Estimate your software project: effort, cost range, and contract model.
Software project cost is effort hours multiplied by an hourly rate, plus third-party fees and tax. This tool estimates the hours from your scope, shows every assumption, and lets you apply your own rate.
Pick what you are building and how sure you are about the scope. You get an effort range in person-weeks, a cost range at an hourly rate you choose, and a suggested contract model with the reasoning behind it.
Free. No email needed to see the result. Your answers are saved only in this browser, and Clear my answers removes them. Anonymous counts of choices such as product type go to Google Analytics. Nothing you type is sent unless you submit the optional review form.
This is a rough planning range, not a quote. Diffwise.app does not publish prices, and nothing here is what Diffwise.app or any supplier will charge.
How this estimate works
No public, peer-reviewed dataset gives hours per app feature. So every effort range here is labelled a planning assumption, with its reasoning shown. Vendor guides that sell development services publish similar figures. We use them only to check that our assumptions are not unusually low or high.
- Effort hours per item, as a low and a high figure. Each size line and feature line has its own range.
- Extra client work for each additional platform. Only the client-side share of a line grows when you add a second mobile platform or a mobile client.
- Design, testing and coordination added as percentages of the build hours.
- The range widened to match how settled your scope is, using the Cone of Uncertainty. A well-defined scope rarely widens it. An idea-stage scope is shown with a warning.
- Cost equals hours times the rate you choose. The tool never supplies a rate for you.
Hours are not calendar time. The tool shows person-weeks of effort, at a productive-hours figure you can edit. Calendar time depends on team size, decisions, and reviews. As a check, the booking web app described in our cost guide comes to 330-680hours here, against 396-665 hours in the guide's own illustrative estimate. Both are planning assumptions.
We do not apply a blanket saving for AI coding tools. We found no public primary evidence of a stable figure, so a discount would be a guess. If a supplier applies one, ask what it is based on.
Why the range can be wide
Steve McConnell's Cone of Uncertainty (Construx white paper, version 1, January 2010) describes how estimate error shrinks as decisions are made. It is the best case for skilled estimators. It narrows only when decisions remove variability, and it widens again if the product definition changes. The tool never makes your range narrower than the band for your scope stage.
| Level | McConnell milestone | In everyday words | Band in the paper | Applied here | Basis |
|---|---|---|---|---|---|
| A | User Interface Design Complete | Final designs exist, every user task is written down with a 'done when' check, and integrations are known. | 0.8x to 1.25x | 0.8x to 1.25x | McConnell (2010), Figure 1: McConnell Figure 1, skilled estimators, best case. |
| B | Requirements Complete | Every user task is written down with a 'done when' check, and integrations are mostly known. | 0.67x to 1.5x | 0.67x to 1.5x | McConnell (2010), Figure 1: McConnell Figure 1. |
| C | Approved Product Definition | The product is described, for example as a feature list, but tasks are not yet testable. | 0.5x to 2x | 0.5x to 2x | McConnell (2010), Figure 1: McConnell Figure 1. |
| D | Initial Concept | The scope is an idea with little written down. | 0.25x to 4x (a 16x spread) | 0.5x to 2x | Planning assumption: McConnell's band is 0.25x to 4x. A 16x spread is not useful for a planning decision, so this tool shows the level C range and a visible warning. |
Your scope is at the idea stage. At this stage, McConnell's Cone of Uncertainty puts even skilled estimates at anywhere from a quarter of the true effort to four times it. Treat this range as a conversation starter. Writing down user tasks and 'done when' checks narrows it more than any calculator can.
Every number we use
The page and the tool read from the same table, so the text and the maths cannot drift apart. All hours are low and high figures. Client share is the part of a line that grows with the platform multiplier. Sources were accessed on 2026-11-18.
Effort lines
| Line | Applies to | Low (h) | High (h) | Client share | Basis | Reasoning | Cross-checked against |
|---|---|---|---|---|---|---|---|
| Discovery | Marketing or content website | 8 | 24 | 0 | Planning assumption | Smaller than for apps: page inventory, content owners, form routing. | None |
| Discovery | Web app, Mobile app, Web app and mobile app | 24 | 40 | 0 | Planning assumption | Equal to the Diffwise illustrative estimate (24 to 40 h). | Diffwise.app (opens in a new tab) |
| Discovery | AI feature added to an existing product | 16 | 32 | 0 | Planning assumption | Narrower scope, but the existing system has to be read first. | None |
| Hosting and release | Marketing or content website | 8 | 16 | 0 | Planning assumption | Hosting, domain, redirects, release checks. | None |
| Hosting and release | Web app | 20 | 35 | 0 | Planning assumption | Equal to the Diffwise illustrative estimate (hosting and release, 20 to 35 h). | Diffwise.app (opens in a new tab) |
| Hosting, build and release | Mobile app | 24 | 40 | 0 | Planning assumption | The web figure plus build signing and a mobile CI pipeline. | Diffwise.app (opens in a new tab) |
| Hosting, build and release | Web app and mobile app | 36 | 60 | 0 | Planning assumption | Web and mobile pipelines, with some shared work. | None |
| Environment access and onboarding | AI feature added to an existing product | 24 | 56 | 0 | Planning assumption | 8 to 16 h environment access plus 16 to 40 h reading and onboarding to an existing codebase. The width reflects unknown code quality. | None |
| App store release (per platform) (per platform) | Mobile app, Web app and mobile app | 8 | 16 | 0 | Planning assumption | Store listing, review submission, and fixes after store review. Multiplied by 1 or 2 platforms. | None |
| Main user task (per unit) | Apps and AI features | 30 | 80 | 0.5 | Planning assumption | The upper end equals the Diffwise illustrative booking flow (50 to 80 h) and admin queue (50 to 80 h). The lower end covers simple list, detail and edit tasks. | Diffwise.app (opens in a new tab) |
| Page layout (per unit) | Marketing or content website | 6 | 16 | 1 | Planning assumption | One responsive layout built from a design, with content wiring and states. | None |
| Email sign-in with password reset | Apps and AI features | 24 | 48 | 0.4 | Planning assumption | The Diffwise estimate has 30 to 50 h for accounts and roles together. RapidDev has 40 to 80 h, which includes more (OAuth, MFA); those are separate lines here. | Diffwise.app (opens in a new tab); RapidDev (opens in a new tab) (vendor) |
| Sign-in with Google, Apple, or company single sign-on | Apps and AI features | 16 | 40 | 0.3 | Planning assumption | Each provider needs configuration, account-linking rules and testing. Enterprise SAML sits at the high end. | None |
| Two-step verification | Apps and AI features | 12 | 24 | 0.4 | Planning assumption | Enrolment, recovery codes, and the extra test paths. | None |
| Extra user roles with different permissions (per unit) | Apps and AI features | 12 | 24 | 0.3 | Planning assumption | Each role adds permission rules, tests, and admin views (Diffwise cost-driver table). | Diffwise.app (opens in a new tab) |
| Team or organisation accounts with invites | Apps and AI features | 40 | 80 | 0.3 | Planning assumption | Invites, membership, data separation between organisations, and ownership transfer. | None |
| One-off card payments (hosted checkout) | All types | 40 | 70 | 0.3 | Planning assumption | Equal to the Diffwise illustrative card deposits line (hosted checkout, webhooks, refunds, failed payments). Lower than RapidDev's 60 to 120 h because hosted checkout is assumed. | Diffwise.app (opens in a new tab); RapidDev (opens in a new tab) (vendor) |
| Subscriptions and billing | Apps and AI features | 60 | 120 | 0.3 | Planning assumption | Plans, trials, proration, failed-payment retries, and the customer portal. Matches RapidDev's payments range. | RapidDev (opens in a new tab) (vendor) |
| Marketplace payouts to other people | Apps and AI features | 100 | 200 | 0.3 | Planning assumption | Onboarding sellers through the payment provider, split payouts, refunds across parties, and reporting. | None |
| Admin screens to review and correct records | Apps and AI features | 50 | 80 | 0 | Planning assumption | Equal to the Diffwise illustrative staff admin queue. Client share is 0 because admin is assumed to be web-only. | Diffwise.app (opens in a new tab) |
| Advanced admin: reports, bulk actions, audit log | Apps and AI features | 120 | 240 | 0 | Planning assumption | Inside RapidDev's 120 to 300 h. Replaces the basic admin line. | RapidDev (opens in a new tab) (vendor) |
| Reports or data export | Apps and AI features | 16 | 30 | 0.1 | Planning assumption | Equal to the Diffwise illustrative report export line. | Diffwise.app (opens in a new tab) |
| Email notifications | All types | 16 | 30 | 0 | Planning assumption | Equal to the Diffwise illustrative email notifications line. | Diffwise.app (opens in a new tab) |
| Push notifications | Mobile app, Web app and mobile app | 20 | 40 | 0.5 | Planning assumption | Equal to RapidDev's range (20 to 40 h). | RapidDev (opens in a new tab) (vendor) |
| In-app chat or live updates | Apps and AI features | 120 | 250 | 0.5 | Planning assumption | Equal to RapidDev's range (120 to 250 h). | RapidDev (opens in a new tab) (vendor) |
| File or image uploads | All types | 16 | 40 | 0.4 | Planning assumption | Storage, size and type checks, previews, and permission to view. | None |
| Search and filters | All types | 24 | 60 | 0.5 | Planning assumption | Filters and indexing. Full-text or relevance tuning sits at the high end. | None |
| Content editor for non-developers | All types | 24 | 60 | 0.3 | Planning assumption | Content model, editor setup, preview, and editor training notes. | None |
| Forms (contact, lead, booking request) (per unit) | Marketing or content website | 8 | 16 | 0.5 | Planning assumption | Validation, server handling, delivery, and confirmation. | None |
| More than one language (per unit) | All types | 12 | 30 | 0.6 | Planning assumption | The first extra language carries the setup work. Translation itself is excluded. | None |
| Maps or location | Apps and AI features | 60 | 120 | 0.6 | Planning assumption | Equal to RapidDev's range (60 to 120 h). | RapidDev (opens in a new tab) (vendor) |
| Works offline and syncs later | Mobile app, Web app and mobile app | 60 | 160 | 0.8 | Planning assumption | Local storage, a sync queue, conflict rules, and testing with flaky connections. Wide because conflict rules vary greatly. | None |
| Connections to other systems (CRM, ERP, calendar, accounting, and so on) (per unit) | All types | 24 | 60 | 0 | Planning assumption | Aalpha uses about 32 h each. The high end covers poor documentation, no sandbox, rate limits, and retries. | Aalpha (opens in a new tab) (vendor) |
| Move data from an existing system | All types | 24 | 80 | 0 | Planning assumption | Mapping, a scripted import, verification, and a rehearsal. Unknown source data quality explains the range. | None |
| Product analytics events and a simple dashboard | All types | 16 | 40 | 0.3 | Planning assumption | Event plan, instrumentation, and one dashboard. | None |
| AI that drafts, summarises, or classifies text using a hosted model | Apps and AI features | 40 | 100 | 0.3 | Planning assumption | Prompting, model API integration, output checks, an evaluation set, fallbacks, and usage limits. Below the low end of RapidDev's 80 to 300+ h. | RapidDev (opens in a new tab) (vendor) |
| AI that answers from your own documents or data (retrieval) | Apps and AI features | 100 | 240 | 0.3 | Planning assumption | Ingestion, chunking, retrieval, access control on documents, an evaluation set, and citations. Inside RapidDev's range. | RapidDev (opens in a new tab) (vendor) |
| AI that takes actions in your systems (tool use, agents) | Apps and AI features | 120 | 300 | 0.3 | Planning assumption | Tool definitions, permission boundaries, confirmation steps, logging, an evaluation set, and failure handling. Equal to RapidDev's upper range. | RapidDev (opens in a new tab) (vendor) |
| Health, financial, or children's data, or another regulated area | All types | 40 | 120 | 0 | Planning assumption | Extra security review, audit logging, data-retention work, and evidence for the buyer's compliance process. External legal, audit or certification fees are excluded. | None |
Design, testing and coordination
| Line | Low (%) | High (%) | Basis | Reasoning | Cross-checked against |
|---|---|---|---|---|---|
| Design: clear, accessible defaults | 5 | 10 | Planning assumption | Adapting a library still needs layouts and states. | None |
| Design: custom interface | 15 | 22 | Planning assumption | The Diffwise estimate's design line is about 25% of its build lines. Purrweb puts design at about 25% of the total. Set slightly lower because that design line also covered empty and error states, which the testing line partly holds here. | Diffwise.app (opens in a new tab); Purrweb (opens in a new tab) (vendor) |
| Design: new brand-level system | 25 | 35 | Planning assumption | Adds identity and system work on top of custom screens. | None |
| Testing | 20 | 25 | Planning assumption | Techsy uses 20%. The Diffwise estimate's testing is about 26 to 30% of its build. Purrweb is about 13% of the total (about 20% of development). Includes accessibility checks to WCAG 2.2 AA, device checks, and failure paths. | Diffwise.app (opens in a new tab); Techsy (opens in a new tab) (vendor); Purrweb (opens in a new tab) (vendor) |
| Coordination and management | 10 | 12 | Planning assumption | Applied to fixed lines, build, design and testing. Aalpha uses 10%, Techsy 15%, and the Diffwise estimate is about 10 to 11%. | Diffwise.app (opens in a new tab); Aalpha (opens in a new tab) (vendor); Techsy (opens in a new tab) (vendor) |
Platform multipliers
| Case | Low | High | Basis | Reasoning |
|---|---|---|---|---|
| Mobile, both platforms, one shared codebase | 1.15 | 1.25 | Planning assumption | One codebase plus platform-specific fixes, device testing, and native modules. |
| Mobile, both platforms, separate native apps | 1.8 | 2 | Planning assumption | Two client codebases. Under 2.0 at the low end because design, API and decisions are shared. |
| Mobile, both platforms, approach not chosen | 1.15 | 2 | Planning assumption | Cross-platform for the low figure and native for the high figure. The choice of approach alone moves the high figure this much. |
| Web and mobile: mobile client as a share of the web client | 0.75 | 1 | Planning assumption | The API and design are reused, but every screen is rebuilt. The total multiplier is 1 plus this share times the mobile multiplier. |
Other assumptions
| Assumption | Default | Allowed range | Basis | Reasoning | Cross-checked against |
|---|---|---|---|---|---|
| Productive hours per person per week | 30 h | 20 to 40 h | Planning assumption | A 37.5 to 40 hour week less meetings, reviews and interruptions. Aalpha uses 32. | Aalpha (opens in a new tab) (vendor) |
| Hours per day (converts day rates to hourly) | 7.5 h | 6 to 9 h | Planning assumption | A typical working day less a short break. Only used for day-rate presets. | None |
| Largest scope (high estimate) for a single fixed price | 800 h | 100 to 5000 h | Planning assumption | Above roughly this size, a single fixed price usually has to cover risks that cannot yet be tested. Pricing and accepting one phase at a time keeps each fixed scope small enough to verify. The threshold is a planning assumption, not a published standard. | Diffwise.app (opens in a new tab); US Federal Acquisition Regulation (acquisition.gov) (opens in a new tab) |
Hourly rate reference figures
These are public reference figures. Employee wages are what a developer is paid, not what an agency or contractor charges. A supplier's rate also covers employer taxes, benefits, equipment, management, sales time, unbillable hours, and profit. Use the rate from the quotes you receive whenever you can. No rate is preselected. A preset is a reference point, not a recommendation.
| Reference | Figure | What it measures | Data date | Source | Caveat |
|---|---|---|---|---|---|
| US software developer median wage (employee) | 65.38 USD per hour | Median hourly wage of employed software developers (SOC 15-1252), all US, excluding benefits. Annual median $135,980. | May 2025 | US Bureau of Labor Statistics (opens in a new tab) | Official data. A wage, not a rate a supplier charges. |
| US software developer 90th-percentile wage (employee) | 103.21 USD per hour | 90th-percentile hourly wage, SOC 15-1252. (10th percentile 39.64, 25th 50.58, 75th 82.68, mean 71.20.) | May 2025 | US Bureau of Labor Statistics (opens in a new tab) | Official data. A wage, not a rate a supplier charges. |
| US in-house developer employer cost (derived) | 94.48 USD per hour | Derived: median wage 65.38 divided by 0.692. 0.692 is the share of wages in total compensation for private-industry professional and related occupations ($50.98 of $73.69 per hour worked). Adds benefits and payroll taxes only. | Wage May 2025; compensation June 2026 | US Bureau of Labor Statistics (opens in a new tab); US Bureau of Labor Statistics (opens in a new tab) | The inputs are official; the division is ours. It still excludes overhead, unbillable time and profit, so it is not an agency rate. |
| UK contract software developer, median advertised day rate | 529 GBP per day | Median daily rate in UK contract job adverts titled Software Developer. 10th percentile 363, 90th percentile 626. Based on 275 adverts that quoted a rate. Converted to hourly with the editable hours-per-day assumption. | 6 months to 2026-11-18 | IT Jobs Watch (opens in a new tab) | Advertised asking rates, not agreed rates. An individual contractor, not an agency. |
| Canada software developer median wage (employee) | 50 CAD per hour | National median hourly wage for software developers (NOC 21232), Labour Force Survey. Low 30.87, high 79.33. | 2024-2025 | Job Bank, Government of Canada (opens in a new tab) | Official data. A wage, not a rate a supplier charges. |
| Clutch: typical hourly rate of listed development companies | 25 to 49 USD per hour | Clutch says most software development companies listed on it charge between $24 and $49 an hour (the page also shows $25 to $49). | Page updated 2026-09-21 | Clutch (opens in a new tab) (vendor) | Published by a directory that profits from listings. Sample size and period are not stated, and many listed companies are offshore. Use as one reference only. |
Considered and not offered as presets
- UK ONS ASHE (SOC 2020 code 2134) is the right official UK employee source, but its occupation tables were not checked for this tool. Until then, the National Careers Service gives a UK range of 30,000 pounds (starter) to 75,000 pounds (experienced) a year for 37 to 40 hours a week. It is context, not a preset.
- Eurostat Structure of Earnings Survey 2022 reports only broad groups (EU-27 mean gross hourly earnings for all professionals was 24.40 euros), so it is not a software figure. Euro users type their own rate.
- Upwork's published freelancer bands could not be opened and checked for this tool, so they are not offered.
- Arc, Toptal and Accelerance publish rate guides without fixed figures, dates, or a reproducible method. Uncited ranges in competitor articles are excluded.
Not included in any figure
- Third-party fees: hosting, payment processing, email or SMS delivery, model API usage for AI features, app store developer accounts
- Tax: VAT, GST, HST, or sales tax
- Content writing and translation
- Legal or compliance audits
- Ongoing maintenance and support after release
- Your own team's time
Fixed price, time and materials, or fixed capacity?
The tool checks five rules in order and shows the first that matches, with the reasoning. Fixed price fixes the scope and the cost. Fixed capacity fixes the team and the cost per month and lets the scope move. Time and materials charges for hours worked, so you carry more of the effort risk. This is general information, not legal advice.
- Paid discovery first if the scope is at the idea stage, if systems you must connect have no documented APIs or test accounts, or if the product handles regulated data without written requirements.
- Fixed price if the scope is settled (levels A or B), unlikely to change, and its integrations are known. If the high estimate, before the uncertainty allowance, is above 800 hours, fixed price per phase instead. That threshold is a planning assumption, not a published standard.
- Fixed capacity if you expect ongoing work and some change in priorities.
- Capped time and materials, in phases if your budget is a hard ceiling.
- Time and materials with approval points otherwise.
Paid discovery first, then choose
Your answers suggest the scope is not settled enough to price the build reliably. A short, separately priced discovery phase should produce a written task list, an integration review, a risk list, acceptance checks, and a revised estimate that you own and could take to any supplier. The GOV.UK Service Manual says you should not start building during discovery, and that stopping after discovery is not a failure. Run this estimator again with what discovery produces.
Sources: GOV.UK Service Manual (opens in a new tab); Steve McConnell, Construx (opens in a new tab)
Fixed price
Your scope is written as testable tasks, is unlikely to change, and its integrations are known. That is when a fixed price works best: the supplier takes on the effort risk for the agreed scope. Expect the price to include a buffer for that risk, and every change to go through a change request. US federal guidance describes firm-fixed-price as suitable when specifications are 'reasonably definite' and performance risks can be identified and costed (FAR 16.202-2).
Sources: US Federal Acquisition Regulation (acquisition.gov) (opens in a new tab); GOV.UK (opens in a new tab); Diffwise.app (opens in a new tab)
Fixed price per phase
Your scope is settled, but large. Split it into phases, each with its own fixed price and acceptance checks, so each commitment covers work that can be fully tested. You can also change supplier between phases.
Sources: US Federal Acquisition Regulation (acquisition.gov) (opens in a new tab); Diffwise.app (opens in a new tab)
Fixed capacity (monthly team or retainer)
You expect ongoing work and some change in priorities. A fixed-capacity agreement buys a set amount of team time each month, and you decide what it works on. You do not fix the scope in advance. Compare what capacity each offer buys, how priorities are set, what you see each week, and what happens to unused time. This differs from fixed price: fixed price fixes the scope and the cost, while fixed capacity fixes the team and the cost per period and lets the scope move.
Sources: Diffwise.app (opens in a new tab)
Capped time and materials, in phases
You need a firm budget limit, but your scope will still move. Pay for time worked up to an agreed cap for each phase. Check what the contract says happens at the cap: under the withdrawn UK government framework guidance, the supplier finished the work at its own cost, while many private contracts simply stop work until you approve more. US federal time-and-materials contracts must include a ceiling price (FAR 16.601).
Sources: GOV.UK (opens in a new tab); US Federal Acquisition Regulation (acquisition.gov) (opens in a new tab)
Time and materials with approval points
Your plan will change as you learn, so a model that prices change cheaply fits best. You pay for hours worked, which puts the effort risk on you. Ask for a weekly report of hours against planned work, a working demo at each milestone, and an approval point before each new budget. US federal guidance reserves time and materials for work that cannot be estimated accurately up front (FAR 16.601), which matches an early product.
Sources: US Federal Acquisition Regulation (acquisition.gov) (opens in a new tab); GOV.UK (opens in a new tab)
US Federal Acquisition Regulation sections are US federal procurement guidance, not law that applies to private contracts, and the regulation is being rewritten. We cite each section as read on 2026-11-18. The withdrawn GOV.UK guidance is cited as historical. For the full discussion, read fixed price vs time and materials software contracts and the software project handover checklist. To turn a result into a fuller brief, use the project brief builder.
Questions about the estimator
How much does it cost to build an MVP?
There is no honest single figure. Cost is effort multiplied by rate. With this tool's planning assumptions, a small booking web app (sign-in, one booking flow, an admin queue, card deposits, email, export) comes to about 330-680 hours before fees and tax. Multiply that by the rate in the quotes you receive. Your scope will differ.
How do I estimate the cost of a software project?
List the user tasks, give each a low and high hour range, add design, testing and coordination, widen the range for how settled the scope is, then multiply by a rate. Add third-party fees and tax separately. Read the estimate method in our cost guide.
Should I choose fixed price or time and materials?
Fixed price suits a small, stable scope written as testable checks. Time and materials suits work you expect to change, ideally with a cap and approval points. If you cannot yet write testable checks, pay for discovery first. Read the contract model guide.
What is the difference between a fixed capacity model and fixed price?
Fixed price fixes scope and cost. Fixed capacity fixes the team and the cost per month, and you choose what the team works on, so the scope can move.
Why is the estimate a range, and why is it so wide?
Early estimates are uncertain. McConnell's Cone of Uncertainty shows even skilled estimates at the idea stage can be off by a factor of four. The range narrows as you write requirements and agree designs.
What hourly rate should I use?
Use the rate in the quotes you receive. If you have none yet, the tool offers public reference figures, such as US Bureau of Labor Statistics wages. A wage is what an employee earns, not what an agency or contractor charges, which also covers overheads and profit.
How do I compare quotes from MVP development companies?
Give every supplier the same brief, then compare scope, assumptions and exclusions before totals. The brief this tool generates is built for that. Read how to compare web app quotes.
Does the estimator store or send my answers?
The estimator runs in your browser. It saves a draft of your answers in this browser only, so the page remembers them when you come back. The Clear my answers button removes that draft. Anonymous usage events go to Google Analytics: they record choices such as the product type, scope stage and suggested contract model, and never hours, costs, rates or anything you type into a text field. Your full answers reach Diffwise.app only if you submit the optional review form with your email.
Sources
Every link was read on 2026-11-18. Sources marked vendor are published by companies that sell development services or hiring. We use them only as cross-checks.
- Steve McConnell, Construx, Software Development's Cone of Uncertainty (white paper, version 1). Data date: January 2010. Accessed 2026-11-18. https://www.construx.com/wp-content/uploads/2019/02/CxWhitePaper_ConeOfUncertainty.pdf (opens in a new tab)
- US Bureau of Labor Statistics, Occupational Employment and Wage Statistics: Software Developers (15-1252). Data date: May 2025. Accessed 2026-11-18. https://www.bls.gov/oes/current/oes151252.htm (opens in a new tab)
- US Bureau of Labor Statistics, Employer Costs for Employee Compensation, Table 4. Data date: June 2026 (released 2026-09-09). Accessed 2026-11-18. https://www.bls.gov/news.release/ecec.t04.htm (opens in a new tab)
- Job Bank, Government of Canada, Software developer wages (NOC 21232). Data date: 2024-2025. Accessed 2026-11-18. https://www.jobbank.gc.ca/marketreport/wages-occupation/22548/ca (opens in a new tab)
- IT Jobs Watch, UK contract Software Developer rates (job-advert data). Data date: 6 months to 2026-11-18. Accessed 2026-11-18. https://www.itjobswatch.co.uk/contracts/uk/software%20developer.do (opens in a new tab)
- Clutch, Software development pricing. Data date: Page updated 2026-09-21. Accessed 2026-11-18. Vendor-published. https://clutch.co/developers/pricing (opens in a new tab)
- National Careers Service (UK), Software developer job profile. Data date: Not dated on page. Accessed 2026-11-18. https://nationalcareers.service.gov.uk/job-profiles/software-developer (opens in a new tab)
- GOV.UK, How to pay for Digital Outcomes and Specialists services (withdrawn 31 December 2025; historical guidance). Data date: Withdrawn 2025-12-31. Accessed 2026-11-18. https://www.gov.uk/guidance/how-to-pay-for-digital-outcomes-and-specialists-services (opens in a new tab)
- GOV.UK, The Digital, Data and Technology Playbook. Data date: Not dated on page. Accessed 2026-11-18. https://www.gov.uk/government/publications/the-digital-data-and-technology-playbook/the-digital-data-and-technology-playbook (opens in a new tab)
- GOV.UK Service Manual, How the discovery phase works. Data date: Not dated on page. Accessed 2026-11-18. https://www.gov.uk/service-manual/agile-delivery/how-the-discovery-phase-works (opens in a new tab)
- US Federal Acquisition Regulation (acquisition.gov), FAR 16.104, Factors in selecting contract types. Data date: As read 2026-11-18. Accessed 2026-11-18. https://www.acquisition.gov/far/16.104 (opens in a new tab)
- US Federal Acquisition Regulation (acquisition.gov), FAR 16.202-2, Firm-fixed-price: application. Data date: As read 2026-11-18. Accessed 2026-11-18. https://www.acquisition.gov/far/16.202-2 (opens in a new tab)
- US Federal Acquisition Regulation (acquisition.gov), FAR 16.601, Time-and-materials contracts. Data date: As read 2026-11-18. Accessed 2026-11-18. https://www.acquisition.gov/far/16.601 (opens in a new tab)
- Diffwise.app, Web app and MVP development cost in 2026 (illustrative booking MVP estimate, labelled planning assumptions). Data date: 2026-11-03. Accessed 2026-11-18. https://www.diffwise.app/blog/web-app-and-mvp-development-cost (opens in a new tab)
- Diffwise.app, Fixed price vs time and materials software contracts. Data date: 2026-11-04. Accessed 2026-11-18. https://www.diffwise.app/blog/fixed-price-vs-time-and-materials-software-contracts (opens in a new tab)
- RapidDev, App development cost guide. Data date: Page says updated November 2026. Accessed 2026-11-18. Vendor-published. https://www.rapidevelopers.com/md/cost/app-development-cost (opens in a new tab)
- Aalpha, MVP development cost calculator. Data date: Not dated on page. Accessed 2026-11-18. Vendor-published. https://www.aalpha.net/calculators/mvp-development-cost-calculator/ (opens in a new tab)
- Techsy, MVP cost calculator. Data date: Not dated on page. Accessed 2026-11-18. Vendor-published. https://techsy.io/en/resources/tools/mvp-cost-calculator (opens in a new tab)
- Purrweb, App cost calculator. Data date: Not dated on page. Accessed 2026-11-18. Vendor-published. https://www.purrweb.com/app-cost-calculator/ (opens in a new tab)
Have a scope you want checked?
Tell us what you want to build or improve. We will help you find a useful next step.