Pixel
The system that learns to read your store.
Pixel measures what visitors do, sets it next to every order and finds, every day, the patterns that cost or earn revenue.
Everything that matters, on one screen.
The feed puts what needs attention now at the top. The watchers set every hour against the profile of the same weekday. And behind every insight sits the calculation that carries it.
- AlarmWatchersCheckout conversion deviates since 14:00
Orders this hour sit far below the profile of the same weekday.
Poisson testp < 0.01Open for 2 hours - OpportunityBasketsTop and bottoms in the same print sell together
Candidate for a bundle and a cross-sell in the cart drawer.
Fisher's exactp < 0.01Open for 6 days - OpportunityStockSize M sold out on a running seller
Revenue is left on the table until this size is back.
Sales velocityRangeOpen for 3 days - InfoSearchSearch without results: “linen trousers”
Add a synonym or pass it on to buying.
Zero resultsExits afterwardsOpen for 9 days - RiserRisersThe summer dress speeds up week on week
Give it a more prominent place in the collection.
Poisson ratep < 0.01New today
- 1
A feed that tidies itself
What matters now sits at the top. Whatever no longer holds at the next run disappears on its own.
- 2
Hourly watchers
Every hour is set against the same weekday over recent weeks. A dip shows up the same afternoon.
- 3
Evidence with every insight
Test, threshold and duration sit next to the conclusion. Nothing has to be taken on instinct.
Eight questions every store asks again every day.
Each module answers one question, with a statistical test that decides whether the answer holds. Raw percentages without volume never reach the feed.
- 01
Baskets
Which products are bought together more often than chance explains?
Fisher's exact test with minimum counts - 02
Viewed together
What is viewed together but not bought together: comparison shopping or a missed bundle?
Co-view lift against co-purchase lift - 03
Risers and fallers
Which products speed up or slow down, week on week and year on year?
Poisson rate comparison, weekday-aligned - 04
Watchers
Does today deviate from the same weekday over recent weeks?
Robust z-score on median and MAD - 05
Dead ends
Where do visitors drop off more than their place in the session explains?
Exit rate against expectation per session depth - 06
Stock
How much revenue do running sellers leave on the table through sold-out sizes?
Sales velocity as a range - 07
Gateway products
Which first purchase demonstrably brings customers back?
Proportion test on repeat purchase - 08
Searches without results
What are visitors searching for that the store does not return?
Zero results and exits afterwards
Built on three agreements.
Every number is traceable.
The engine is deterministic: the same data always gives the same answer. A language model only explains, it never invents a figure.
One definition of revenue.
Order total minus refunds, without cancelled and test orders, in store time and over complete days. What the feed calls revenue, the report calls revenue too.
Built for each brand.
Pixel runs as a dedicated app per client, tuned to that brand's theme and measurement points. Not a standard plug-in, but a system that knows your store.
Designed and built by Webfluencer.
Pixel grew out of our work on Shopify stores for leading brands. We build the theme, stamp the measurement points into it and turn every insight into a change we make ourselves.
Your vision, our craftsmanship
We work with ambitious brands to build digital experiences that truly make an impact.