One brain.
Seven brands.
A New York venture firm asked us to build one shared customer brain for their D2C portfolio. Every brand gets deeper insight than they could build alone, and chooses per experiment whether the learning travels to the rest of the fund or stays in house. Rolled in eight weeks.
A portfolio of brands.
No shared brain. No shared learning.
Every brand in the portfolio was building its own view of the customer. Different tools, different definitions, different dashboards. When one brand cracked a retention loop, the others paid to learn it again from scratch.
Founders got siloed reports. Marketers got dashboards that ignored their day-to-day. The fund saw retrospective decks instead of a live picture.
The constraint was not data. It was a central brain the whole portfolio could share, that still felt native inside each brand.
One brain.
Every brand plugs in.
The intelligence sits in the middle. Every brand plugs in their own stack and gets their own tenant on the same shared brain.
- ›Shopify · Recharge · Klaviyo · Segment
- ›Meta · TikTok · Google Ads
- ›Zendesk · Gorgias · Notion
- ›Foundation-model layer
One brain.
Native inside every brand.
Flick through the tenants. Chrome, type, tone and colour switch; the underlying brain stays the same. Every brand opens what feels like a product built for them.
auto-cycling · tap to lock
The marketer, the founder,
and the board see it differently.
Same underlying data. Each role gets it framed for the decision they actually make. Nobody scrolls past somebody else’s view.
Which cohort is worth another dollar today?
Share the learning
or keep it in house.
Every brand chooses per experiment whether the learning stays private or joins the portfolio pool. When a shared experiment lands, the central brain matches the audience shape against every other brand and surfaces the winning play as a suggestion where it’s likely to work.
The data never travels. Only the plays do, and only the ones a brand chose to share.
Central brain. Native feel.
Learning that compounds across the whole portfolio.
“We’ve never had this level of detail about our customers. And we get to choose what travels to the rest of the fund. The plays go, our data stays with us. It has changed how we run growth.”
Cited in AI search.
Bidding on Google.
Yamaha’s global online music school launched in three regions at once. We ran the deep research, did the manual audit legwork, then automated keyword analysis, wired live competitor monitoring, and built the internal tool the marketing team now runs themselves.
Three regions.
Two search systems. One launch.
Parents were searching Google in one register (piano lessons for kids, online music school). Adult learners were asking ChatGPT and Perplexity for the best online music school. The brand needed to show up in both, in three languages, at launch.
A human SEO team can handle one region in one language at a time. The platform reality required all three at speed, and it needed to keep running after we handed the keys back.
The constraint was two things at once: depth of coverage across markets, and a system the in-house team could actually run.
One audit engine.
Every question shoppers ask.
The engine runs continuously against every AI search engine and every paid keyword. Same queries, every day, in every market.
- ›ChatGPT · Claude · Perplexity · Gemini · Google AIO
- ›Google Ads · Google Search Console
- ›Yamaha CMS · Schema layer
- ›AI translation pipeline (en · de · fr · ja)
One audit.
Three locales, three truths.
Flick through the markets. Same audit engine, different query demand, different competitors, different fixes to close each gap.
auto-cycling · tap to lock
Add HowTo schema on lesson-plan pages · adult-learner testimonial block
Organic AI.
Paid Google. One strategy.
The same query universe drives every discipline. Where AI cites Yamaha organically, paid can dial back. Where it doesn’t, paid moves in and the content brief queues for next week.
Where the brand shows up when shoppers ask ChatGPT.
Watch competitors move.
Respond the same day.
The audit engine watches every competitor across every AI engine and every locale. When one of them starts getting cited for a query Yamaha targets, the system flags it as an opportunity or a risk and briefs the fix.
The marketing team opens the tool every morning and sees the overnight movement.
Search plus AI discovery.
Running continuously in every market.
“We used to guess what parents were asking AI about music school. Now we see it live, in every market, and our team runs the system themselves.”
Want work like this on your stack?