A web store produces more numbers than a game needs. Five of them run the store; the rest are diagnostics. This page is the five, where each comes from in an Appside setup, and how to read them without a dashboard project.
The five
| # | Number | Tells you | Source |
|---|---|---|---|
| 1 | Store visits | Is traffic growing? Which channel sends it? | Your traffic tag, or the store's visit counts; channel from promo codes (attribution) |
| 2 | Visit → checkout | Is the store convincing? | The Studio funnel (store → checkout) |
| 3 | Checkout → paid | Does checkout work for these players? | The Studio funnel (checkout → paid); payment methods per region |
| 4 | Repeat purchases per paying player (30 days) | Is there a reason to return? | Your order table |
| 5 | Refund rate | Are offers clear and delivery reliable? | Your order table (order.refunded), the Studio's refund count |
Revenue is the result of 1 × 2 × 3 × average order × (1 + repeat) − refunds. Watch it, but fix the factor, not the product.
The three sources
1. The Studio. Studio Insights shows store visits, checkout starts, paid orders, gross revenue, refunds and net of refunds by day or week. It also shows visit-to-paid conversion, ARPPU, first-purchase rate and repeat buyers within 30 days. Break figures down by source, campaign, pack and identity method. Revenue reconciles with the order ledger; currencies stay separate, and memberships are excluded. Visits require browser engagement and can slightly undercount slow connections.
2. Your order table. Every paid order reaches your backend as a signed order.paid webhook with the order ID, player ID, pack, credits, price, currency and time, and every refund as order.refunded with the same order ID (reference). One table, one row per order, with the code or channel your backend knows. This is the source of truth for money, cohorts, repeat rate and refunds, and it joins to your in-app purchases on the player ID. Build it before launch; it's also the reconciliation table.
3. Your traffic data. Studio Insights records visits and sources from store links, SDK placements and referring sites. Your own traffic tool can add acquisition context. For revenue and refunds, use the order ledger rather than browser tags.
Reading them weekly
Same day, same order, one page:
Week 44 visits 1,240 →checkout 31% →paid 62% repeat/payer 0.4 refunds 1.8%
Week 45 visits 1,610 →checkout 29% →paid 64% repeat/payer 0.6 refunds 1.1%
Rules for reading:
- One number changed a lot? Find the change in the calendar that week (live ops) before you look anywhere else.
- Visits up, checkout rate down? New traffic from a weaker channel. Fine, if it pays.
- Paid rate down in one country? Payment methods (localized pricing).
- Repeat flat for three weeks? The store has no reason to return to (retention).
- Refunds up? Offer copy or delivery; read the tickets.
What not to measure (yet)
- Lifetime value of web buyers. Needs months of data; in month one it's noise.
- A/B tests on store design. Not until visits are in the thousands per week; before that, decide.
- Share of in-app vs web revenue. Leadership wants it, finance needs it, but it doesn't tell you what to fix. Compute it monthly from the order table; don't manage by it.
- Anything per hour. Weekly, or daily at launch.
The order table, minimal schema
order_id · player_id · sku · credits · price · currency · paid_at · refunded_at · code · channel · game
Ten columns, every question above, and the finance team's month-end, from one webhook handler.
Create your store free: use Studio Insights and signed webhooks to follow your store. No card needed to start; 3.9% per successful payment.