Build vs Buy Decisions
A structured framework for build, buy, partner or open source decisions: differentiation, total cost of ownership, time to market, risk, lock-in and exit costs.
Executive explanation
Build where software creates competitive advantage; buy (or use SaaS/open source) where it is a commodity. The hard part is honest assessment of differentiation and full lifecycle cost: building is not a one-time cost, and buying is not free of integration and lock-in costs.
Business scenario
A fintech needs (a) a card issuing processor, (b) a fraud detection system, (c) an HR system, and (d) a lending decision engine that is its core product.
Decision framework
| Criterion | Build favoured when | Buy favoured when |
|---|---|---|
| Differentiation | Core to competitive advantage | Commodity / parity capability |
| Fit | No product fits; unique workflows | Products fit 80%+ with configuration |
| Time to market | Can wait | Need it now |
| Total cost of ownership (5 years) | Lower at our scale | Lower including integration |
| Skills | We can hire and retain the team | Scarce skills |
| Risk & compliance | We need full control | Vendor certified (PCI, SOC 2) reduces our burden |
| Lock-in & exit | Strategic dependency unacceptable | Exit cost acceptable, data portable |
TCO for build = development + infrastructure + maintenance (often 15–25% of initial build per year) + operations/on-call + security/compliance + opportunity cost of engineers. TCO for buy = licences/subscription (with growth) + implementation/integration + customization + vendor management + exit cost.
Result for the scenario: (a) partner with a card processor, (b) buy a fraud platform initially with in-house rules on top, (c) buy HR SaaS, (d) build the lending decision engine.
Example
[!FACT] Many large retailers use Shopify, Salesforce or SAP for commerce and CRM while building proprietary capabilities in pricing, supply chain and data, consistent with the build-for-differentiation principle.
CTO interview questions and model answers
The engineering team wants to build an internal tool that exists as SaaS. How do you decide?
Ask what is differentiated, compute 5-year TCO for both including maintenance and opportunity cost, assess fit and lock-in, run a time-boxed proof of concept of the best SaaS option, and record the decision in an ADR with revisit criteria.
When should you revisit a buy decision?
When scale makes licence costs exceed build costs, when the capability becomes differentiating, when the vendor's roadmap diverges, or when lock-in becomes a strategic risk.