How Do You Stack Up?

You have a decision to make and four options that all demo well. Everyone in the room has an opinion, the loudest one usually wins, and six months later somebody is quietly paying to migrate off whatever got chosen. The cost of the decision turns out to be much larger than the cost of the thing.

The instinct is to search, and search now returns the most-marketed answer rather than the best one. Ask an AI instead and you get something worse in a specific way: a confident recommendation with no provenance, no alternatives considered, and no way to tell whether it is describing a live project or one abandoned in 2021.

The alternative is to do the research before you need it. A library of surveys, each one covering a domain — the real alternatives, compared, with what it costs to leave again and an honest read on whether the project will still be maintained in three years. Built once, consulted in the moment.

So we do it live, with a real decision from the room. Not a demo of a tool — an actual choice somebody in the audience is facing, worked through on stage against the library, arriving at an answer they can take back to work. It is the part of the talk I cannot rehearse, which is also why it is the part worth watching.

Audiences leave with the method rather than the library: how to tell a researched recommendation from a confident one, which questions actually separate two tools that look identical on their landing pages, why exit cost belongs in the decision and almost never appears in it, and how to spot the answer that was optimized for being findable rather than for being right.