Build vs Buy vs Rent Decision Framework
Stop deciding build-vs-buy on gut feel. Run the NPV.
Finance · 60 sec
●●●
NPV = −I₀ + Σ (CFₜ ÷ (1 + WACC)ᵗ) · compared across Build / Buy / Rent
- Inputs
- Upfront $ · Annual run-rate $ · WACC · Horizon (yrs) · Switching cost
- Output
- NPV per option + qualitative scorecard
- When to use
- When the AI/vendor debate is stuck on opinions, not numbers.
- Replaces
- Two-option Build-vs-Buy decks that ignore Rent entirely.
Every product org is being asked the same question this year: do we Build it ourselves, Buy a pre-built solution, or Rent a SaaS / API? There are three options — not two — and the honest answer depends on five inputs discounted to today's dollars at your WACC.
The five factors
Strategic differentiation (does this become a moat?). Time-to-market delta (months). Total cost of ownership over 36 months. Switching cost if the vendor disappears. Internal expertise depth.
Worked example at 12% WACC
Build (in-house): $200K upfront + $50K/yr × 5 → NPV ≈ −$380K. Buy (pre-built): $150K upfront + $65K/yr × 5 → NPV ≈ −$376K. Rent (SaaS): $0 upfront + $80K/yr × 5 → NPV ≈ −$288K. Rent wins on cash, Build wins on control — the framework forces you to price the trade.
The trap
Teams routinely under-estimate the maintenance tail. A model you trained costs roughly 1.6x its initial build cost over the following two years just to keep pace.