July 18, 2026 · 5 min read
Fixed-Price vs Retainer, What I Learned Pricing MarTech Engagements
I priced the same kind of MarTech engagement both ways across different clients. The lessons about risk, incentives, and who actually benefits surprised me.
I spent a stretch quoting nearly every engagement fixed-price, because it felt more professional — a defined deliverable, a defined number, no ambiguity. After enough projects where the fixed price turned out wrong in both directions, I started paying closer attention to what each pricing model actually incentivizes.
Fixed-price puts all the estimation risk on me. If a GTM migration takes twice as long as scoped because the client's existing setup is a disaster nobody mentioned in discovery, that's my problem to absorb, not theirs. I've eaten weeks of unpaid work this way more than once. The upside is clients love it — a number they can budget against with no surprises.
What retainers actually change
A retainer shifts the risk the other direction and changes the incentive entirely. Instead of racing to finish a fixed scope, the incentive becomes sustained usefulness — which, for the kind of ongoing tag governance and measurement work I do most often, is actually the more honest match to the work itself. Nobody needs their GA4 setup 'finished.' They need it maintained, monitored, and adjusted as the business changes, indefinitely.
Fixed-price fits work with a real finish line. Retainer fits work that was never going to have one, no matter how the SOW is worded.
How I split it now
- Fixed-price for anything with a genuinely discrete deliverable — a migration, an audit, a specific build
- Retainer for anything ongoing by nature — tag governance, measurement maintenance, monthly reporting infrastructure
- A fixed-price discovery phase before any retainer, specifically to kill the estimation-risk problem before locking into a recurring number
- Never fixed-price work where the client's existing environment is unknown and unaudited — that's exactly the scenario where fixed-price punishes me for someone else's technical debt
The real lesson wasn't that one model is better. It's that picking the model based on the actual shape of the work, rather than on which one sounds more professional, prevents most of the pricing regret I used to deal with project after project.