I walked a seasoned consultant through mapMyCulture last week. Decades in the field. Sharp, direct — not the kind of person who softens feedback to be polite.
His first reaction: "That's a high acquisition cost just for scraping Glassdoor."
Fair. A lot of tools stop there.
One Source Isn't a Diagnostic
Public review data is a strong signal, but on its own it's incomplete. It captures the people motivated enough to post — and misses everyone who never does.
So I explained that we also integrate an internal survey layer: anonymous campaigns an organization runs alongside the public data. Two sources, unified in a single dashboard.
He sat with that for a moment. Then: "Okay. That's much more interesting."
That shift is the whole point. Public reviews tell you how your culture looks from the outside. Anonymous internal surveys tell you how it feels from the inside. Put them in one view and the gaps between the two become the most useful thing on the screen.
The Question That Stayed With Me
I flew out of that call feeling good. But he wasn't done.
At the end, he said something I haven't stopped thinking about: "My network operates across Europe, Latin America, Asia. If the AI is primarily trained on American data — and most AI is — then the analysis is going to be biased toward American cultural norms. That's a real problem for international use."
He's right. And I didn't have a clean answer.
Same Dysfunction, Different Signal
Here's why it matters. A disengaged employee in Antwerp writes a measured, professional review. The same situation in Texas generates three paragraphs of frustration.
Same underlying dysfunction. Completely different signal on the page.
An AI trained mostly on English-language, US-centric data might read those two very differently — flagging the Texan as a crisis and the Belgian as fine, when the reality underneath is identical. Get that wrong and the whole diagnostic tilts.
Language Support Is Not Cultural Calibration
We support four languages. But language support and cultural calibration are not the same thing. Translating the words is easy. Reading them through the right cultural lens — knowing that restraint in one place carries the same weight as open frustration in another — is the hard part.
This is an open validation question I'm taking seriously. Because a diagnostic that misreads cultural nuance isn't just imprecise. It's potentially misleading — and a misleading culture map is worse than no map at all.
Reading culture honestly means reading it in context — combining public and internal signal, and calibrating for where people actually come from. That's the work mapMyCulture is built around: two data sources, one clear picture, interpreted through the right lens.