Page one
What the scraper saw this morning for each tracked keyword: every listing's rank, price and promo badge, compared with yesterday.
Tracked listings
Click a listing for its 14-day price and rank history.
Share of search · top 10
How many of the top-10 results our brand holds, per keyword.
How it worked
The production system this demo rebuilds ran every morning for sales teams across 6 countries. Two things about it were unusual.
It was an email app, on purpose
We wanted to build fast and fail fast, but internal IT rules made standing up a proper internal web service slow. Email needed no approvals, so email became the interface, in both directions.
↺ …and the next morning's briefing reflected what you replied. No servers to approve, no UI to host: the mail client everyone already had was the app.
Automation was earned step by step
Each stage shadowed what salespeople already did before automating it.
- shippedAutomate the lookupCollect the data reps were searching for by hand, every week, and have it assemble itself overnight.
- shippedAutomate the noticingCatalogue the situations a rep would spot in that data (a quiet price cut, a vanished listing) and detect them automatically.
- shippedAutomate the playbookCollect how reps reacted to each situation, and attach those reactions to every insight as ready-made suggestions.
- where it was headingLearn from the repliesTrack how reps interacted with the briefings, to find which reactions could safely become automatic decisions.
The daily mechanics
- ScrapeThe search pages for every tracked keyword, every morning: rank, price, promo badge, rating.
- IngestOur own sales and margin data, refreshed daily and weekly.
- MapOur products to rivals that share keywords inside a ±40% price band.
- DecideA decision tree turns overnight changes into typed insights, each with a priority and the weekly sales it puts at stake.
- BriefOne email, ranked so the most expensive signal is always on top.