every month, a configurable product lets you learn cheaply. Bespoke software is best built around a process you understand well enough to write down.
And buy when the product fits well enough that the gaps are cosmetic. A field with the wrong name or a report laid out differently is not a reason to build. Workarounds that eat hours every week are.
When does bespoke software earn its cost?
There are four situations where we see bespoke software pay back clearly.
Your process is your advantage. If the way you match supply to demand, price a job, manage rights or onboard a client is what customers choose you for, a generic tool will bend it towards the average. Software built around your process protects the thing that makes you different.
The workarounds already cost more than a build. Count the spreadsheets that sit beside the product, the copy and paste between systems, and the person whose job is mostly reconciling them. That is a monthly cost, and it grows with the business.
The products do not talk to each other. Often the gap is not inside any one product but between them. A thin bespoke layer that joins your CRM, your accounting system and your delivery tools onto one set of records can be worth more than replacing any of them.
Per-seat pricing outgrows the value. A product priced per user is cheap for ten people and expensive for a hundred, especially when most of those users only need to read or approve. Bespoke software has no per-seat pricing, so growth raises hosting and usage costs, not a licence fee for every head.
What does buying the core and building the edge look like?
For most growing businesses the answer is a mix, and it follows a simple rule: buy the core, build the edge.
The core is the work that is the same everywhere. Keep it in well-known products: bookkeeping in your accounting package, email in your email provider, payments with a payment provider.
The edge is the work that is specific to you: the workflow your team runs every day, the client portal your customers use, the agent that reads supplier invoices and drafts the payment run, the dashboard that joins finance and delivery. Build that, and connect it to the core.
An operating platform is this pattern applied to the whole back office, delivered as a platform build in phases. Clients, pipeline, finance and documents sit on one shared set of records, while bookkeeping stays in your accounting system and the platform reads from and writes to it so finance and delivery agree.
Use this decision guide for each job in turn:
- Does a well-known product do this job, the way we do it, today? If yes, buy it.
- Is this job part of why customers choose us? If no, buy the closest fit and live with the gaps.
- Are the workarounds around a product costing hours every week? If yes, consider building the missing piece rather than replacing the product.
- Will per-seat pricing or the vendor's roadmap constrain us as we grow? If yes, price a bespoke version over three years.
- Do we understand the process well enough to write it down? If no, buy something configurable now and build later.
How do you compare the cost over three years?
Comparing a monthly subscription with a one-off build price is comparing two different shapes of cost. Put both over the same period (three years is a sensible horizon) and list every category, not just the headline price.
| Cost category | Off-the-shelf | Bespoke |
|---|---|---|
| Setting up | Configuration, data import, training | The build, at a fixed price |
| Licences | Per seat or per tier, every month, usually rising | None |
| Hosting | Included in the subscription | Billed by the provider to your own accounts |
| Workarounds | Spreadsheets, manual re-keying, extra tools | Designed out, where the build covers them |
| Integrations | Paid connectors or add-on tiers | Built in, to the systems you choose |
| Changes | Wait for the vendor, or pay for a higher tier | Priced in writing when you need them |
| Support | Included, at the vendor's response times | A monthly plan quoted to scope, or your own engineers |
| Leaving | Export limits, migration and retraining | Nothing to hand back |
Two lines are easy to miss. Workarounds are a real cost even though no invoice arrives for them, so count the hours. And the cost of leaving matters, because a product that is hard to leave can raise its price.
How do lock-in and per-seat pricing creep in?
Lock-in rarely arrives in one decision. It accumulates.
Your data ends up in a format only the vendor reads. Your team learns the vendor's way of working. Integrations are built against the vendor's API. Each step is reasonable, and together they make the product expensive to leave, which gives the vendor pricing power.
Per-seat pricing works the same way. The first ten seats are an easy yes. By the time the whole company needs access, including people who only approve or read, the subscription can be one of the larger lines in the budget.
None of this means you should avoid products. It means you should check, before you commit, what you can export, in what format, and what the price will be at twice your current headcount. We cover the ownership side in more detail in how to own your code and data with a software partner.
What should you ask a vendor or a builder before you commit?
The same handful of questions separates good options from bad ones on both sides.
Ask a software vendor:
- What will this cost at twice our current headcount, and at the tier we will need in two years?
- What can we export, in what format, and is there a charge for it?
- Which of the features we need are on the current roadmap, and which are only promised?
- Where is our data stored, and who at the vendor can see it?
Ask a builder:
- Will the code and data live in our accounts from the first commit?
- Is the price fixed in writing, and what happens when the scope changes?
- What do we get besides the software: tests, documentation and a handover?
- What happens after launch, and who fixes a defect?
- Have you told us which parts we should buy instead?
The last question is the one that tells you most. A builder who never recommends a product is selling builds, not advice.
What should you check before you sign either?
Run this alongside the decision guide above, before you sign a subscription or a build contract.
- Have we written down, in plain language, how this process works today?
- Have we trialled the two closest products with the people who will use them?
- Have we counted the hours spent on workarounds each week?
- Have we priced both options over three years, with every category in the table above?
If a no-code tool is on your shortlist too, read no-code or custom build, which covers when to move on from one.
How Rare Agentics can help
We will tell you when a product is the better answer. When it is not, our bespoke software builds are priced in writing before work starts, with working software every week and the code in your own accounts. Your code lives in your repositories and your data in your cloud accounts from the first commit. There is nothing to hand back, because it was never ours, and the system stays with you if we ever part.
If you want to test the idea on one piece first, a proof sprint is from £3,000 and delivers one bounded feature or workflow in 1 to 2 weeks. Tell us what you are weighing up, and within two working days you get a first view.