The four classifications
For every (product × property) pair where there's evidence from both brand-controlled sources and independent market sources, Theia computes:
| Brand says | Market says | Classification | Action |
|---|---|---|---|
| Yes (high mentions) | Positive | VALIDATES | Keep — proof of message-market fit |
| Yes (high mentions) | Negative | OVERSELLS | Stop — claim is being rejected |
| Yes (high mentions) | Silent | IGNORED | Test — message isn't landing |
| Silent | Positive | UNTAPPED | Activate — market loves something you're not saying |
The most strategically valuable cell is UNTAPPED. It's where the brand is leaving free attention on the table.
How "brand-controlled" is defined
For each source, Theia tags whether it's brand-influenced:
- Brand-owned YouTube channels (filtered by channel name)
- Brand PDPs and product pages
- Press releases distributed by brand PR
- Paid placements (where detectable)
Everything else — editorial, independent YouTube, Reddit, forums, retailer reviews — is market voice.
Two parallel aggregations are then run: brand sentiment and market sentiment per property. The four-quadrant classification is computed from the difference.
Example: Canon EOS R8
| Property | Brand mentions | Market mentions | Classification |
|---|---|---|---|
| Autofocus | 24 | 187 (positive) | VALIDATES |
| Video | 18 | 142 (positive) | VALIDATES |
| Display | 12 | 8 (mixed) | OVERSELLS |
| Ease of use | 9 | 6 (mixed) | OVERSELLS |
| Portability | 1 | 89 (positive) | UNTAPPED |
| Value for money | 0 | 76 (positive) | UNTAPPED |
| Ergonomics | 0 | 54 (positive) | UNTAPPED |
| Image stabilisation | 0 | 31 (negative) | IGNORED (and probably correctly — every rival has IBIS) |
The R8 marketing was pushing Display and Ease of Use (Canon's chosen angles) when the market was praising Portability, Value, and Ergonomics — and was negative on IBIS. A complete misalignment between brand message and market validation.
Why this matters
Three reasons brand teams find this analysis uncomfortable but invaluable:
01 — It exposes wasted messaging spend. Every OVERSELLS row is budget being spent on claims the market is rejecting.
02 — It surfaces free attention. Every UNTAPPED row is content that already has consumer momentum — the brand just needs to amplify, not invent.
03 — It defends the messages that work. VALIDATES rows give the marketing director the evidence they need to keep funding the messages that are working.
Persisted at scale
For the Canon EU deployment, brand_gaps has 30K+ rows: every (product × property) combination where both brand and market have sufficient mention evidence.
L3 Situation Analysis reads this table directly. L4 Content Generation uses UNTAPPED rows to seed new content briefs. The MCP server exposes it for ad-hoc queries.
Related concepts
Related terms