I audited 15 sales from my paid mobile game. All 15 increased purchases, but two made less money.
retained text
A few days ago I posted here about my paid mobile game falling to roughly one sale a day after more than two years of updates. The responses helped me look at the situation a little less emotionally. A lot of people said that I had supported the game long enough, that existing players weren’t owed free content forever, and that it might simply be time to move on. I’m not quite ready to do that yet. Before deciding what comes next, I went back through the App Store and Google Play reports to see whether all the discounts, price changes, updates, and press outreach had actually helped. I expected the answer to be fairly simple. It wasn’t. **Every discount increased purchases. Not every discount increased revenue.** I compared 15 Google Play promotions with equal-length periods immediately before them. Purchases per day increased in all 15. Developer proceeds per day increased in 13, but decreased in two. One Halloween sale increased purchases per day by 2.66x while developer proceeds fell to 0.80x of the previous period. I sold more copies and made less money. The best result was almost the opposite: 25.15x the purchases and 9.36x the proceeds per day. So discounts definitely created activity, but the size of the sales spike didn’t tell me whether the event was financially successful. Discount depth, refunds, country mix, taxes, and store fees all mattered. **The biggest sales spike looked like a recovery. It wasn’t.** During a longer $0.99 period on iOS, I saw: - First 30 days: 1,031 net units and $759.09 in developer proceeds - Days 31–60: 120 net units and $89.94 - Days 61–90: 114 net units and $88.60 For that first month, it felt like the game had come back to life. Then it dropped straight back toward the old baseline. About 65% of the first month’s iOS units came from Taiwan. I had contacted Taiwanese gaming sites directly, one of them published an article in Traditional Chinese, and it was republished or summarized by other sites. The game also reached #1 in Taiwan’s paid puzzle category during that period. At the time I wanted to give the lower price most of the credit. Looking at it now, the spike was probably a mixture of price, localization, local press, direct store links, and chart visibility. I didn’t use a separate campaign link for each publication, so I can’t tell how many purchases came from any one article. That was a clear tracking mistake on my part. **Even the comparison window changed the answer.** On Android, comparing the 60 days before and after the same reference date gave me: - Purchases: 361 to 435, up 20.5% - Developer proceeds: ₩477,342 to ₩521,065, up 9.2% But using only the nearest 30 days: - Purchases were down 1.6% - Developer proceeds were up 3.3% The earlier periods already contained other discounts, so this wasn’t a clean $2.99 versus $0.99 experiment. I also found that the result changed depending on whether I used Google’s original Pacific dates or converted every transaction to Korean time before creating the windows. Some transactions moved from one side of the comparison to the other. **I also had to throw out my first retention analysis.** I originally used a server-side `last_login` field to estimate D7 and D30 retention. Later I discovered that saves and other account changes could update that field. It wasn’t measuring what I thought it was measuring. I recalculated the cohorts using actual GA4 activity. D30 changed from 2.81% to 2.46% on iOS and from 3.33% to 2.63% on Android. That still doesn’t prove lower-price users retained worse. These were first-visit cohorts, not confirmed groups of $2.99 and $0.99 buyers. I can’t connect an individual player’s activity to the price they paid. The iOS cohort also became much larger. The D30 percentage fell, but the actual number of people returning on day 30 increased from 10 to 29. Depending on which number I looked at, I could tell myself two completely different stories. **What I’m trying now** After more than two years without IAPs, I’m currently running a one-month experiment on iOS: - No upfront price - One optional permanent class-unlock purchase - All classes can still be earned through normal play - No ads - Android remains paid, because Google Play doesn’t let the same app return from free to paid This isn’t a victory lap or a claim that free-to-play is the answer. It’s a reversible test on iOS to see whether removing the upfront price brings in enough new players, and whether a single optional purchase can support continued maintenance. I don’t have the result yet. A few early downloads or purchases won’t mean much. I need the full month, refunds, proceeds, and mature retention cohorts before deciding anything. The biggest thing I learned from reviewing the paid-game period is that temporary success is very easy to mistake for a new baseline. A sale can increase purchases and reduce revenue. A retention rate can fall while the number of returning players rises. A game can reach #1 in a category and still return to one sale a day later. I put the full monthly history, all 15 promotion comparisons, retention methodology, and the corrections to my earlier analysis here: https://dungeontracer.com/pricing-report For anyone who has done a similar pricing or monetization test: what result would convince you that a free-entry model is actually sustainable rather than just another temporary spike?
document fetched 2026-09-28 00:27 · canonical https://reddit.com/r/gamedev/comments/1wrbf1g/i_audited_15_sales_from_my_paid_mobile_game_all