Crux

AI-first ecommerce analytics for scaling Shopify brands

Crux builds ecommerce analytics around how your business runs, validated every night, then installs a proven growth system that tells you where to focus next.

How it fits together● Live
YOUR PLATFORMS YOUR CONTEXT VALIDATED AND OPTIMISED FOR AITransformedraw to cleanModelledaround youValidatedevery nightDefined onceCAC, new CVR AI analystin Claude Dashboardswhere to look Growth skillsone way to use AI
  • £22M

    Fussy, £2.2M to £22M GMV in three and a half years, using these growth fundamentals.

  • 1

    Version of revenue, CAC, new customer conversion rate, CM1, CM2, CM3 and the rest.

    See 01 
  • Up to 5x

    Better experiment success rate, because every test feeds learnings into the system.

    See 02 
  • £0

    To take the warehouse, models and Google Cloud project with you if you leave.

    See 03 

Fussy: the founders' operator record. Up to 5x: brands on the Crux growth system.

Brands we've worked with

  • Fussy
  • Seep
  • Two Dudes
  • S3 Sensitivity
  • Kate Barton Design
  • Rover
  • The North Face
  • Augmentum Media

Built by operators. Fussy, £2.2M to £22M GMV. DogBuddy, acquired by Rover.

01

Your AI-optimised data foundation

Better data in, better answers out, from every AI you use

Your AI is only as good as what it runs on. We bring your numbers and your customers' words into one foundation, validated, modelled around your business and with every metric defined once, so your AI agents answer with what happened and why.

Fig. 1 · Your AI analyst, in Claude
AI analyst · in Claude● Live
Which channel brought in our best new customers last quarter?

Paid Search. Its new customers spent £118 on average in their first 90 days, against £74 from Meta.

Paid Search£118
Email£96
Organic£88
Meta£74
TikTok£51

Tables used: orders, customers, attribution

Illustrative: example questions and figures.

  1. Built for AI from the ground up: every table and metric is defined and described, so the AI analyst gets the right answer first time, on fewer tokens.

    Data definitions
  2. Modelled from 15+ years of ecommerce operator experience. Each number comes from the platform best placed to know it. Shopify for revenue and orders, GA4 for visits and behaviour, the ad platforms for spend, all tied back to real Shopify orders. The result is one simple foundation your team and AI can both read, validated every night.

    Ecommerce data model
  3. The numbers say what happened; your customers' words say why. We bring reviews, survey answers and cancellation reasons in beside the numbers, so the AI analyst can answer both questions.

    Context layer
  4. Ask in plain English and get the answer in seconds. Your AI agent plugs directly into an MCP connector we build for you.

    AI analyst
  5. You can check any number. Every model and test is readable and traced to its source.

    Open models
02

Clarity in the chaos

Know what's driving growth, and where to invest next

Ten dashboards, three versions of revenue and a different answer from every platform. We replace the noise with one shared language for growth. Every channel, initiative and experiment is measured in one place, so you know what's working, what isn't, and where your resources should go.

Explore the growth system →
Fig. 2 · KPI health, then pacing
Spend
£7,428▲ 17%
v the last week (£6,355)
CPUStrong
£0.66▼ 13%
v the last week (£0.76)
Session to engagement %Investigate
35%▼ 3.8%
v the last week (36%)
Engagement to checkout %Watch
15%▲ 0.6%
v the last week (15%)

Illustrative figures, drawn the way the dashboard suite shows them.

  1. However your product costs, fulfilment fees and margins work, we model them in, set in a sheet your team can change without code. We build data models bespoke to how your business runs, not forced into a legacy platform template.

    Bespoke build
  2. No single source gives you the whole truth, so we attribute every order (opens in a new tab) through a six-source waterfall, then re-attribute the orders left as Direct or Unknown to what actually drove them.

    Attribution
  3. Every KPI is triaged and given a status against its own 12-week history, so you know straight away what's on and off track. And you track how you're pacing against budget, daily, down to CM3.

    KPI health
  4. Your team learns faster: brands on the growth system improve their experiment success rate by up to 5x, because every test builds on the decisions, learnings and customer evidence logged before it.

    Experiment log
  5. One way to use AI across the whole team: we set everyone up with the growth skills and workflows we use.

    Growth skills
03

Yours to keep

You own it and run it

No renting a platform with a black box. We build your data and AI infrastructure to best practice, so when you're ready to hire in-house or take the next step, everything stays with you.

Book a call →
Fig. 3 · What you take with you

Your Google Cloud projectYours from day one

  • Warehouse
  • Models
  • Tests
  • Definitions
  • Dashboards
  • AI analyst

We host and run it while we work together. If you leave, all of it transfers to you, at no charge.

  1. The models we build for you are yours from the day we write them.

    Ownership
  2. We host and run your warehouse while we work together, at our cost.

    Hosting
  3. If you leave, for any reason, the warehouse, the models and the Google Cloud project transfer to you at no charge.

    Free exit
  4. Half the build fee on signing, the other half only when you confirm in writing that the build met the scope.

    Fee on sign-off
04

Next

Start with a call

We go through your reporting before we price anything.