August 2026 - Page 23 of 44

In-house developers, a software agency and freelancers are not three prices for the same thing — they are three different products, sold in three different units, carrying three different kinds of risk. This guide normalises all three to a comparable annual cost for the 2026 UK market, sets out the on-costs that make a salary roughly 1.5 times its headline figure, and compares the routes on speed to first release, delivery risk, control, knowledge retention and intellectual property. It includes a weighted scoring method you can run in twenty minutes, the hybrid core-plus-capacity model most UK companies end up with, the IR35 and copyright traps that catch contractor engagements, and the hiring mistakes that make the in-house route the least reversible of the three.

Software development company pitches converge fast: by the third meeting everyone is agile, everyone has a dedicated senior team, and every quote sits within twenty per cent of the others. The differences that decide whether your project ships only surface when you ask specific, verifiable questions. This guide gives you thirty of them, grouped into the seven areas where engagements actually break down — team and continuity, delivery process and evidence, price and change control, code ownership and exit rights, security and data protection, support after launch, and reference checking — with UK day-rate bands by supplier type, a scoring method that weights the things which predict outcomes, the contract clauses worth holding out for, and the red flags that should end a shortlist.

Software projects rarely fail in a single moment. They drift — a sprint slips, integration is deferred, a status report stays amber for eleven weeks — until the gap between reported progress and real progress can no longer be closed inside the original budget. Most of those projects are still recoverable, and for far less than the cost of starting again. This guide covers the practical rescue: the four failure signatures that separate a struggling project from a genuinely failing one, a ten-day audit that produces defensible numbers instead of percentage-complete fiction, the five root cause families and how to tell which two you have, the four rescue options with UK cost bands, the governance changes that stop a rescued project relapsing, and the three signals that mean you should stop and salvage instead.