The most honest answer to this question is: it depends but what it depends on is knowable, the ranges are real, and understanding them before you go to market is worth more than any number you could get from a quick Google. This is a complete, no-fluff breakdown of what custom software actually costs in 2026, what drives the variance, and how to use that information to budget realistically.
In this article
- Why pricing is so opaque and what that means for you
- Cost ranges by project type
- The six factors that drive cost variance
- What agency location actually means for cost and quality
- The hidden costs most budgets miss
- How to get a meaningful estimate
Custom software pricing has a transparency problem. Most agencies do not publish rates, most projects do not disclose budgets, and the variance between a credible estimate and a lowball one designed to win the work can be enormous. The result is that most buyers approach the market without a realistic sense of what things cost, which makes them vulnerable to both overcharging and to cheap estimates that balloon once real complexity is discovered.
This guide is an attempt to fix that with ranges that reflect what production-quality software actually costs in 2026, what drives the variance within those ranges, and how to use that knowledge to ask better questions before you commit.
1. Why pricing is so opaque and what that means for you
Software development is priced opaquely for a structural reason: the cost of a project is almost entirely determined by its complexity, and complexity is genuinely difficult to assess without a detailed technical discovery. A custom web application can cost £15,000 or £500,000 depending on integrations, compliance requirements, performance demands, and architectural decisions that are not visible in a brief.
This creates a market dynamic where low estimates win competitive pitches and then change orders extend the project to its real cost. The buyers who get the best outcomes are the ones who understand the real cost ranges before going to market, can recognise an underestimate when they see one, and choose partners who are honest about complexity rather than optimistic about it.
The underestimate trap
A proposal that comes in significantly below market rate is not a good deal it is a signal that either the scope has not been understood, the complexity has not been assessed, or the quality of the output will reflect the price. Agencies that win on price and recover their margin through change orders are a well-established pattern in the market. Knowing what things should cost is the most reliable defence against it.
2. Cost ranges by project type
The following ranges reflect production-quality development by experienced teams in 2026. They assume a UK or US agency rate, full discovery and design phases, proper testing and QA, and post-launch stabilisation. They do not include ongoing maintenance or feature development after launch.
Simple internal tool or admin dashboard£15,000 – £35,000
A single-purpose tool used by internal staff: a reporting dashboard, a simple workflow tool, a data management interface. Limited integrations, no complex authentication, no public-facing components. Well-defined scope that does not change mid-build.
Typically includes: discovery, design, development, basic QA, deployment. Excludes: complex integrations, compliance work, ongoing maintenance.
Customer-facing web application£40,000 – £120,000
A web application used by external customers: account management, booking systems, client portals, marketplaces. Requires proper authentication, user management, payment integration where applicable, and the higher design and performance standard that external users expect. Wide range driven by integration complexity, feature depth, and compliance requirements.
Typically includes: full discovery, UX design, frontend and backend development, integration work, QA, deployment. Payment integration adds £5,000–£15,000.
Mobile application (single platform)£35,000 – £100,000
A native or cross-platform mobile application for iOS or Android. Cost is driven primarily by the number of screens, the complexity of native features required (camera, GPS, biometrics, offline functionality), and the integration with backend systems. App store submission and review adds timeline but minimal direct cost.
Cross-platform (both iOS and Android from a single codebase) adds 20–40% to single-platform cost. Native development for both platforms roughly doubles it.
SaaS platform with multi-tenancy and billing£80,000 – £300,000+
A software product sold to multiple clients, each with their own data, permissions, and configuration. Multi-tenancy architecture, subscription billing integration, onboarding flows, admin dashboards, and the level of reliability and performance required for a commercial product all contribute to the wide range. Most SaaS builds at the lower end of this range are MVPs; meaningful additional investment is required to reach a product that is genuinely ready to sell.
Lower-end estimates typically exclude: advanced analytics, API for third-party integrations, white-labelling, enterprise SSO, and the iteration required after early customer feedback.
AI agent or automation system£25,000 – £120,000
A purpose-built AI system: a voice agent, a document processing pipeline, a customer support AI, a workflow automation agent. Cost is driven primarily by the number and complexity of integrations, data quality and preparation requirements, compliance constraints, and the monitoring infrastructure needed for production operation. Simple, well-scoped AI agents with clean data and minimal integration complexity sit at the lower end. Complex multi-integration systems with compliance requirements sit at the upper end.
Data preparation and cleaning are almost always required and frequently underestimated budget an additional £5,000–£20,000 depending on the state of your existing data.
Enterprise system or legacy replacement£150,000 – £1,000,000+
A system that replaces a core business function: an ERP, a practice management system, a claims processing platform. The wide range reflects the enormous variance in complexity, integration depth, data migration requirements, and the regulatory and performance demands of business-critical systems. Very few enterprise builds come in below £200,000 when scoped honestly.
Data migration from legacy systems is a major cost driver; budget separately and early. Phased delivery is strongly recommended to manage risk at this scale.
3. The six factors that drive cost variance
01Integration complexity
The single largest driver of cost variance in most projects. Connecting to a modern, well-documented API with good test environments is straightforward. Connecting to a legacy system with limited documentation, restricted access, or unpredictable behaviour is not. Every additional integration adds cost not just for the connection itself, but for the error handling, testing, and ongoing maintenance it requires.
Before you get an estimate, list every system your new software needs to connect to and find out whether each has an accessible API with documentation. This one exercise will save significant estimate variance.
02Compliance and regulatory requirements
GDPR, HIPAA, FCA, PCI-DSS, and ISO 27001 compliance requirements add cost at the architecture, development, and testing stages. They are not optional extras that can be added after launch. If your system handles regulated data or operates in a regulated industry, compliance must be designed in from the start, and the cost must be in the original budget.
03Data quality and preparation
For systems that import, migrate, or learn from existing data, particularly AI systems, the state of the existing data is a major cost driver. Clean, structured, well-documented data is a foundation. Inconsistent, incomplete, or siloed data requires significant preparation work before it can be used. This cost is almost always underestimated because it is invisible until the data is actually examined.
04Performance and scalability requirements
A system that needs to handle ten concurrent users and one that needs to handle ten thousand are different engineering challenges with different costs. Performance requirements response time targets, concurrent user capacity, uptime SLAs need to be stated explicitly in the brief. Retrofitting performance after launch is significantly more expensive than designing for it from the start.
05Design complexity and UX depth
A customer-facing product used daily by external users requires a meaningfully higher level of UX design investment than an internal tool used occasionally by staff who are motivated to learn it. The cost of UX design is not just the design work itself; it includes user research, prototyping, usability testing, and the iteration cycles that follow. For consumer-facing products, underinvesting in design is one of the most reliable ways to build something that works, but nobody uses.
06Scope stability
Projects with stable, well-defined scope deliver predictably within budget. Projects where requirements evolve significantly during development cost more because every change in development requires rework of things already completed. A well-invested discovery phase that produces a stable, signed-off specification before development begins is the single most cost-effective thing you can do to keep a project within its original budget.
4. What agency location actually means for cost and quality
| Location | Typical day rate range | Considerations |
|---|---|---|
| UK / Western Europe | £500 – £1,200/day | Same timezone, direct communication, strong regulatory familiarity, higher rate |
| US / Canada | $600 – $1,400/day | Same as UK higher rate, same communication and quality benefits |
| Eastern Europe | £250 – £550/day | Strong technical quality, 1–3hr timezone difference, growing AI/SaaS capability |
| South Asia | £80 – £200/day | Large variance in quality, significant timezone gap, coordination overhead, higher management burden on client |
| South East Asia / LATAM | £120 – £300/day | Growing quality in specific markets, moderate timezone gap |
The day rate difference between UK and Eastern European agencies looks significant in isolation. Over a full project, the effective cost difference is smaller than it appears, for several reasons: Eastern European projects often require more client-side project management, timezone gaps slow feedback loops and extend timelines, and the coordination overhead of distributed teams adds cost that does not appear in the day rate.
For complex AI systems, compliance-sensitive projects, and anything requiring deep integration with UK or US business systems, the timezone and communication advantages of a UK or Western European agency typically justify the rate premium. For well-defined projects with stable scope and limited compliance complexity, Eastern European agencies with strong track records represent genuine value.
5. The hidden costs most budgets miss
- Discovery and scoping. A proper technical discovery the phase that produces the specification your estimate should be based on typically costs £3,000–£15,000 depending on project complexity. Agencies that provide free estimates without a discovery phase are estimating against incomplete information. The estimate will change once the discovery is done.
- Ongoing maintenance. Budget fifteen to twenty-five percent of build cost per year for routine maintenance, security updates, dependency management, and hosting. This is not optional; unmaintained software degrades in security and performance.
- Feature development. Software evolves. Budget at least one significant development phase per year. Most systems at twelve months post-launch have changed meaningfully from what was delivered.
- Internal time. Your team will spend time on requirements, testing, training, and change management. At senior hourly rates, this is typically ten to twenty percent of the build cost and is rarely in the budget.
- Third-party services. AI APIs, payment gateways, SMS providers, email delivery platforms, monitoring tools: the recurring costs of the services your software depends on add up quickly and are often excluded from initial cost discussions.
- Data preparation. If your project involves AI or data migration, auditing and cleaning your existing data is a high cost that rarely appears in the initial estimate because the data has not been examined yet.
6. How to get a meaningful estimate
A meaningful estimate is one based on enough information to assess the actual complexity of your project, not one produced from a thirty-minute conversation and a one-page brief. The steps that produce meaningful estimates:
- Invest in a proper discovery. A paid discovery phase, typically two to four weeks, produces a technical specification that can be estimated accurately. It is the only way to get a number that reflects the real cost rather than the assumed cost.
- List every integration before you ask for a price. Name each external system your new software needs to connect to, describe what the integration needs to do, and find out whether an accessible API exists. This is the information that most affects the estimate and most often arrives late.
- Be explicit about compliance requirements. State which regulatory frameworks apply upfront. Discovering compliance requirements mid-build is one of the most reliable ways to blow a budget.
- Ask for assumptions to be stated in writing. Every estimate rests on assumptions. Ask the agency to write them down. An estimate without stated assumptions is not an estimate it is a placeholder that will change the moment reality diverges from the assumptions.
- Compare proposals on scope, not just price. Two proposals that look like they are pricing the same thing may be pricing very different scopes. Check what is included and what is explicitly excluded before concluding that one is cheaper than the other.
The bottom line
Custom software in 2026 costs what it costs because good engineering is skilled, time-consuming work and the complexity of real business systems is routinely underestimated at the point of initial briefing. The ranges in this guide reflect production-quality development by experienced teams. They are not what you will find at the bottom of the market, and they are not what you should be trying to find.
The businesses that get the best return on custom software investment are the ones that budget realistically from the start, invest in a proper discovery before committing to a number, and choose partners who are honest about complexity rather than optimistic about it. The ones that struggle are the ones that choose the lowest estimate and spend the difference plus a premium on change orders, remediation, and rebuilding what was built too cheaply.
If you have a project in mind and want a realistic sense of what it would cost with the assumptions stated clearly and the complexity assessed, the SmartWayLabs team is honestly happy to work through it with you. We will tell you what we think it will take, and why.
Want an honest cost estimate for your software project?
SmartWayLabs provides realistic estimates based on actual discovery, with assumptions stated in writing and complexity assessed before a number is given. Talk to the team ↗