$27M
Modeled 2027 ARR, up from ~$2.7M in 2025. Projection, never launched.
CASE STUDY · AI + STRATEGY · PRE-PRODUCT
The company plan came before the product.
A pre-product AI personalization platform, given a business before it ever shipped.
There was a real idea and no company around it: AI personalization for Shopify Plus. I built what a pre-product idea actually needs before code. The category, the wedge, the pricing, the beta plan, the competitor map, and a modeled ARR curve from about $2.7M to $27M. Then I left before launch. So read this as strategy and category IP, not a track record.
Everything on this page is a modeled projection. The product never launched.
Read this first
What this page is, and is not.
- Every number here is a modeled projection I built, not a measured result.
- The product did not launch. I left the company before it went live.
- What is real: the strategy, the category, the pricing architecture, and the model itself.
- Treat the ARR curve as an untested financial scenario.
The problem
A real idea with no business around it.
The technology worked. AI personalization for Shopify Plus stores, the kind of engine that reshapes a storefront around who is shopping. Good idea. But an engine is not a company. There was no category to sit in, no wedge to lead with, no pricing anyone had sanity-checked, and no story an investor could underwrite.
That is the gap I was brought in to close. Not to write model weights. To build the thing an idea needs before it can raise money or sign a logo: a reason to exist that a board can nod at.
The wedge
Personalize the first visit, before you know the buyer.
Some personalization depends on browsing or purchase history that a first-time visitor has not supplied. The proposal focused on that cold-start problem. It was a commercial hypothesis, not a measured conversion result.
So I pointed the whole company at session zero. Personalize the first visit using signals present before any account exists: source, intent, device, context. Win the buyer on the visit you are already paying for.
The category was never better personalization. It was cold-start conversion, won before the second session ever happens.
The positioning hypothesis was more revenue from existing traffic by improving the first visit. The beta plan was meant to test whether the proposed product could deliver it.
What I built
A board-ready plan, four parts.
None of this was product work. It was the commercial architecture an idea needs to become a fundable company.
GO-TO-MARKET
The roadmap
The path from zero to first revenue on Shopify Plus. Who to sell, in what order, with what motion, and the sequencing that gets a pre-product company its first reference logos.
PRICING
The architecture
How the product captures value. Packaging and tiers tied to the outcome the buyer actually feels, cold-start conversion lift, not a flat seat fee that ignores the win.
BETA + MAP
Proof and position
A design-partner beta to test the proposed lift, alongside a competitor map to examine whether first-visit personalization offered a useful distinction.
ARR MODEL
The scenario
The full ARR model behind the raise. Assumptions, drivers, and a curve a board can interrogate line by line, from about $2.7M to $27M.
The model
The ARR curve, from ~$2.7M to ~$27M.
This is the centerpiece, and it is a projection. It is the scenario I modeled if the plan executes: session-zero conversion lift compounding into new logos and expansion across the Shopify Plus base. A board can pull every assumption apart. That is the point of building it.
Modeled ARR · 2025 → 2027
~$2.7M → ~$27M projected. Modeled scenario, product did not launch.
2025 ~$2.7M · 2027 ~$27M. Modeled projection, never launched. Not a measured result.
First-visit conversion
2.5% → 3.8% modeled lift from session-zero personalization.
Modeled projection, product did not launch.
Average order value
$75 → $95 modeled, alongside a 30% bounce reduction.
Modeled projection, product did not launch.
The model · by the number
The drivers, tagged for what they are.
Four numbers carried the story into the room. Each one is modeled. None of them happened.
$30M
Added annual revenue modeled for one major department store, at the top end. Modeled, not measured.
30%
Modeled reduction in first-visit bounce from session-zero personalization. Projection only.
52%
Modeled first-visit conversion lift, 2.5% to 3.8%. Projection, product did not launch.
The point
My deliverables were the positioning, pricing, design-partner plan, competitor research and financial model. I left before launch, so this case cannot establish adoption or conversion lift.
The plan was ready to test. The projected results were still unproven.