Measuring D2C: the nine numbers to report

The nine numbers, and how to calculate each

A web shop report for the board fits on one table. Measure every row by platform and by market, because store fees, payment methods and dispute rules all differ by both.

The sections below cover what published evidence exists for each, and where the evidence stops.

Web share of revenue: what listed publishers report

Several listed publishers now report direct sales each quarter. Each measures on its own basis, so the list is a direction of travel rather than a league table. The D2C share tracker keeps the full series.

Measure your own share on gross revenue in both channels. If you compare web gross revenue with store revenue booked net of the store’s cut, the web share looks larger than it is.

Payer conversion, ARPPU and repeat rate

Expect a web visit to convert less often than an in-app purchase prompt. The only controlled public test, run by RevenueCat on the Dipsea subscription app in May 2025, converted 27.0% of users with in-app purchase only and 18.1% with a web-only flow. It is one app and not a game, so treat it as direction, not a target. Measure conversion by source, because a logged-in link from an email to a known spender and a cold social click are different funnels.

Compare ARPPU between channels on the same cohort. A raw comparison flatters the web shop, because the players who go to the trouble of buying on the web are already your higher spenders. Appcharge, a web shop vendor, reports that average web store orders in the US were three times the size of orders through its payment links. That is a difference between two web products, driven partly by who uses each.

Repeat rate is the metric most tied to web revenue. Appcharge reports that 97% of the web store revenue it processed came from repeat buyers, and that repeat rates rose above about 84% once a player reached a third web purchase. Report the share of first-time buyers who reach purchase three, by acquisition source.

Net revenue per $100 after all costs

This is the number the business case rests on. Work it out per channel and market:

Playtika says in its 10-Q that payment processing and related costs for its direct sales “are typically 3 to 4%, compared to the 30% platform fee”. Stripe’s US pricing is 2.9% + 30¢ on domestic cards, and Stripe states that processing fees on refunded payments are not returned. The hidden fees page lists the lines that headline rates leave out.

Expect the P&L to move in two directions at once. Stillfront describes its direct sales as “negative for net revenue but accretive to gross profit and gross margin”, because the store’s cut no longer comes off the top.

Authorization and refund rates

Authorization rate is the share of payment attempts that the issuer approves. No neutral benchmark for games is published, so track your own by payment method, card country and amount. The levers are technical: routing, retries and local methods. For example, Stripe says Cartes Bancaires, France’s card network, “can have a positive effect on your acceptance rate in France” and retries technical declines over Visa or Mastercard.

Refunds usually cost you the original processing fee. A rising refund rate can point to the same problems that drive chargebacks: unclear offers, delivery failures, or purchases by minors.

Dispute ratio and Visa’s monitoring program

If your studio is the merchant, card disputes count against your own merchant account. Visa’s Acquirer Monitoring Program (VAMP) measures:

It covers card-not-present transactions, so it applies to every web shop sale.

The thresholds apply once a merchant has at least 1,500 fraud reports plus disputes in a month (in CEMEA, 150 and USD 75,000). Stripe notes that the same transaction can count twice, once as a fraud report and once as a dispute. Visa does not publish fine amounts in the fact sheet; your acquirer sets them in its contract.

Mastercard runs a separate program. A 2019 acquirer guide from J.P. Morgan puts its first tier at 100 chargebacks and a 1.5% chargeback rate in a month; check the current Mastercard rules with your acquirer. With a merchant of record, both programs apply to the provider’s account, not yours, but a provider will still act on a game that drives its ratio up. The chargebacks page covers prevention.

Cannibalization and incrementality: run a holdout

A web shop can grow revenue, or move the same revenue to a cheaper channel, or both. Moving revenue is still worth money, since each dollar keeps more margin. But the board case differs, and only an experiment separates the two.

Studios surveyed by Appcharge and GDC reported a median revenue uplift of 15% from direct sales, and 35% for the leaders. Those are self-reported figures in a vendor survey, not measured effects.

To measure your own:

1. Randomize by player. Assign a fixed share of eligible players to a holdout that sees no web links, web offers or web emails.
2. Measure total spend. Compare revenue per player across store and web combined, net of each channel’s fees, not web revenue alone.
3. Run a full cycle. Cover at least one full event or season cycle so the result is not a timing shift.
4. Keep the holdout clean. Exclude holdout players from web-offer emails and creator campaigns, or the groups blur.
5. Read two answers. The difference in net revenue per player is the value of the web shop. The difference in gross spend tells you whether it grew the pie.

Keep a small permanent holdout after launch. Store fees, link rules and offers change, and the incremental value changes with them.

Monthly reporting checklist

1. Web share of gross revenue, by platform and market.
2. Conversion by traffic source, and share of first-time buyers reaching a third purchase.
3. ARPPU for web and store on the same cohort.
4. Net revenue per $100 by channel and market, with every cost line shown.
5. Authorization rate by payment method and card country.
6. Refund rate and dispute ratio against your provider’s and Visa’s thresholds.
7. Incremental net revenue from the holdout, with the test dates and group sizes.