Player Lifetime Value on Day 3 to Bid on Acquisition
A mobile game studio spends €2M a month on player acquisition campaigns. 95% of players never pay and 1% generate 60% of revenue. To optimise bids by campaign and country it needs to know, in the first 3 days, how much each player will spend over 180 days: waiting 6 months is not viable.
Where the money comes from
What to do now
- 1Calculate LTV net of store commissions by country.
- 2pLTV model with data from the first 3 days.
- 3Send the pLTV to the ad networks for bidding.
For your committee
One page with the decision, the impact, the risks, the steps and the KPI to track.
The print dialog opens: choose "Save as PDF".
Watch out: more volume is not always more profit
Optimising for cheap installs or for gross revenue: app stores keep a share of every purchase and certain countries have revenue with high server and support costs.
KPI for the committee
Net 180-day ROAS by campaign.
Detail for your data team
Proposed technique: XGBoost
Why: It is a numerical prediction on massive tabular data, with non-linear relationships (completing the tutorial matters more in some countries than others) and a highly skewed distribution. XGBoost, with a two-part formulation or Tweedie loss, is the industry standard for pLTV.
Cost of getting it wrong: Overestimating LTV leads to bidding high for players who do not pay and burning budget; underestimating it means losing the most valuable players to the competition.
Data needed: Game events from the first 3 days: sessions, levels, tutorial completed, social interactions, first purchases.; Source: campaign, ad network, country, device.; Actual 180-day spend of 5M historical players.