How Store Insights works
Store Insights reads your own store data and suggests campaigns worth running, with the exact numbers behind each one. This page explains what it looks at, how each recommendation is decided, and why some insights take time to unlock.
The short version: Store Insights looks at your store profile, your catalogue, and the sales your store makes after you install the app, then suggests campaigns worth running. Every suggestion can show the exact numbers that triggered it, tucked behind a "Why this?" link, and nothing publishes on its own. It uses your own store data only, never customer information. The thresholds below are current defaults that we tune over time, not fixed promises.
The three kinds of data we use
Recommendations are built from three layers, all of them your own store data. None of them touch customer information.
Store profile
Your Shopify plan, country, currency, product count, and product types. This is read fresh every time you open the page, and it decides which recommendations make sense for you. A store with a single product sees quantity-break ideas; a store spanning many product types sees collection-sale ideas.
Catalogue scan
A background scan of your products and variants: stock levels, prices, how long each product has been listed, and whether it has ever been part of a Stackable campaign. It runs on demand when you press Re-scan, and it refreshes itself weekly.
Sales tallies
Per-product units sold and revenue, counted forward from the day you install the app, from your store’s new orders. No customer details are ever stored, and there is no backfill of history. That is why some insights take time to unlock: the app can only count sales it has watched since you installed it.
The recommendations
Store Insights checks the rules below in ranked order and shows the ones that fit your store right now. These are the current rules; we tune the thresholds over time.
| Recommendation | When it appears |
|---|---|
| Clear your slow movers | When your slow-mover data is ready (see below) and at least one product qualifies. Suggests a volume discount targeting those exact products. |
| Clear aging stock | Shown before your sales data unlocks: products listed for 60 or more days, holding 10 or more units of stock, that have never been on a Stackable offer. |
| Reward your best sellers | When your best-seller data is ready (see below) and at least one best seller still holds enough stock to discount safely. Suggests a buy-more-save-more ladder on those products. If every best seller is running low on stock, this is held back and the Selling fast panel says so instead. |
| Pair a slow mover with a best seller | When both your best-seller and slow-mover data are ready and each has at least one product. Suggests a buy-together deal that pairs your top seller with a slow mover, moving the slow one on the strength of the fast one. |
| Prepare a Black Friday sale | From mid-October through the end of November. |
| Add a spend goal | Once you have roughly 20 or more recent orders and no live spend-goal campaign. The suggested threshold comes from your own higher-value orders (about the top quarter of order values, rounded to a friendly number). |
| Add a free-shipping threshold | When you have no live free-shipping campaign and your store is in a country where shoppers broadly expect free shipping (currently US, CA, GB, AU, NZ, IE). The threshold is anchored to your order values, or a sensible default when there is not enough data yet. |
| Add quantity breaks | For single-product stores. |
| Run a collection sale | For catalogues spanning more than 5 product types. |
| Launch your first campaign | When nothing is live yet. |
Every recommendation shows its title and a one-line reason. The exact numbers that triggered it sit behind a "Why this?" link on the card, which expands a short "Why you are seeing this" explanation, so nothing is hidden, it is just tucked away until you want it. "Create draft" opens a prefilled draft campaign in the builder for you to review; nothing publishes automatically. When a recommendation targets specific products, the draft prefills up to 30 of them, largest stock value first.
What counts as a slow mover
A slow mover is a product that has been listed at least 30 days, still holds at least 5 units of stock, and whose sales over the trailing 30 days fall in the bottom quarter of your comparable products. The list is ranked by how much stock value is tied up in each product, and it currently caps at the top 50. These are current defaults that we tune over time.
What counts as a best seller
A best seller is a product that has sold at least 5 units in the recent window and whose sales fall in the top quarter of your selling products. The Selling fast panel shows each one with its units sold, revenue, stock on hand, and days of cover (how long the stock lasts at the current pace). The list is ranked by units sold and currently caps at the top 20. These are current defaults that we tune over time.
Best sellers unlock sooner than slow movers, at 7 days of sales history rather than 30. The reason is honesty about what the numbers can claim. Calling a product a best seller only needs proof that it IS selling, which a shorter window can show; the panel always states the basis, for example "based on 14 days of sales". Calling a product a slow mover claims the opposite, that it is NOT selling, and the only honest way to know that is to watch a full month. So the two unlock at different times on purpose.
We never suggest discounting stock you are about to sell out of
A best seller must still hold at least 10 units before we suggest a discount on it, so a volume deal never speeds a near-empty product to zero. When every one of your best sellers is running low, the Reward your best sellers suggestion is held back and the Selling fast panel tells you your best sellers are running low instead. This stock floor is a current default we tune over time.
Why some insights unlock later
Nothing is guessed
Sales counting starts the day you install, with no backfill. Best-seller insights unlock after 7 days of history and slow-mover insights after 30, because claiming a product sells needs less watching than claiming it does not. Until each unlocks, the page tells you exactly what it is still collecting and when it unlocks. Numbers are never extrapolated or invented.
Dismissing and restoring a recommendation
Dismiss hides that recommendation. Dismissed recommendations are gathered under a "Dismissed" link at the bottom of the page, where you can restore any of them with one click. Recommendations are re-evaluated from current data every time you open the page, so a dismissed one can return if it still fits, and a suggestion whose conditions no longer hold drops off on its own.
What Store Insights never reads
Store data only
Recommendations use your store data only: your catalogue, stock levels, per-product sales counts, and your own campaigns. They never use customer information of any kind.
Helping a merchant (for support)
Why do I see no slow movers?
Three causes, in order. First, there is no catalogue scan yet, so run one with Re-scan. Second, there are fewer than 30 days of sales history, in which case the page shows the date the insight unlocks. Third, nothing currently qualifies, which is good news: no product is sitting slow.
Why do I see no best sellers?
Three causes, in order. First, there is no catalogue scan yet, so run one with Re-scan. Second, there are fewer than 7 days of sales history, in which case the panel shows when it unlocks. Third, nothing has sold enough yet to stand out as a best seller. If best sellers show but there is no "Reward your best sellers" suggestion, they are all running low on stock, which the panel notes.
Why is there no free-shipping recommendation?
Either the store already runs a live free-shipping campaign, or its country is not one of the ones where we treat free shipping as an expectation (currently US, CA, GB, AU, NZ, IE).
The stock value looks wrong.
Stock value is the units on hand times the product’s lowest variant price, taken from the last catalogue scan. If the catalogue changed recently, run a re-scan so the numbers reflect the current state.
Does this read customer data?
No. It uses the store profile, the catalogue, and per-product sales counts only. No customer information is read or stored.
A recommendation disappeared.
Either it was dismissed, or its conditions stopped holding. Recommendations are re-checked from current data every time the page opens, so both are expected behaviour.