Satire
The Tech Stack That Could Not Be Agreed On
The selection criteria document had 23 criteria. Every vendor scored 6.7.
The organisation needs to standardise on a database. A committee is formed. An evaluation framework is produced. The framework has 23 weighted criteria.
Three vendors are shortlisted.
What Actually Happens
Oracle, because it is already here. MongoDB, because the developers want it. PostgreSQL, because someone read an article.
Each vendor is scored by different evaluators using different interpretations of the same criteria. Oracle scores highest on "enterprise support" and "compliance readiness." MongoDB scores highest on "developer experience" and "schema flexibility." PostgreSQL scores highest on nothing but nobody wants to say that out loud because the person who suggested it is still in the room.
The final scores are within 0.4 of each other across all 23 criteria. The committee cannot reach consensus. A fourth vendor is suggested. The evaluation framework is revised to include two new criteria that happen to favour the fourth vendor. The original shortlist is reopened. Six weeks pass.
The organisation continues using Oracle. The working group is disbanded. The Confluence page is archived.
“I've modelled this as a distributed consensus problem. We need a Raft-based voting algorithm with a quorum of stakeholders. Current quorum is unavailable.”
DivineForge Advisory
Recognise this pattern in your organisation? I help teams cut through the governance, make the right technical calls, and actually ship.
Let's talk architecture →How a Builder Should Respond
You agree on the primary constraint before building an evaluation matrix — not 23 criteria, but one question: what is the single most important thing this database must do well for this specific workload? Everything else is secondary. You run a time-boxed proof of concept against each shortlisted option against that one constraint, with real data and real query patterns, and you let the result make the decision. A selection process that produces equal scores has not produced a framework — it has produced a draw. The tie-breaker is always operational: which one can your team actually run at 2 AM when something breaks.
20 articles about enterprise dysfunction. None of it billable. Buy me a coffee.
Buy me a coffee