Software Project Cost Calculator

Estimate what a custom software project costs in the UK — with the arithmetic shown, so you can check it rather than trust it.

No email required. The result appears immediately.

How this calculator works

Most “cost calculators” are lead capture with a number attached. The number comes from nowhere and exists to start a conversation.

This one shows its working.

The basis

Every honest UK software quote is built on a day rate. ITJobsWatch reports a median advertised day rate of £525 for UK software developers in the six months to 1 October 2026, with the 10th percentile at £366 and the 90th at £606. Software Engineer roles sit higher at £550. London carries a premium of 15–30%.

Agency pricing builds on that: contractor rate, plus overhead, plus margin, plus risk. A typical UK agency day rate lands £450–£900 depending on size and specialism.

The calculation

Estimated days  ×  Day rate  =  Range

Estimated days come from the scope you describe. The calculator adjusts for the four factors that genuinely move a software price:

How to sanity-check any quote you receive

This works on anyone’s quote, including ours.

Quote total  ÷  weeks quoted  ÷  5  =  implied day rate

Then ask:

Under £400/day? Ask who is actually writing the code and where they are based. That may be entirely fine — offshore development is a legitimate choice — but you should know.

£450–£900/day? A normal UK agency range.

Over £900/day? Ask what you are paying for beyond engineering. Specialist regulatory experience or genuine enterprise scale can justify it. A London postcode on its own does not.

Divide again by the number of people assumed and the picture gets clearer still. A £60,000 quote over 10 weeks implies £1,200/day — which means either two people or a very expensive one.

Typical projects and what they actually cost

Real shapes rather than abstract bands, so you can find the one nearest yours.

Internal tool replacing a critical spreadsheet — one user type, one or two integrations, moderate data migration. Job tracking, stock, scheduling, approvals. £12,000 – £28,000, 5–9 weeks.

Customer portal over an existing system — logged-in customers seeing order status, documents, invoices, stock. Usually one deep integration with an ERP or accounting package. £18,000 – £45,000, 8–14 weeks.

Operational system with multiple roles — office, warehouse and management each with their own screens and permissions, several integrations, reporting. £35,000 – £80,000, 14–24 weeks.

Operational layer over existing accounting — keep Xero or Sage, build stock, jobs and costing above it, integrate properly. The most common thing we ship. £25,000 – £60,000, 8–16 weeks.

Full custom ERP — replacing the systems the business runs on. £60,000 – £150,000+, 20–36 weeks.

Agentic AI feasibility then one production agent — measured accuracy on real data, then one process live with audit and escalation. £24,000 – £57,000 combined, 8–11 weeks.

Data migration is quoted separately in every one of these, after a real extract has been inspected. Bundling it means guessing, and guessing wrong on migration is a well-trodden route to a 50% overrun.

What this calculator cannot tell you

Being straight about the limits.

It cannot see your data. The difference between clean and messy records is weeks of work, and nobody knows which you have until someone looks at an extract.

It cannot test your integrations. “They have an API” and “we have authenticated and pulled real records” are different statements, and the gap between them is where 54.9% of over-budget projects originate.

It cannot tell you whether to build at all. Often the right answer is an off-the-shelf product, a process change, or three integrations rather than a full system.

It produces a range, not a quote. Anyone giving you a firm fixed price before discovery has either padded it or is planning to recover the difference through change requests.

What happens after the estimate

If the number is roughly what you expected: the next step is discovery — one to two weeks, £3,000–£8,000 — producing a requirements document, a technical approach and a costed plan. That output is yours, including the right to take it to another supplier.

If the number is higher than your budget: useful information, cheaply obtained. Options are narrowing scope to a first phase, considering a product, or building around what you already run instead of replacing it.

If the number is lower than you expected: check the scope you entered. People frequently under-describe integrations, which is where the cost hides.

How to use the number

Treat it as an order of magnitude, not a quote. Its job is to tell you whether you are looking at a £15,000 problem or a £150,000 one, so you know which conversations are worth having.

Use it to sanity-check what comes back. Brief three firms identically, then compare each quote against this range. A quote at half the estimate usually reflects a different understanding of the scope — ask which parts they have assumed out. A quote at triple usually includes something you did not ask for.

Use it to decide what to cut. If the number is uncomfortable, the fastest reduction is almost always fewer integrations in phase one, followed by fewer user types. Features are rarely where the cost is.

Do not use it as a negotiating position. “A calculator said £30,000” is not an argument against a quote built after someone looked at your data. The calculator has not seen your systems; a firm that has done discovery has.

Related tools and reading

• ERP readiness check — twelve questions predicting whether an ERP project will overrun, scored against the Panorama failure data
• Our published pricing — full ranges, five-year comparison against licence fees, what is included
• How to write a spec — a vague brief is expensive before anyone writes code
• Fixed price vs time and materials — which model actually protects you
• How to choose a firm — seven stages, nine questions, contract checklist

Our published prices

We do not hide these behind a call. Full ranges here.

Minimum build engagement: £10,000. Below that, an off-the-shelf product will serve you better and we will usually name it.

What actually drives a software price

Four factors move a quote more than the feature list. Understanding them tells you where to push when a number comes back too high.

Integrations — the largest variable

Connecting to a modern, documented API is routine work. Connecting to a fifteen-year-old system with no API, undocumented behaviour and a vendor who charges for access can take longer than the rest of the build combined.

The Panorama 2026 ERP study found that among projects exceeding budget, 54.9% needed technology that was only discovered after the project began. That is overwhelmingly integrations.

What to do: insist any supplier actually connects to your systems during discovery — authenticates, pulls real records, confirms the fields exist. “They have an API” and “we have read data from it” are different statements.

Data quality — the most underestimated

Migrating 40,000 clean records is a day. Migrating 40,000 records where the same customer appears four times under three spellings, dates sit in two formats and a third of phone numbers are in a notes field is weeks.

What to do: export a sample and look at it before anyone quotes. Count duplicates. Check null rates on fields that matter. This single exercise turns a guess into an estimate.

Number of user types

Each distinct role means its own screens, its own permissions and its own testing. One user type is a system; four is closer to four systems sharing a database.

What to do: ask whether any roles can share an interface in phase one. Often two of them can.

How many people must agree

Not technical, and it genuinely affects cost. A project with one decision-maker moves at the speed of the build. One needing four departments to agree moves at the speed of the slowest calendar, and elapsed time is money on any engagement.

What to do: name one person with authority before the project starts.

If the number is higher than your budget

Three routes, in order of how often they work.

1. Narrow phase one. The highest-pain area only, live and in use, before anything else. You get value in weeks instead of months, and phase two is funded by the saving from phase one.

2. Keep what you already run. Most businesses convinced they need a complete system need their existing accounting package, two or three integrations, and one custom module. That is typically a quarter of the cost.

3. Check whether a product fits. Genuinely check — trial two with your real data for a fortnight. Here is the comparison with five-year cost models.

What does not work: finding the cheapest quote and hoping. A quote substantially below the others usually reflects a different understanding of the scope, and you meet that difference in month three.

Why no competitor publishes one of these

Two reasons, and both are commercial rather than technical.

The honest answer is a range, and ranges look weak in a sales process. A number with a £40,000 spread invites the question “why?”, and answering it properly means admitting how much depends on things nobody has looked at yet.

It removes pricing flexibility. A firm that has not published anything can anchor the number to what it believes you can pay. That is legal, common, and the reason most of this industry asks about your budget before telling you their rates.

We publish because the enquiries it produces are better. People who cannot afford us do not call, people who can arrive already knowing roughly what they are buying, and the conversation starts at the problem rather than at the price.

The arithmetic behind every UK quote

Any quote you receive, from anyone, decomposes like this:

Engineering days  ×  (contractor rate + overhead + margin + risk premium)

Contractor rate — ITJobsWatch median £525/day for UK software developers, six months to 1 October 2026. London 15–30% higher.

Overhead — employment costs, software licences, insurance, premises, non-billable time. Typically adds 30–60% to the raw rate.

Margin — the firm’s profit. 15–30% is normal and it is not a scandal; a firm making no margin does not exist in three years when you need a change.

Risk premium — only on fixed price, and only where scope is uncertain. 20–40% before discovery, far less after. This is why discovery makes a fixed price meaningful.

What this tells you: a quote implying £300/day from a UK supplier has either squeezed margin to nothing, is using juniors, or is not UK-based. All three can be acceptable — if you know which one you are buying.

Frequently asked questions

Built and maintained by Jas Singh, Founder. Rates last reviewed 8 October 2026.

Day rate data from ITJobsWatch UK contract listings for the six months to 1 October 2026. Advertised rates reflect what is offered, not necessarily what is agreed.