Notes
Things we learned the expensive way.
Written from work already done rather than from a content calendar. Each one answers a question a firm owner or their technical evaluator has actually asked us.
01
Your cost-per-case benchmark is probably wrong
Not because the arithmetic is bad. Because of what went into it, and what quietly did not.
Read it →02
The pagination bug that drops your oldest records
It passes the type checker. It passes the build. Nothing in the interface looks wrong. Only counting the answer catches it.
Read it →03
One writer per rail
When two systems can write the same record and the platform gives you nothing to deduplicate on, agreement between teams is the only control you have.
Read it →04
What it costs to migrate ten years of history
Every proposal underestimates this, including ours once. The cost is not in moving the records. It is in deciding what they meant.
Read it →05
Fixed fees: the risk is the term
The instinct when a fixed-fee matter loses money is to raise the fee. In the data we have looked at, that is usually the wrong lever.
Read it →