What Is The Difference Between Bespoke And Off-the-Shelf Software?
Most businesses reach this question at a specific moment. A spreadsheet finally breaks, an acquisition doubles the headcount, or a renewal quote lands with a number nobody in finance was expecting.
The honest answer depends less on the software than on two things: how unusual your process really is, and how many people need a seat.
Anyone telling you bespoke always wins is selling something. Salesforce, Microsoft Dynamics 365, Sage and SAP exist because thousands of organisations genuinely do run the same processes as each other.
What packaged products carry is a cost profile that keeps moving. Vertice’s SaaS Inflation Index recorded software price rises of 16.4% in June 2026, its highest reading yet.
Set that against UK consumer price inflation of 2.9% in the twelve months to July 2026, and software is rising at roughly five and a half times the rate of everything else you buy.
What Is Off-the-Shelf Software and What Are You Buying?
Off-the-shelf software is a pre-built product licensed to many customers at once. You are buying the right to use a shared codebase, not the codebase itself, along with the vendor’s data model, release schedule and pricing decisions.
That bundle is worth a lot. A mature product carries thousands of person-years of development, a decade of other customers’ edge cases, and a security team you could never justify hiring.
What the Licence Actually Covers
- Access to the application for a named or concurrent user, at a tier the vendor defines.
- Hosting, infrastructure and a support level that improves as you pay more.
- Feature releases you did not commission and cannot decline.
- An integration surface: an API, webhooks, and usually a rate limit on both.
Prices are public and easy to check. Salesforce UK list pricing runs from £20 to £280 per user per month across its Sales Cloud tiers, with newer AI tiers above £400, while Sage prices its small business accounting per company rather than per seat.
Where Off-the-Shelf SaaS Quietly Limits You
The constraint is rarely the feature list. It is the data model underneath it.
Every packaged product makes assumptions about what an entity is. If your business treats a customer as a vessel, a well site or a clinical cohort rather than a company with contacts, you will spend the implementation bending custom fields around the gap.
Those workarounds become technical debt you pay for at every upgrade. It is the same problem as legacy modernisation, except you are renting the legacy rather than owning it.
What the Sprawl Costs
The bill is rarely one product. Cledara’s 2026 analysis of 1.8m purchases put a 50-person UK company at around $209,000 a year on software, or roughly $4,180 per employee.
The median company runs 25 active subscriptions, and the top quartile runs 49 or more. Vertice’s spend research finds 14% of apps go entirely unused and 51% are underused, which is the number that makes a build case on its own.
What Is Bespoke Software and How Is It Different in Practice?
Bespoke software is built to your specification, so the data model, workflow and integrations follow your process rather than the reverse. In a properly drafted UK contract, you own the intellectual property in the delivered code and can take it to another supplier.
We have set out what bespoke software is in full elsewhere, so this is the short version of how it behaves differently once it is live.
What Changes When You Own the Code
You set the roadmap, so the next release fixes your bottleneck rather than a feature the vendor’s largest customer asked for. You hold the source code, which changes what happens if the relationship ends.
Ownership only exists if the contract says so. Under UK copyright rules work by a contractor stays with the contractor unless it is assigned in writing, which is why IP assignment on payment is the clause to check first.
What You Take On
You also pick up the work a vendor would otherwise absorb: hosting, patching, monitoring, backups and support. None of it is difficult and all of it costs money every year.
Budget it as a maintenance retainer rather than a contingency. The widely used industry rule of thumb is 15% to 20% of build cost a year, and cloud hosting sits on top of that.
The R&D Tax Difference
Bespoke development that seeks a technological advance and resolves genuine technical uncertainty can qualify under HMRC’s merged R&D expenditure credit scheme, which applies to accounting periods beginning on or after 1 April 2024.
The headline credit is 20% above the line, and it is itself taxable, so the net benefit is 15% to 16.2% depending on your corporation tax rate. Configuring a packaged product almost never qualifies, because no technological advance is being sought.
How Do Bespoke and Off-the-Shelf Software Compare Side by Side?
The two models differ most sharply on ownership and exit risk, and least on quality. A well-built bespoke system and a mature packaged product can both score well against a standard such as ISO/IEC 25010.
| Factor | Off-the-shelf software | Bespoke software |
| Cost profile | Low entry cost, recurring per-seat fees that compound with headcount and annual uplifts | High one-off build cost, then a maintenance budget typically 15% to 20% of build cost a year |
| Time to value | Days to a few months. Configuration, data migration and training are the long poles | Three to nine months for a first release, and an MVP-first approach can shorten that considerably |
| Fit | Good for standard processes, poor where your process is the differentiator. Gaps get filled with workarounds | Fit is the point. The risk is over-specifying and building things nobody uses |
| Ownership | You licence access. The vendor owns the code and the roadmap | You own the IP, the code and the backlog, subject to the contract saying so explicitly |
| Support | Tiered vendor support, priced by tier. Predictable but impersonal, and you queue behind larger accounts | Your supplier or in-house team. Faster and more contextual, but you are paying for the capacity |
| Scalability | Scales technically without effort. Scales commercially very badly, because cost is a function of seats | Scales commercially well, since adding users costs infrastructure rather than licences. Technical scaling is yours to engineer |
| Compliance | Vendor evidences its own certifications, and you inherit its data residency and telemetry decisions | You decide residency, retention and logging, and you carry the work of proving it |
| Exit risk | Vendor lock-in: proprietary data formats, migration cost, and exposure to repricing or product retirement | Supplier lock-in unless documentation, tests and handover are contracted. Mitigated by holding the code |
Fit Is the Factor Buyers Underestimate
ISO/IEC 25010:2023 defines nine software product quality characteristics, and functional suitability is the first of them. The 2023 revision renamed usability as interaction capability, replaced portability with flexibility, and added safety as a new characteristic.
Packaged products score well on the ones you can test in a demo: reliability, security, performance efficiency. They score less well on flexibility, which is precisely the characteristic that matters as requirements move.
A useful test during procurement: count the processes your team would have to change to adopt the product. If it is two or three, buy the product, and if it is a dozen and one of them is how you actually win work, think harder.
Exit Risk Cuts Both Ways
Lock-in is the risk most often raised against off-the-shelf SaaS, and it is real. Data sits in a proprietary schema, exports arrive in a shape nothing else reads, and switching costs grow with every year of history.
Bespoke has its own version. If one supplier holds all the knowledge, undocumented and untested, you are just as stuck, which is why documented development practices matter more than the technology choice.
The difference is that you own the asset, so the fix is a handover rather than a migration. A handover takes weeks, and a migration between two vendors’ data models takes months.
What Does Each Option Cost Over Five Years?
Below roughly 50 users, off-the-shelf almost always wins on cost alone. Above about 150 users on a mid-tier licence, a bespoke build usually costs less over five years, because a per-seat licence is a cost that grows with your success and then grows again at every renewal.
Take Microsoft Dynamics 365 Sales Enterprise at its UK list price of £80.70 per user per month excluding VAT, with a 7% annual uplift, which is conservative against recent software inflation.
| Seats | Off-the-shelf: 5-year licences | Plus implementation | 5-year total | Bespoke 5-year total |
| 25 users | £139,200 | £40,000 | £179,200 | £490,000 |
| 60 users | £334,100 | £75,000 | £409,100 | £490,000 |
| 150 users | £835,300 | £150,000 | £985,300 | £510,000 |
| 300 users | £1,670,700 | £250,000 | £1,920,700 | £540,000 |
The Assumptions Behind Those Numbers
- Licence at £80.70 per user per month, with a 7% compound annual uplift.
- A bespoke build of £250,000, plus four years of maintenance at £45,000 a year.
- Hosting at £12,000 a year, scaling modestly with load.
- Implementation cost for the packaged product rising with seat count, as it does in practice.
Your figures will differ, and the shape of the curve will not. One line is flat and the other compounds.
Why Software Costs Rise Faster Than Everything Else
A 7% uplift is a cautious assumption. Recorded SaaS price inflation reached 16.4% in June 2026, so a renewal built on last year’s number is optimistic rather than prudent.
Model the five-year figure with a realistic uplift before you compare it to anything. It is the same discipline as costing a build properly, which how much bespoke software development costs sets out on the other side of the equation.
What the Table Does Not Show
R&D relief. A qualifying build can recover a share of eligible development spend through the merged scheme, which changes the effective build cost.
Internal cost. Bespoke needs product ownership from someone senior in your business, and if you do not budget that time honestly the project drifts.
Risk. Packaged software has a known failure mode, which is poor adoption, and bespoke has a worse one, which is a build that never ships.
A 2026 longitudinal review of IT project outcomes puts 31% of projects successful, 50% challenged and 19% failed outright.
That last point is why the cost comparison alone should not decide it. A £250,000 build abandoned at month seven costs more than any licence.
Where Is the Break-Even Point for Your Seat Count?
Break-even is not a fixed number of seats. It depends on the licence tier, so a £150,000 build breaks even against about 156 seats at £80 per user per month, and against about 45 seats at £280.
Every article on this subject quotes a seat count as though it were a law. It is arithmetic, and the arithmetic moves with the price you are actually paying.
| Licence tier (per user per month) | Break-even seats vs a £150k build | Break-even seats vs a £300k build | Typical products at this tier |
| £20 | About 625 seats | About 1,250 seats | Entry CRM, helpdesk, storage |
| £50 | About 250 seats | About 500 seats | Mid-tier CRM, project tools |
| £80 | About 156 seats | About 313 seats | Enterprise CRM, mid ERP |
| £150 | About 83 seats | About 167 seats | Premium CRM, vertical platforms |
| £280 | About 45 seats | About 90 seats | Top-tier suites, AI tiers |
Figures compare five years of licence fees against build plus maintenance plus hosting, before implementation cost and before any R&D relief. Both of those push the crossover lower.
How to Run the Number Yourself
Multiply your seat count by your monthly licence, then by twelve, then apply your renewal uplift for five years. Add implementation and internal admin time.
Put that total next to a build quote plus five years of maintenance at 18% and hosting. If you want the build side of that sum to be realistic, getting a proper quote takes a scoped conversation rather than a ballpark.
Why the Crossover Moves
- Headcount growth pushes it toward building, because licences scale with people and builds do not.
- Tier upgrades push it toward building, and vendors price the one feature you need into the tier above.
- Falling build costs push it toward building, particularly where AI-assisted delivery shortens the work.
- Heavy configuration pushes it toward building, because you are already paying build-level costs without the ownership.
How Do You Score the Decision for Your Own Business?
Score each of the eight factors below from 1 to 5, where 5 favours building. A total of 32 or more points to a bespoke build, 20 to 31 points to a hybrid, and under 20 means buy the product and move on.
| Factor | Score 1 (buy) | Score 5 (build) |
| Process distinctiveness | Standard for your sector | The process is how you win work |
| Seat count and growth | Under 50 users, stable | Over 150 users, growing |
| Integration demand | One or two mainstream systems | Four or more, including legacy or offline |
| Data and compliance | Vendor already evidences what you need | Unusual residency, retention or minimisation duties |
| Existing customisation | Product used close to out of the box | Upgrades are already a project each time |
| Internal ownership | Nobody has time to own a build | A named senior owner with real authority |
| Time pressure | Live this quarter or nothing | Six to twelve months is acceptable |
| Time horizon | Requirement may not exist in three years | Core to the business for five years or more |
How to Read Your Score
The internal ownership row is a veto rather than a score. If nobody senior will own the build, a low score there should override a high total.
Scores in the middle band are the common case, and they are pointing at a hybrid rather than a compromise. That is the next section.
Discuss Your Project Today
When Should You Choose Off-the-Shelf Software?
Choose off-the-shelf when the process is standard, the seat count is modest, the requirement is regulated in a way that packaged vendors already handle, or you need something working this quarter. In those cases a bespoke build is an expensive way to arrive at the same answer.
The Categories That Are Already Solved
Payroll, general ledger accounting, HR records, helpdesk ticketing and document storage are solved problems with strong products at every size. Rebuilding double-entry bookkeeping is a waste of money, and so is rebuilding Making Tax Digital submissions that your accounting vendor already files for you.
The same applies to commodity infrastructure. Payment processing and checkout are better bought and integrated properly than built from scratch.
The Clearest Signals Packaged Is Right
- Your process matches everyone else’s in your sector, and you are not trying to change that.
- Fewer than about 50 people need a licence and headcount is stable.
- The vendor already meets and evidences your compliance obligations.
- You need to be live in weeks because of a funding round, an audit or a contract start date.
- Nobody internally has time to own a build, and hiring for it is not realistic this year.
Related: What Is Software Development?
When Does Bespoke Software Make More Sense?
Bespoke pays when your process is the competitive advantage, when licence costs scale faster than the value the software delivers, or when the systems you need to connect have no packaged equivalent.
It also pays when the alternative is four products stitched together by manual re-keying. That is a cost nobody puts in the comparison, and it is usually the largest one.
Integration Is the Strongest Case
If your operation depends on pulling company data from the Companies House register, matching it against a legacy pricing engine and pushing the result into a field application used offline, no vendor sells that. They sell three of the four parts and leave you the gap.
That gap is where a custom web application or a field-ready mobile app earns its cost, because the value is in the join rather than in any one system.
Data Control and UK Compliance
Where your UK GDPR obligations are unusual, where you need specific residency guarantees, or where a vendor’s telemetry sits awkwardly against data minimisation, owning the system removes an argument you would otherwise keep having.
This is not a theoretical point in regulated sectors. A vendor that will not tell you where data rests, or which sub-processors touch it, is a procurement problem before it is a technical one.
Scale Economics
If 200 of your 300 licence holders log in twice a month to approve something, you are funding a vendor’s growth rather than your own. Occasional users are the worst value in per-seat pricing.
The same arithmetic drives most sector-specific builds, which is why bespoke CRM and bespoke ERP projects tend to start as licence renewal conversations.
Signals It Is Time to Look at a Build
- Your team maintains spreadsheets that reconcile two systems that should talk to each other.
- You are paying for a licence tier you do not need, to get at one feature you do.
- The vendor has announced a repricing, a forced migration or a product retirement.
- Onboarding a new client takes days of manual setup that a workflow could do in minutes.
- You have customised the packaged product so heavily that upgrades are a project in themselves.
That last one is the point where the comparison stops being theoretical. If you are already paying build-level costs to keep a packaged product bent into shape, you have bespoke effort with none of the ownership.
Tell us what you need and we will scope it with you
Is There a Middle Path Between Building and Buying?
Yes, and it is where most well-run UK technology estates end up. Buy the packaged product for the commodity work, then build only around the part that differentiates you and integrate the two.
Competitors treat this as a footnote. For a mid-market business it is usually the answer.
| Pattern | What you buy | What you build | Fits when |
| Buy plus integration layer | Mainstream CRM, ERP or finance system | The middleware that joins them and the data warehouse | Several products hold pieces of one process |
| Buy plus differentiating module | The commodity platform | The one workflow that is your competitive edge | One process is distinctive and the rest are not |
| Configure and extend | An enterprise platform with a development surface | Extensions inside the vendor’s framework | The platform genuinely fits and you need a few gaps closed |
| Build the front, buy the back | Accounting, payments, identity | The customer-facing application | Your customers experience the product directly |
| Buy now, build later | Whatever ships fastest | Nothing yet, but keep the data exportable | You do not yet know what the process needs |
Buy the Commodity, Build the Differentiator
The rule is simple: never build anything a customer would not notice. Accounting, identity and storage are invisible to your buyers, and the workflow that gets a quote out in an hour instead of three days is not, which is where bespoke development should be pointed.
Engineering-led businesses show this most clearly. Our CAD engineering work with QuickWells built the part no product covered and left the rest alone.
Where Configuration Stops and Development Starts
Configuration is anything the vendor supports through their own interface and will carry through an upgrade. Development is everything else, whoever writes it.
The line matters because heavy configuration breaks at upgrade time and nobody warns you in the sales cycle. Ask the vendor which of your intended changes survive a major release, and check the answer against their platform documentation rather than the sales deck.
Starting Small Deliberately
Where the answer points at a build but the scope is unclear, an MVP first limits what you can lose. A first release in three months tells you more than another quarter of requirements workshops.
The point of a launchpad approach is to make the expensive decision later, with evidence, rather than earlier with a business case.
Related: What Is an MVP in Software Development?
What Should Be in the Contract Before You Commission a Build?
IP assignment on payment, source code and documentation as named deliverables, a defined handover, and rates fixed by role. Without those four, you have bought a service rather than an asset.
No competitor guide on this subject covers procurement, and it is the largest unmanaged risk on the bespoke side of the decision.
The Clauses That Decide Your Exit
- IP assignment on payment, covering employees, contractors and any subcontracted work.
- Source code, tests, infrastructure-as-code and deployment scripts as contractual deliverables.
- A defined handover: documented environments, a runbook, and paid knowledge transfer.
- Rates by role rather than a blended figure, with a fixed term and a rate review date.
- Escrow where the system is business-critical and the supplier is small.
None of this is unusual and all of it gets forgotten. What should be in a software development contract covers the rest, including acceptance criteria and warranty periods.
How to Check the Supplier
Pull the accounts and directors before you sign, and see whether the balance sheet supports the team size they claim. A small supplier is not a problem, and a small supplier holding your only copy of the code is.
Ask for two references on projects of comparable size and ring them. Then ask what happens if the engagement stops next month, because the answer tells you whether you are buying an asset or renting a dependency, which is the same question a digital transformation programme has to answer at every stage.
When Is Bespoke the Wrong Call?
When nobody senior will own it, when the requirement is standard, when the timeline is shorter than the build, or when the real problem is a process nobody has agreed. Software does not settle arguments about how the business should work.
Every article on this topic is written by a company that sells builds, including this one. Here is where we tell clients not to.
The Four Patterns That Waste Money
- No product owner. The build drifts, scope arrives by email, and nobody can approve a decision inside a week.
- Rebuilding a solved category. Payroll, ledger accounting and ticketing have mature products at every price point.
- Automating a process nobody has agreed. Two departments disagree, and the build becomes the negotiation.
- Building for a requirement that may not exist in two years, such as a single client contract or a pilot scheme.
What a Failed Build Actually Costs
The write-off is the smallest part. You also lose the year, the internal credibility of the next technology proposal, and the licence savings the business case promised.
Size the first phase so that failure is survivable. A costed discovery and a first release beat a twelve-month programme, and how much software development costs should be read as a range rather than a number when scope is still moving.
Related Guides
Related: What Is Bespoke Software?
Related: How Much Does Bespoke Software Development Cost?
Related: How Much Does Bespoke CRM Software Cost?
Related: How Much Does Case Management Software Cost?
Related: What Is the Cost of CMMS Software?
Related: What Should Be in a Software Development Contract?
Related: What Is Legacy Software Modernisation?
Frequently Asked Questions
No, because bespoke costs more up front but has no per-seat licence, so total cost of ownership can be lower once seat count and licence tier push five years of fees past the build cost. Below about 50 seats on a mid-tier product, off-the-shelf is almost always cheaper.
ARTICLES









