Monitor prices across retailers

This tutorial follows the Market Radar workspace used throughout these docs. It starts from a spreadsheet of products and ends with a set of scheduled pipelines, a canonical product catalogue matched across twelve retailers, and a dashboard on top. You will build a smaller version, then grow it.

You need

A product list (CSV or sheet)

Sources

2 to 12 retailer sites

Time

About 20 minutes plus the build

Result

Daily prices, promotions and stock per retailer

1. Catalogueyour product list
2. Discover pagesfind each product on each retailer
3. Match productsone canonical product across sites
4. Track pricesdaily price, promo, stock
5. Analysedashboards and alerts

1. Start from your catalogue

1. Prepare a product list. One row per product with whatever identifiers you have: SKU, brand, name, pack size, and any known retailer URLs. A dozen products is enough to start.

2. Describe the job and attach the file. On the setup page, write the brief and attach the catalogue.

   Track the price, promotion and availability of the products in the attached catalogue across Amazon, Walmart, Best Buy and Target, every morning. Match the same product across retailers so I can compare them in one row per product per retailer. Keep the product page link, the list price, the current price, any promotion text, and whether it is in stock. Flag price drops of 10% or more.

3. Review the proposal. Jason proposes the fields, the retailers it can reach, the matching rule (identifier first, then brand plus name plus pack size) and a daily schedule. Refine here if a retailer is missing or a field is wrong.
4. Build. Jason explores each retailer, builds the pipelines, runs them once on a sample of your catalogue and sends the result for review. Expect this to take a while for four sites; you will get a notification when it is done.

2. Understand what Jason built

Open Pipelines. Rather than one large pipeline, Jason split the job into stages that depend on each other. That is deliberate: a retailer changing its page layout only breaks the stage that reads it.

Pipeline	What it does	Reads	Publishes
Catalogue discovery	Finds the product page for each catalogue item on each retailer	your catalogue	product pages
Product identity matching	Decides which pages are the same product	product pages	product matches
Price & promotion tracking	Collects list price, current price, promotion text	product matches	price observations
Inventory monitoring	Collects stock state	product matches	inventory signals
Change detection	Compares today with yesterday and flags meaningful moves	observations	daily rollups
Retail Signal Fusion	Joins everything into one product-level view	all of the above	live product offers

Open Price & promotion tracking and read its statement of work. Note the parameters strip: the drop threshold you asked for is a parameter, so you can change it without asking Jason.

3. Check the first rows

Open the Live product offers dataset. Each row is one product with a column per retailer link, the current and previous price, availability and a confidence score for the match.

Sort by confidence ascending to review the weakest matches first. If a match is wrong, tell Jason:

The Galaxy Watch 7 match on Best Buy is the 44mm model; we track the 40mm. Match on the model number in the title, not the name alone.

Jason updates the matching pipeline as a new version and re-runs it.

4. Let it run, then look back

The pipelines are scheduled at staggered times so each stage runs after the one it depends on. After a few days, open Latest runs to see the rhythm, and open any run to see rows, duration and pages fetched.

5. Put a dashboard on top

1. Open Analytics and press New Dashboard.

2. Describe the dashboard and link the data. Type @ and pick Live product offers and Price observations, then say what you want to see:

   Products tracked, price changes over the last 30 days, promotions live today, in-stock coverage by retailer, and a table of today’s biggest price moves. Add Retailer and Category filters.

3. Press Start building. A conversation with Jason opens. He writes and checks a query for each card, then draws the dashboard. It shows as a blue Building card in All Analytics until it is ready.

The same dashboard appears in the Analytics section of each pipeline whose datasets it reads. To rearrange cards, change one or add a filter, press Edit dashboard; see Editing dashboards.

Click any bar or point to ask Jason about that value, for example why a retailer’s coverage dropped on one day.

6. Grow it

• More retailers. “Add Costco and Home Depot with the same fields.” Jason extends discovery and matching; the other pipelines pick up the new pages automatically.
• More products. Upload a new catalogue version to the input dataset, or append rows. The next run picks them up.
• Regional prices. Add a parameter for region or store, and run the pipeline once per value.
• Alerts. The next tutorial sends a Slack message when a price drops. Alert rules can also send the rows that matched to a Google Sheet or a warehouse.