Options enter a matrix with weighted criteria and justified scores. And each one declares what comparison charts leave off the page: the cost, the timeline and the path to leave within 18 months.
The problem
The platform comparison comes from the sales material of one of the candidates, or from a generic benchmark that knows neither your volume nor your team. The cost of leaving appears in neither, and it is what turns a choice into lock-in.
The matrix scores the options with your criteria and your weights, with the justification of each score in sight. The exit-cost column declares reversibility, timeline and what exactly locks you in; an option without that answer shows up as a gap instead of disappearing.
How it works
Every criterion has a declared weight and every score has its reasoning written down. You can disagree with a weight, and that is the point: the discussion becomes about the criterion, and stops being about the intuition of whoever spoke loudest.
The recommendation ships with its flip condition declared: in this scenario, this option; if volume crosses the threshold, the comparison flips. That condition becomes a watched assumption.
Where the analytics layer should live
4 weighted criteria · 40 GB active · team with no dedicated engineer
| Option | Score | Monthly cost | Cost to leave within 18 months |
|---|---|---|---|
| Analytical Postgres on the database you already run | 8.4 | R$ 4,400 | Low Standard SQL, dump and restore. No proprietary format. |
| BigQuery + BI serverless | 7.9 | R$ 2,600 | Medium Export to Parquet handles the data. Vendor-specific SQL and scheduled routines have to be rewritten. |
| Managed lakehouse full platform | 5.1 | R$ 11,900 | High Catalog, permissions and notebooks have no equivalent elsewhere. Estimated 4 to 6 months. |
Recommendation: BigQuery + BI at this volume. Above 1 TB active, the comparison flips.
The monthly cost of each option is computed per tool, with the source of each price in the table itself, and the 12- and 24-month TCO comes out by arithmetic.
With no vendor commission, the matrix is free to recommend the cheap option when the scenario calls for the cheap option. That is what makes it possible to recommend the Postgres you already run.
Analytics layer: monthly costs
BigQuery + BI scenario · source declared per price
| Item | Monthly cost | Source |
|---|---|---|
| Storage and querying | R$ 1,200 | vendor website |
| BI tool, 25 users | R$ 1,050 | vendor website |
| Light orchestration | R$ 350 | estimate with the client |
| Monthly total | R$ 2,600 |
TCO: R$ 37,800 over 12 months · R$ 69,000 over 24 months (R$ 6,600 setup + recurring)
The TCO is the renderer’s arithmetic over the table, never a number the model wrote from memory.
Under the hood
The rules that make the comparison worth anything.
What comes next
12- and 24-month TCO computed by arithmetic, a declared source per price, and the deviation checked later against the real invoice.
See it from the inside →AdversarialObjections consideredAn adversarial reviewer attacks through cost, risk and simplicity; answers get verdicts, and an objection that stands is kept in the document.
See it from the inside →MemoryMemory and assumptionsFacts, decisions and assumptions outlive the conversation, and each assumption is watched with the condition that would invalidate it.
See it from the inside →Paste the structure, get the most serious findings in about two minutes, and decide later whether an engagement is worth opening. No account, no card.