The most expensive mistake is comparing a platform invoice against an engineering salary. Both are incomplete; TCO needs the whole system.
The four costs that tend to disappear
Build
Engineering time to design and build the path.
Run
Maintenance, upgrades, on-call and day-to-day operations.
Risk
Incidents, rollbacks, lock-in and the cost of switching later.
Speed
The value of getting product into users' hands sooner.
Relative cumulative cost by path70%
Q1Q2Q3Q4Q5Q6Q7Q8
How to read it
The cheapest path up front doesn't always win once the platform becomes a permanent responsibility.
An honest decision template
BlockIn-housePlatform
BuildInternal hoursSetup + adaptation
RunOn-call and upgradesFee + internal ownership
RiskConcentrated in the teamContract + lock-in
SpeedDepends on the backlogCan shorten the path
The questions that close out the spreadsheet
Which part of the work would we stop doing if we bought?
Which critical capability do we need to keep in-house?
Which cost shows up in month 13 but not in the demo?
Codifly takeaway
Calculate TCO for the stage you want to reach, not just the size you are today. The right decision leaves a clear exit if your constraints change.
If your first question is cloud cost, start from a concrete baseline: estimate your AWS savings opportunity .