Dave Carruthers

Honest answer: because Hotspex was a profitable service business with smart people delivering smart engagements. The technology to deliver an emotional intelligence API at scale — LLMs that can interpret nuanced consumer language, MCP that lets AI agents call external context, vector embeddings that make 25 years of qualitative data queryable — none of it existed in a usable form until 18 months ago.

The unlock is now. AI agents need contextual frameworks, and HMF is the only validated emotional one. The full RBC ad-testing platform was built in 18 hours with Claude Code. The friction has collapsed. The window opened. We are moving.

The IP isn't the categorisation scheme. The IP is the validation against outcomes. 25 years of correlating emotional response data against actual brand performance — share, penetration, growth, churn — across 800 brands and 30 countries. That dataset doesn't exist on the open web. It can't be scraped. It exists in our warehouse, tied to NDA'd brand performance data we got because we ran the studies.

A frontier model can produce a plausible-sounding emotional taxonomy. It can't tell you which zone of the map predicts CPG growth in markets with declining category demand. We can. That's the moat. Anyone can build a map. Only one company has 25 years of map-versus-outcome data.

Mark made this point on stage at MM&E: AI commoditises the production layer. The validation layer — the empirical proof that this signal correlates to that outcome — is what becomes more valuable, not less.

Performance-based grant. 3–8% range depending on the level of ownership we agree on. Vesting tied to specific commercial milestones — BCG signed, Adobe POC live, Tracksuit live, ARR thresholds at year-1, year-2, year-3. Same structure used for the other operator equity grants at the company.

Cap table: clean. The media subsidiary is on a separate cap table — the June 2026 sale doesn't dilute Hotspex Insights. You join a debt-free company entering a platform-build phase with cash to fund it.

Exact percentage and milestone schedule is a second-conversation detail. I'm not playing games with that — it's just that the right number depends on what you want to own (NA only, or AI strategy too) and what milestones you're comfortable signing up for.

I'm going to be precise here because you'd see through anything else.

BCG: Head of research practice attended the MM&E keynote in person. Active scoping conversation about HMF integration into their brand consulting practice. No signed LOI. Signed scope of work is the next step — that's what the President closes.

Adobe: Verbal interest from a senior product person via warm intro. The HMF API specification is being drafted now. Next milestone is technical integration POC in their Brand Brain layer.

Tracksuit: Conversation with their leadership about HMF as the emotional layer on their dashboards. Earlier than BCG. They've expressed clear interest; commercial structure is unscoped.

System1: Earliest of the four. Real. Not yet structured.

So: warm interest, scoping, and one keynote-validated relationship. Not signed deals. That gap is the job.

First 30 days: we run the existing NA client book together so you understand the delivery playbook and the math. You meet the BCG, Adobe, and Tracksuit relationships personally. You inherit the platform team and get briefed on the MCP server build state.

Days 30–90: three KPIs. (1) Lock the BCG scope of work. (2) Move Adobe from interest to scoped technical POC. (3) Hire the NA commercial team — your VP Sales, two BDRs, a marketing lead. We have budget for those hires the day you start.

By month 6: Tracksuit live, System1 scoped, the HMF MCP server is in production with at least one paying enterprise integration.

I'm not asking you to figure out the strategy. The strategy is on this page. I'm asking you to execute it.

Voxpopme is video research. Hotspex is emotional motivation methodology + an emerging platform layer. Different categories, different buyers. We've already had outside counsel review and they don't see an issue. If your specific non-compete language is more restrictive than the market norm, we'll review it together — but I haven't seen a Voxpopme non-compete that would block this.

Founder tax residency reasons that don't bind your domicile. Park City stays Park City. NA = the entire continent. We expect Toronto presence for monthly leadership reviews and quarterly board meetings — call it 4–5 days a month max. We'll set up a flexible NYC presence as the partnership pipeline matures, because Adobe / BCG / System1 work pulls you there anyway.

You should ask people who've worked with me. I'll send you references — including operators I've made wealthy and operators I've parted ways with. Both are useful data points.

The honest version: I move fast, I'm direct, I don't waste time on process theatre. I expect the same back. I will not micromanage NA — I literally can't, I'm building the platform side and running global. What I will do is push hard on goals and back you completely on how you hit them. If we don't agree on a goal, we argue until one of us changes their mind. Once we agree, I'm out of your way.

If "you let me run it" means "no one challenges my decisions" — that won't work here. If it means "operational autonomy with goal alignment and no second-guessing" — that's exactly the model.

The multiples are the comp range from public insights/SaaS deals: Qualtrics, Quantum Metric, Numerator, Glassbox, Brandwatch acquisitions. The 1.5× floor is service-firm reality. The 6–8× ceiling is what platform-licensing companies command when they have proprietary IP and recurring contracts.

The inputs that determine where we land: (1) recurring licensing as % of revenue, (2) named platform partners shipping in production, (3) annual contract value vs. project revenue. Those are the variables. The President of NA owns most of them.

So: multiples are real. Whether we hit the 7.5× scenario depends almost entirely on whether the partnerships ship. Hence the equity structure is performance-based — the upside scales with whether you actually deliver them.

Because if I sent this through a recruiter, you'd never see the version of it that I actually wanted you to see. The recruiter would summarise it into a generic JD and you'd skim it on your phone and pass.

I built this page specifically for you. The kind of person who notices that — who reads the parallels section and thinks "wait, he actually mapped this against my career" — is exactly the kind of person I want running North America. It's not a gimmick. It's the screening test.