Playables and Rewarded Placements, Built and Bought Together

A playable that converts everybody who opens it has failed. Rewarded inventory hands you a user who tapped through a video to collect a reward, and the playable's job is to lose the ones who were never going to stay before you pay for the install. The shape we aim for is an engagement-to-install rate between 8 and 15 percent with D7 retention holding within four points of the same app's non-rewarded paid traffic. A 40 percent engagement-to-install rate with D1 falling off a cliff is the same money spent worse, and it will look like a win for about a fortnight. The build: Playables are written as HTML5 against MRAID 2.0, packaged inside 5 MB for Unity, AppLovin, Mintegral and Liftoff, with the end card calling mraid.open rather than a hardcoded store URL so one file ships to every network without a rebuild. We build in Playable Factory or Luna Playworks when there is a Unity build worth reusing, and by hand when there is not, which for a listings app or a subscription app is most of the time. Turkish and English versions are produced together rather than translated afterwards, because a localised call to action inside a 20 second interaction reads as one. Non-game playables, which is most of our work: A property portal's playable is a three step search: district, budget, one result the user taps. A marketplace's is a single listing flow ending at the point of enquiry. A pet commerce app's is a breed or weight selector that returns a real recommendation. Each shows the actual first run of the app in the actual interface. We do not build fake UI, a false gameplay loop, a tap-here arrow over a screen the app does not have, or anything that misrepresents what the product does. That is not a taste position. It moves the loss from a cost you can see to a D1 number you cannot fix. The buying: Rewarded video and rewarded interstitial on Unity Ads, AppLovin and Mintegral, bought separately from your non-rewarded activity and with separate targets, because blending them makes both unreadable. Rotation runs on engagement-to-install and on retention by creative, never on click-through, which on a rewarded placement mostly measures how much the user wants the reward. Fatigue on these networks appears as engagement rate decay before it appears in CPI, so the replacement schedule is set off that curve rather than a calendar. Not included: Video production at volume, UGC creator sourcing, brand film, and any arrangement where we ship a large number of variants and let the network sort them out. We build a small number of concepts that each test a different proposition and kill them on evidence. If you want a supplier producing dozens of cuts a month, we are the wrong shop and we will say so on the first call rather than after the contract. Who this is not for: Apps whose first run takes longer than 30 seconds to reach anything worth showing, since there is nothing honest to put in the playable. Teams with no design system or build we can pull assets from, unless that work is being bought elsewhere and delivered to us. And anyone whose measure of a playable is its click-through rate. What you receive: Source, not just packaged bundles. Each playable ships as its own repository with assets, build script and network manifests, so your next agency can change a headline without coming back to us. Alongside it, the weekly buying note: engagement to install, install to event, retention by creative, and what is being replaced next.