Fabio “Petrucio” Stange shipped Marvel’s Iron Man VR and Batman: Arkham Shadow at Camouflaj before going solo as Lone Piggy. Between late April and the end of August 2026 he spent about $5,100 of his own money advertising his roguelite deckbuilder Kegs of Eternity on X and Reddit, with every click tagged so Steam’s UTM report could tell him which visits turned into wishlists. The result was roughly 1,770 wishlists at about $2.88 each — and a set of findings that invert most of what indies assume about paid ads.

The tactics he will defend with data:

  • Mobile beats desktop by an order of magnitude. A desktop click opens the Steam page in a browser where almost nobody is signed in (7% were), so it converts at 0.8 wishlists per 100 visits. A phone click deep-links into the Steam app where the user is already signed in: 4.5 per 100 on iOS, 11.5 on Android. Turn web/desktop off.
  • Split Android and iOS and bid them separately. Android clicks wishlist at 1.2–3.8x the rate of iOS clicks in every country with enough data. Bid to equalize cost per wishlist, not cost per click.
  • Use ads to choose your Steam capsule. Steam’s own CTR metric is contaminated — external visits inflate it with no matching impressions, and success makes it fall. Run each capsule variant as its own UTM-tagged ad, multiply the ad platform’s CTR by Steam’s wishlist-per-tracked-visit rate (WLTR = CTR × V2WR), and pick the highest.
  • Judge everything on cost per wishlist from Steam’s UTM page, never on CPC from the ad dashboard. He killed Germany on day three because the clicks looked expensive; when Steam’s numbers caught up it was his second-best market at $1.64 per wishlist with a 27% conversion rate.
  • The first 48 hours of a new ad cell are a lie. Ad platforms front-load delivery to the most engaged users, so four of his cells ran at $0.64–0.80 per wishlist for two days and $1.44–4.01 afterwards. Day one is the ceiling, not the forecast — and the “50 cents per wishlist” figures people quote are day one.
  • Sustained spend in one cell decays. France/Android went from $1.65 per wishlist to $5.02 across July and August; a five-week pause bought two cheap days again before the decay resumed.
  • Non-English markets beat the Anglosphere. Japan was his best market at $1.22, matching the US. Germany, Spain and France beat the UK, Canada and Australia. Korea did not work.
  • A static JPEG beat the trailer. 1.8 wishlists per 100 visits versus 0.8 for the trailer, same week and targeting — a trailer gets clicks from everyone; a boring image with a one-line pitch gates the click on intent. Your Steam capsule is the wrong ad creative, because it was designed for a place where clicks are free.
  • Reddit is not 20x worse than X. His first campaign, on recommended settings, cost $28.57 per wishlist. His second, with manual bids, cost $2.65 — the same ballpark as X.
  • Budget exploration separately from exploitation. $1,521 of deliberate tests at $4.54 per wishlist, $1,667 running the winners at $2.23.

His own framing on whether any of this is worth it: a wishlist is a promise, not a sale. Break-even is roughly price × 0.45 × conversion rate, which puts a $2.95 blended cost per wishlist underwater on week-one conversion rates and comfortably positive on year-one ones — and he says so before anyone can accuse him of a victory lap.

That is the point of publishing pre-launch. Almost every indie marketing write-up is survivorship bias: the tactics come from devs whose games worked, filtered through hindsight. He has no idea yet how the game will do, and Part 2 will test whether removing the Steam AI-disclosure banner (his placeholder art was AI-generated; it is being replaced with commissioned work) moves conversion — argued with numbers instead of on the internet.