Product analysis
Pick one product. Product analysis follows everyone whose first order with you included that product, and shows what happened next: did they come back, when, what did they buy, and how much were they worth?
Open My Shop → Analytics → Product network → Product analysis and choose a product under Choose a product. You can also get here from:
- a product's details panel on the Network map,
- any row in Starting products,
- the Purchase journey button on a product's detail page.
This page is rebuilt every Monday. The line under the title shows which order dates it covers. Recompute now starts a rebuild without waiting for Monday.
The three cards at the top
| Card | What it tells you | Example |
|---|---|---|
| Buyer LTV · 180d | On average, how much profit one of these buyers brought you in their first 180 days. It counts everything they bought, not just this product, and includes people who never came back. Hover the card to see how it varies from buyer to buyer. | ฿85: the average buyer who started with this product was worth ฿85 of profit within six months. |
| Next-basket AOV | AOV means average order value. For buyers who came back, how much their next order was worth compared with their first. | ฿250, up from ฿180 on the first order: returning buyers spent more the second time. |
| Network importance | Of all the repeat orders in your shop over time, the share that are for this product. | Rank 3 of 120: the third-biggest share. Buyers keep coming back to this product, whatever they bought first. |
Profit here means the price minus product cost, platform fees and the product's own ads, at today's prices and costs. Buyer LTV needs your product costs (COGS) and the Sensitive financials permission.
When do they come back?
- Return time. Each bar is a two-week block after the first order. The taller the bar, the more buyers came back for the first time in those two weeks. The headline is the median: half of the buyers who came back did so sooner than this, and half later.
- Return percentage. A running total. For example, if it shows 15% at day 30 and 24% at day 90, then 15 in 100 buyers had come back within a month and 24 in 100 within three months. Where the line flattens, few more buyers come back.
- Return activity. A grid. Each row is one more return visit: the top row (First) is buyers' first time back, the next row (Second) their second time back, and so on. Each column is about one month after the first order, up to a year. A dark square means many orders happen at that point. If only the top row is dark, buyers come back once and stop. Dark squares lower down mean buyers keep coming back.
What do they buy?
Product mix shows what these buyers go on to buy. Blue is what they buy in their first 180 days. Orange is what they're likely to buy over many future orders. If this product's own bar is the biggest, people mostly rebuy it. If other products lead, it works as a way in to the rest of your range.
Purchase flow & sensitivity
Read the flow left to right, like a river. The first column is the first order. Each column after it is the next time those buyers came back, up to five visits within 180 days. Each block is a product. The wider the band between two blocks, the more purchases go that way. Numbers are per 1,000 people who started with this product.
Click a band to see how much profit it brings and how long buyers usually take to make that step.
Two what if cards estimate extra profit per 1,000 people who start with this product:
- Bring more buyers back. What if a few more buyers came back for a second order, for example with a reminder or a voucher? Slide from 0 to 10 points. +2 means, for example, 20% coming back becomes 22%.
- Strengthen one connection. Pick a next product. What if returning buyers chose it 0–20% more often than they do now (10% to start with)? For example, at 10%, 20 in 100 becomes 22 in 100.
These are estimates before the cost of any voucher or discount, not a promise of results. Use them to decide what's worth testing, then compare the estimate with what the offer would cost.
- A visit is one buyer's purchases on one day. Several orders on the same day count as one visit.
- Buyer LTV · 180d uses buyers who've had a full 180 days. If no buyer has known costs for everything they bought, the card shows no number rather than a guess.
- Profit leaves out VAT and shop-wide ads that aren't tied to a product. Missing costs are left out rather than counted as zero.
- Return activity, Product mix, Network importance and the flow come from a model of how buyers move between products, built from your order history. The what-if cards use the same model, with everything else held as it is now.
- Return time and Return percentage count buyers who've had the full window, and stop at the last day the data covers.
Use cases
- Decide what to offer a product's buyers next. The biggest products in Product mix, and the widest bands out of the first column of the flow, are what these buyers actually go on to buy. Follow up with that product, or bundle the two.
- Time the follow-up. The Return time median and where Return percentage flattens tell you when to send a reminder, and the point after which a buyer is unlikely to come back without one.
- Size a retention offer. Bring more buyers back shows what one or two extra points of returning buyers is worth per 1,000 starts. Keep the cost of the offer below that.
- Compare products to win new buyers with. For two products at similar prices, compare Buyer LTV · 180d and Next-basket AOV. The product whose buyers are worth more is the better one to advertise to new buyers.