A spreadsheet, a subscription stack and a handful of disconnected manual processes can look inexpensive until they begin slowing down sales, service delivery or decision-making. At that point, the question is not simply whether to build software. It is how much does custom software cost, and what commercial return should it be expected to create?
There is no credible single figure. A tailored product can range from £40,000 for a focused MVP to £250,000 or more for a substantial SaaS platform or business-critical operational system. The useful question is what needs to be solved, what must be true at launch, and what can sensibly be delivered in phases.
For UK businesses, a well-scoped custom software investment should buy more than code. It should create a clear route from a business problem to an adopted, maintainable product that improves revenue, efficiency or customer experience.
How much does custom software cost in practice?
Most custom software projects sit within a few broad investment bands. These are indicative UK ranges, typically excluding VAT, and assume an experienced product and engineering team rather than individual freelance capacity.
Focused discovery or prototype: £8,000 to £25,000
Discovery reduces uncertainty before significant build spend begins. It may include stakeholder workshops, user journeys, technical architecture, data mapping, product requirements and a prioritised delivery plan. A clickable prototype can also be useful where a founder needs to test a proposition with customers, investors or internal teams.
This phase is often the best use of budget when the opportunity is clear but the right product shape is not. Skipping it can appear faster, but it usually moves expensive decisions into development, where changes take longer and cost more.
MVP or targeted internal tool: £40,000 to £100,000
An MVP is suitable when a business needs to validate a new service, replace a painful manual workflow or prove that customers will use a digital product. It should solve one valuable problem well, rather than imitate every feature of a mature competitor.
A field-service portal, customer onboarding tool, bespoke quoting system or lightweight marketplace may fit this range. The lower end assumes limited integrations and a tightly defined user journey. Costs rise quickly if the product requires several user roles, complex permissions, substantial data migration or native mobile apps.
Established platform or operational system: £100,000 to £250,000
This is where software becomes a meaningful part of the operating model. The product may support multiple departments, customer accounts, reporting, workflow automation, payments, integrations and detailed administration.
For example, a bespoke Laravel platform that replaces fragmented back-office processes may need to connect with a CRM, accounting system, stock data and third-party APIs. The value can be considerable, but the delivery needs careful product ownership and a phased release plan.
Complex SaaS or enterprise-grade product: £250,000 and above
A multi-tenant SaaS platform, regulated system or high-volume product can exceed this level, particularly where it needs advanced security, audit trails, complex calculation engines, AI features, native applications or resilience requirements.
The headline build cost is only one part of the decision. Products at this level need a clear commercial model, senior governance and a realistic plan for support, optimisation and further development after launch.
What drives the cost of custom software?
The number of screens is not a reliable cost indicator. A simple-looking interface can conceal difficult logic, complicated data relationships or high-risk integrations. Equally, a product with many pages may be relatively straightforward if each follows established patterns.
Four areas usually have the greatest influence on investment:
Product complexity: Rules engines, permissions, calculations, multiple user types, bespoke workflows and reporting all add engineering and testing effort.
Integrations and data: Connecting reliable systems is usually efficient. Connecting legacy software, inconsistent data or poorly documented APIs requires more investigation and contingency.
Quality requirements: Accessibility, security, performance, testing and operational resilience are not cosmetic extras. The appropriate level depends on the risk of failure and the users the product serves.
Change and uncertainty: A stable, prioritised scope is easier to estimate. If critical requirements are still emerging, discovery and phased delivery protect the budget better than pretending certainty exists.
Design also affects cost, but not in the simplistic sense of making things look polished. Strong product design clarifies journeys, reduces support demand and helps users complete tasks correctly. For a conversion-led customer journey or a tool used every day by operational teams, that work often pays for itself.
Why cheaper estimates can become expensive
A low quote may be valid for a narrow, well-understood brief. It can also conceal omissions. If a proposal does not explain how requirements, UX, testing, deployment, security, project management and post-launch support will be handled, the initial figure is not the full cost of delivery.
There is a real trade-off between rate and delivery maturity. A lower day rate can be attractive, but it may create more management overhead for your team or leave gaps between strategy, design and engineering. Senior delivery costs more per day, yet can reduce rework by identifying difficult decisions early.
The same applies to off-the-shelf products. A subscription platform is often the right answer for standard needs such as accounting, email marketing or basic ecommerce. Bespoke software earns its place when existing tools force teams into workarounds, limit a differentiated customer experience or cannot support a commercially important process.
Budget for the full product lifecycle
Launching is a milestone, not the end of the investment. Custom software needs monitoring, security updates, backups, infrastructure management and support. It should also evolve as users reveal where a workflow needs refinement.
A sensible annual support and improvement budget is commonly around 15 to 25 per cent of the initial build cost, although this varies by platform complexity and how quickly the product is expected to develop. Some businesses prefer a retained delivery team; others plan quarterly improvement releases. Either approach is better than treating maintenance as an unexpected cost.
AI and automation deserve their own line in the budget. A practical AI feature may be relatively quick to add when it summarises approved information, classifies incoming requests or assists staff within a controlled process. Costs increase where the work involves private data, unreliable source material, human review, complex evaluation or significant usage volumes.
There are also ongoing model and infrastructure costs to consider. The right design keeps those predictable by defining what the AI can access, when it should act, and when it should hand a decision back to a person.
How to get an estimate you can use
The best estimates are not created from a feature wish list alone. They come from enough context to understand the commercial problem, users, constraints and measures of success.
Before asking for a proposal, define the process that is currently failing or limiting growth. Explain who uses it, what it costs in time or lost opportunity, the systems it must work with, and the outcome that would make the project worthwhile. A target such as reducing quote turnaround from three days to one, or increasing completed applications by 20 per cent, gives the team a better basis for prioritising.
Then separate launch-critical requirements from later opportunities. This is not about cutting ambition. It is about getting the most valuable version into users' hands sooner, learning from real behaviour and investing the next tranche of budget with better evidence.
A credible delivery partner should be able to explain assumptions, risks, dependencies and the level of confidence behind the estimate. Fixed-price work can suit a defined scope. For more uncertain products, a discovery phase followed by staged delivery is often more commercially responsible.
Electro approaches custom software as a product investment, combining strategy, design and engineering so that decisions made early support the operational and commercial case later on.
The most useful starting point is not asking for the cheapest build. It is defining the business change worth paying for, then funding the smallest reliable product that can prove it.