We ran the biggest public take a look at of internet vs. in-app purchases. Right here’s what 5,600 customers taught us.
Seems, in-app purchases are good for conversion charges. The truth is, at the least 30% higher. That’s one of many issues we discovered whereas operating the primary large-scale, side-by-side take a look at of in-app vs internet purchases in historical past.
Learn on to search out the complete outcomes, find out how we ran the experiment, what we predict the findings may imply, and what you are able to do right this moment to run an analogous take a look at with RevenueCat. For those who’d prefer to see us assessment the most recent experiment knowledge reside, be part of our subsequent bi-weekly workplace hours this Friday.
The way it all started
Two weeks in the past, there was a court docket order within the Apple vs Epic case that compelled Apple to permit builders to avoid in-app purchases (IAP). As of April thirtieth 2025, builders have been lastly allowed to ship clients in the US to an exterior web site to finish the purchases, and thus keep away from the 30% payment that App Retailer takes. We shortly launched Net Buy Buttons that, mixed with our Net Billing product, create a seamless means for builders to drive purchasers out of their app to the online for take a look at.
RevenueCat’s new Net Buy Button in motion
IAP within the US is a $52B market, representing about half of all App Retailer income, and that is the primary time that Apple has opened up such a serious market to internet buying. Beforehand, comparable court docket orders have been restricted to a lot smaller markets. It has been a lot debated through the years how a lot, or even when, this new-found freedom would assist builders make more cash. The one actual approach to know is to run a take a look at, so we took this chance to run the most important open take a look at of internet purchases in historical past.
Final week, we deployed an experiment on Dipsea (the app we acquired final yr as a testing-ground) and we’re excited to share preliminary outcomes.
Our experiment
We ran a four-way take a look at utilizing RevenueCat’s Experiments software, with the next variants:
| Variant A | Variant B – Clone | Variant C – IAP + Net | Variant D – Net Solely | |
| Paywall | Dipsea’s native paywall | A clone of Dipsea’s native paywall applied with RC Paywalls | Choice to subscribe on internet & save 30% | A clone of Dipsea’s native paywall however purchases can solely be made on internet (no IAP) |
| Cost Choices | IAP | IAP | IAP + Net Buy | Net Buy Solely |
| Costs | $69.99 Annual (IAP)
$25.99 Quarterly |
$69.99 Annual (IAP)
$25.99 Quarterly |
$69.99 Annual (IAP) + $47.99 Annual (Net)
$12.99 Month-to-month (IAP) |
$69.99 Annual (Net)
$12.99 Month-to-month (Net) |
| Notes | This serves as a baseline and to isolate any points probably launched with our implementation of Paywalls | We tried to match the native Paywall pixel for pixel.. Identical costs as within the management. | This variant introduces a reduced providing on the internet, providing 30% off the annual IAP worth. Theoretically, this could make up for the 30% payment Apple prices builders, and builders needs to be detached between these choices | This variant removes all IAP choices. Dipsea is in a position to do that as a result of it falls below the “reader app” exemption. In contrast with variant B, it helps us characterize the inherent friction within the internet circulate, holding all different issues fixed. |
We began the take a look at on Might seventh concentrating on all of Dipsea’s new US clients. Up to now roughly 5,600 members entered the experiment, totalling roughly 1,400 clients per variant.
Outcomes at a look
Early conclusions
The experiment has solely been operating for six days, however we’ve already seen some attention-grabbing outcomes.
IAP vs internet
Variants B and D have the clearest, and best to interpret, variations. They’ve the very same design and format, the solely distinction is in-app purchases vs exterior purchases. And we will get a reasonably clear understanding of the drop-off brought on by kicking somebody out to the online and circumventing IAP.
The IAP Solely variant of the paywall (Variant B) had a 42% elevate in Preliminary Conversion and a 43% elevate in Trial Begins vs. the Net Solely variant (Variant D).
| Variant | Paywall Views → CTA Faucet | CTA Faucet → Preliminary Conversion | CTA Faucet → Trial Begin | Paywall View → Trial Begin |
| Variant B – IAP Solely | 50% | 73% | 67% | 33% |
| Variant D – Net Solely | 52% | 42% | 37% | 19% |
| DIFF | -4% | 42% | 45% | 43% |
This early on, the one metric we will have any confidence in is Preliminary Conversion (trial or subscription begin).
The preliminary conversion charge for variant B is between 27% and 30%, whereas the equal internet circulate in variant D is between 17% and 19%. It is a massive lower, a 25% to 45% relative drop between the 2. Digging into the funnel, most of that drop happens from the fee sheet by to buy. That’s plenty of fall off.
We don’t have knowledge on trial conversion charge but, or every other downstream charges that will have an effect on LTV. It’s potential the online cohort has the next LTV, however they’re beginning out from a deep gap. Decreased charges aren’t the one good thing about internet purchases; they normally have extra instruments to retain and serve these customers. Cost processors like Stripe pay out way more shortly than IAP, lowering cashflow constraints.
Run this experiment in your individual app
That is one knowledge level from one specific app, with a selected person base. If you wish to actually perceive how your app will do with an online primarily based buy circulate, you’ll need to set this up your self. Thankfully, we make this straightforward:
- Arrange RevenueCat Net Billing by connecting your Stripe account to RevenueCat
- Create a Net Buy Hyperlink, connected to an Providing that features Net Billing merchandise
- Add a Paywall to your Providing with internet purchases that makes use of the Net Buy Button part to hyperlink out to a Net Buy Hyperlink.
- Create an experiment evaluating your present IAP Providing & paywall with the brand new Net Billing providing
Limitations and coming knowledge
It’s solely been per week, so this knowledge could be very preliminary. Large sudden swings in issues like trial conversion charge, renewal charge, refund charge or extra may have an effect on the end result, nonetheless with the numerous dropoff in preliminary conversions on internet, this doesn’t appear possible. It appears possible that pushing customers out of the app has a considerable impression on preliminary conversion charge, and isn’t a “free” 27% boon to builders.
If you’re on the Small Enterprise Program, it’s unlikely sending site visitors to the online will internet you forward given Apple’s payment is simply 15% of income. If you’re paying 30%, it’s potential that it’s nearer, however a naive implementation, with no low cost tradeoff, doesn’t appear to be price it. Maybe with extra subtle concentrating on, higher framing, completely different checkout pages, or some center floor on pricing, you could possibly discover a approach to make the drop off in conversion make up for the elevated income, however you might be enjoying a sport of inches.
We’ll proceed to observe the take a look at and can replace with new posts within the coming weeks, together with a roundup when we have now confidence within the mid-term LTV implications of the experiment.
We’ll additionally maintain delivery like loopy so you may have the instruments to make testing like this even simpler.
