Chris Holland
Before founding Consultus AI, I ran data for the fastest-growing pet insurance company in North America — leading the team from early scale through a private equity exit five years later.
That run taught me what actually moves the needle inside a high-growth business: clean pipelines, honest dashboards, testable decisions, and relentless conversion-rate work. The tools are well understood. What's rare is the discipline to wire them together correctly and the operator judgment to know what's worth measuring in the first place.
Consultus AI is how I bring that playbook to small businesses — the ones that can't hire an enterprise data team but deserve enterprise-quality answers. AI is the force multiplier that makes the economics work. A human (me) is on every engagement, every review, every reply.
One rule: it's not about the product — it's about the outcome. If the numbers didn't move, we didn't earn the fee.
What I've spent years getting right.
A decade of in-house work, concentrated into the handful of things that consistently pay off.
AI Systems
Designing and deploying production AI workflows — not demos — that sit inside a real business and earn their keep.
A/B Testing
Test design, power calculations, and reading results honestly. Distinguishing signal from noise when the stakes are real.
Conversion Rate Optimization
Funnel instrumentation, hypothesis generation, and the iterative grind that actually compounds CVR over quarters.
Data Architecture
Warehouse design, ingestion patterns, and the boring infrastructure decisions that determine whether your analytics works in year three.
Dashboards
Dashboards that answer questions instead of asking them. Axes scaled to where the business lives. Nothing on screen that isn't earning its place.
Data Modeling
Clean semantic layers. Reusable transformations. Models a new hire can understand on their second day.
A few things I care about.
Own it end-to-end
If a number on your dashboard is wrong, that's my problem. I don't hand off. I don't disappear. I ship fixes, not finger-pointing.
Work in public
Every transform is in your GitHub. Every dashboard is in your Metabase. No hidden models. No proprietary layer you'd lose if I walked away.
Explain the reasoning
If I'm making a modeling choice, you hear why. If an A/B result is ambiguous, I say so. Good data work is mostly being honest about what you don't know.