Build vs Buy Calculator
Most teams pick build when they should buy. Some buy when they should build. This calculator costs them out side by side over 3 years.
- No signup required
- 100% free
- 3-year horizon modelled
Project type
Pick the closest match. We'll tune the rest in the next step.
Closest match is fine. Scope and complexity fine-tune the estimate on the next steps.
A build-vs-buy decision compares the 3-year total cost of building software in-house, buying SaaS, or running a hybrid model. The honest version includes hidden costs like SaaS renewal creep, integration glue, in-house ramp time, and compliance work. This calculator surfaces all three options as side-by-side bars, with the hidden 40 percent broken out so you can see what most comparison spreadsheets quietly leave off.
01 Hidden costs
The 40 percent most build-vs-buy models miss
Six line items that don't appear in the vendor pitch deck or the engineering estimate. Each one shifts the answer.
- 01
Renewal creep on the SaaS option
Year 1 list price isn't the price. Most vendors lift by 8 to 15 percent per year, and the year-3 invoice usually lands at 1.4 to 1.6x what you signed for. On a $40K/year tool that's $24K of extra spend you didn't model.
- 02
Vendor lock-in and switching cost
When you outgrow the SaaS, migrating off it is rarely free. Data export, retraining, parallel-run periods, and the integration rewrite typically run 3 to 6 months of one engineer's time, or $30K to $90K of contractor cost.
- 03
The integration tax nobody quotes for
Every commercial tool wants to be the system of record. Webhooks break on vendor updates, SSO needs an enterprise tier, and the 'free API' rate-limits at the volumes you actually run at. Budget 15 to 25 percent of the licence cost in glue code per year.
- 04
Opportunity cost of waiting
An in-house build that takes 9 months delays revenue or savings by 9 months. If the workflow saves $20K/month, that's $180K left on the table before line one of code ships. The calculator surfaces this as a separate bar, not a fudge factor.
- 05
Technical debt from speed builds
Building in-house in a hurry buys you a working v1 and a v2 rewrite. Across 100+ projects we see roughly 30 percent of the original build budget spent again in years 2 and 3 on refactors, performance work, and security hardening.
- 06
SOC 2, GDPR, HIPAA documentation
If you sell to enterprise, the audit paperwork is part of the build. A first-time SOC 2 Type II runs $20K to $80K in tooling plus 2 to 4 months of one person's time. Most build estimates leave it out entirely.
02 Decision tree
Which model fits your situation
Four short scenarios. Find yours, then run the calculator with the answer already in mind.
- 01
Speed is your constraint. Buy.
You need to ship in 8 to 12 weeks and the workflow is solved by an existing category. Buy. Accept the feature gaps for the first 18 months. Revisit when you hit either a pricing wall or a workflow your team has clearly outgrown.
- 02
Capability is the constraint, ownership matters. Go hybrid.
You can buy the commodity layer (auth, billing, helpdesk, analytics) and build the layer that's actually yours. This is the right answer 60 to 70 percent of the time for mid-market teams. Rent the boring parts. Own the part you'd refuse to switch off.
- 03
Differentiation IS the software. Build.
If the product is the software (a vertical SaaS, an AI workflow your competitors can't copy, a routing engine that's your moat), build. The TCO on the calculator will look ugly versus SaaS. The TCO on losing your moat is uglier.
- 04
Sub-$200K total budget, no in-house team. Buy plus light customisation.
You don't have the runway or the headcount to build cleanly. Buy the best-fit platform, then spend $30K to $60K on customisation, integrations, and a clean onboarding. The hybrid model fails at this budget because you can't sustain the internal half.
03 Honest framing
When in-house pays off vs. when outsourcing actually wins
Both models can be the right answer. The conditions matter more than the model.
In-house wins when
- The software is the product, not a supporting tool. Your roadmap changes monthly and you can't wait on a vendor.
- You already have 2+ engineers in the relevant stack and a product owner who owns the workflow end-to-end.
- The work compounds. Building a feature now makes the next 5 features cheaper, not more expensive.
- Your differentiation is in how the workflow runs, not what it does. SaaS can't be your moat if everyone else can buy the same SaaS.
Outsourcing wins when
- You need a working v1 in 12 to 16 weeks and hiring will take 16 weeks just to get the first engineer in.
- The project is well-scoped and bounded. Migrations, redesigns, integration builds, a fixed-scope MVP. Less suited to permanent capability work.
- You're testing a hypothesis. Spending $80K to validate before committing to a 4-person internal team is the cheaper version of the decision.
- Your team's expertise sits elsewhere. Backend team taking on iOS, or a product team taking on its first ML build. The ramp tax kills the in-house math.
04 Methodology
How we built the numbers
Based on cost data from 100+ projects across SaaS, fintech, and hospitality builds between 2020 and 2026. Salary bands are blended US/UK/EU senior engineer rates. Vendor renewal curves come from public pricing pages and the 30 to 40 commercial tools our portfolio companies run in production.
Sample size
100+ shipped projects, weighted toward $80K to $750K builds. Outliers above $1.5M trimmed.
Time window
2020 to 2026. We re-baseline rates every 6 months as the senior engineer market moves.
What we don't model
Equity-funded team economics, offshore-only builds, and any vendor under 3 years old. Three known blind spots, called out so you can adjust manually.
05 Related
Where to go next
If the calculator pointed you toward build, these are the two pages worth reading next.
Want a second opinion on the answer the calculator gave you?
30 minutes with a founder. No deck. We'll pressure-test your numbers and tell you if we'd build it, buy it, or wait.
How it works
Compares total cost of ownership across three models (in-house team, outsourced development, and hybrid), including hidden costs like recruitment, onboarding, benefits, management overhead, and long-term maintenance.
Frequently asked questions
- When the work sits outside your team's current strengths and would need 3 to 6 months of ramp-up. A backend team picking up a React Native rewrite, or a product team taking on its first ML workflow. The fully loaded cost of internal ramp (salary + opportunity cost on what they're not shipping) usually beats an external squad only after the second similar project. For one-off work, outsourcing wins on cost and time.
- Around 50 to 80 seats for most B2B SaaS, and earlier if your usage is heavy (high event volume, large data, API-driven workflows). The break-even on building is somewhere between $150K and $400K of annual recurring spend, depending on category. The calculator shows the cross-over year, not just the 3-year total, so you can decide whether to delay the build until you're closer to that line.
- There's a manual override for it. We treat differentiation as a revenue uplift on the build column (default 0 percent, you set the number). A typical defensible workflow lifts close rate or retention by 3 to 8 percent. If you can't articulate the uplift, that's a signal the build isn't actually a differentiator and you should buy.
- For internal tools, yes. Retool, n8n, Bubble, Zapier-class platforms slot into the SaaS column cleanly. For customer-facing products at scale, the calculator will usually flag no-code as a short-term win that becomes the most expensive option by year 3 once the rebuild starts. Use no-code to validate. Switch to build or buy once the workflow is proven.
- Fair question. We tell roughly 30 percent of inbound prospects to buy a SaaS instead of hiring us. The calculator reflects that. If your problem is solved by an existing $50/seat tool and you have 12 people, we'd rather you save the cash and come back when you've outgrown it. The bias we do carry: we think custom is undervalued for differentiation work, because most teams underprice their own moat.
- Yes, separately. New engineers ship at roughly 30 percent productivity for the first 3 months and 70 percent through month 6. The model rolls that into the in-house bar, which is why in-house total cost looks higher than the raw salary math suggests. If you've already got a high-performing team in the right stack, you can flip the ramp toggle off.