Data & Research · Articles
Essays on model-mediated markets.
Our positions, stated plainly — on measurement, independence, and governed execution. Argument, not data; where we cite numbers, we cite method.
Positioning
The referee shouldn’t grade its own homework
The brand-visibility metric is now sold by the same companies that sell you the media, run the campaigns, or bundle the suite it sits inside. That’s not a conspiracy; it’s a conflict. When the party reporting the number also has budget riding on the number, the number stops being measurement and starts being marketing.
Independence isn’t a feature we bolt on — it’s the whole product. We don’t buy your media, we don’t run your agency, and we’re not a tab inside a marketing cloud that sells you seats. That’s what lets a Share-of-Model reading mean the same thing to your CFO as it does to you.
The test is simple: ask whoever gives you a visibility score what else they sell you. If the answer is “the thing this score measures,” you have a scoreboard kept by one of the teams.
Honesty
We don’t claim your revenue. We prove every decision.
It’s tempting to attach a rupee figure to every marketing move. It’s also usually fiction — the attribution chain from an answer-engine mention to a closed sale is too long and too shared to claim honestly. So we don’t claim it.
What we prove instead is the decision. Every suggested action, who owned it, what changed, and the measured lift with its confidence interval — one row at a time, in a decision log. That’s the CFO-grade audit trail: not “marketing drove ₹X,” but “we caught the drop, shipped the fix, and moved the answer by this much, this confidently.”
Honesty here is a commercial asset, not modesty. A number you can defend in a diligence room outlasts a number you have to caveat in a board deck.
Field notes, every so often.
Machine Readable is the short brief for marketing leaders on how AI is representing categories, sources, and brands.