How do you make build-vs-buy decisions?

Question

Accepted Answer

**Build vs. buy** asks whether to develop something in-house or adopt an existing product or service. The instinct to build is strong and often wrong, the real cost of building includes years of maintenance, not just the first version. The decision hinges on whether it's *core* to your business.

## A decision framework

```text
LEAN BUY when:                     LEAN BUILD when:
- It's NOT your core differentiator  - It IS your core value/differentiator
- A mature solution exists           - No good solution fits your needs
- Speed-to-market matters            - You need deep control/customization
- You lack expertise to maintain it  - Long-term cost of buying is prohibitive
- TCO of buying < building+upkeep     - Vendor lock-in is an unacceptable risk
```

## A concrete example

Don't build your own auth system, identity is hard, security-critical, and not your differentiator; buy/integrate a proven provider. *Do* build your core recommendation engine if recommendations are what make your product special. Build what makes you *you*; buy the rest.

## Count the true cost

Teams chronically underestimate building: the first version is 20% of the lifetime cost. Maintenance, security patches, edge cases, and on-call dwarf it. Count total cost of ownership over years, not the demo you can ship in a sprint.

## A pitfall

"Not invented here" leads to rebuilding solved problems; blind buying leads to vendor lock-in and tools that don't fit. Decide deliberately, per case.

## Why it matters

Building the wrong thing in-house drains years of engineering effort into undifferentiated work.

Getting this right concentrates your scarce engineering time on what actually makes the business win.