ShareChat and Moj, bought by language not by geography

ShareChat and Moj are bought by language, not by geography: fifteen Indian languages, two products, one ad account. A Marathi cohort in Pune and a Bhojpuri cohort in Patna are two different media buys that happen to share a country and nothing else. Impressions here cost a fraction of what Meta charges in India, which is precisely why European brands lose money on them. Cheap inventory absorbs badly aimed creative indefinitely without ever telling you it is not working. Language is the buy: I start by choosing three to five languages and defending the choice against your distribution rather than against population size. Hindi is the default answer and usually the wrong first one, because it is the most contested and the dearest of the Indian-language cohorts. Where your product actually sells, the second and third languages often carry better cost per install and materially better view-through. That selection, with the reasoning written down, is the first deliverable in every engagement here. What creative gets made: Vertical video made for this audience, not Reels cuts with burned-in subtitles. The difference is not subtle: pacing, on-screen text density, music licensing and the first two seconds all diverge, and Moj's audience skips a Reels edit on sight. I brief and pay creators and voice talent in each language, and they are named on the invoice. For ShareChat's feed I keep separate static and short-form assets, since the two products behave differently despite sharing an ad account. Brand filters and hashtag activations on Moj exist and I recommend them rarely, when there is a specific reason for participation rather than viewing. Measurement you can trust, and what you cannot: App installs tracked through your MMP are reliable, and cost per install is the number I will stand behind. Video view rate and completion are reliable. Web conversion attribution is weak on both surfaces and I will not build a case on it. If your business is a website checkout with no app, I will say so before you contract, and you should probably spend the money elsewhere. Brand lift is available at meaningful spend levels and I run it when the budget supports a clean read. What I refuse: English creative. A single India audience. Running Moj as an install machine for a category where day-seven retention collapses, because the platform will sell you that volume forever without complaint. And I do not accept a target cost per install benchmarked against Meta India until we have agreed what quality of install we are counting, since the number is easy to hit and easy to make meaningless. Outside the engagement: Media spend, paid into the ShareChat Ads account, which is opened in your name. MMP subscription and SDK integration. App store listing localisation, which matters more than the advertising does and belongs to whoever owns your store presence. Josh, Instagram Reels and YouTube Shorts, all adjacent and none of them included here. Who should not buy this: Brands wanting a cheap top-up of impressions to flatter a blended CPM. Companies with no Indian-language customer support and no plan for any. Anyone whose approval chain requires every asset signed off in English translation, because that review cycle destroys the creative and the schedule together. Cadence: Fortnightly creative refresh through the first quarter, because fatigue on Moj is fast and visible inside ten days. Monthly reporting by language cohort rather than aggregated, showing spend, cost per install, day-seven retention where your MMP exposes it, and the cohort I would drop next.