Ten days after launch: ads made $3, the "no ads" pack made ₩38,700

Dungeon Division is an idle dungeon defense game where you side with the dungeon and stop the adventurers who keep breaking in. I put it on the App Store on October 23. Today, November 2, is day ten. This post is about the money from those ten days: AdMob revenue and in-app purchases, in exact amounts.

Most solo dev revenue posts have a big monthly number in the title. Mine are small. I didn't round them up or leave anything out. They come from the admin dashboard I built for the game: purchases from the server ledger, ads from the AdMob and LevelPlay reports, downloads from the App Store Connect sales report. The 18 test purchases are excluded. Prices are in Korean won (₩1,000 is roughly US$0.70).

The first 10 days of AdMob and in-app revenue

Converted roughly, the ad revenue is a little over ₩4,000. About one café latte. In-app purchases made more than ten times that.

The first week: zero

Launch day brought 2 downloads. Then 1, then 1. October 26 had 8 and the 27th had 9, and after that it moved between 4 and 7 a day.

Over the same period the game showed up in App Store search results 10,121 times, and 626 of those turned into a product page view.

A full week of zero. On the afternoon of the 26th, GA's realtime view showed a "payer", so I opened the server ledger thinking someone had finally bought something. All it held were my own 17 test purchases. I'd misread the GA label, and nobody had bought anything yet.

October 26: AdMob approval and my first 68 cents

Ads were worse. As I wrote in part 12, my AdMob account wasn't approved on launch day, so for the first three days ad revenue was zero. Tapping a rewarded ad button just said there was no ad right now.

On the morning of the 26th the AdMob console said the account was approved and ad serving was on. That day 9 ad requests went out and all 9 were filled. First-day ad revenue: $0.68. After that it moved between $0.14 and $0.74 a day. November 1 was the best day with 14 impressions.

LevelPlay, the backup ad source I bolted on during launch day, served exactly one ad in ten days, for $0.00. It did keep one player from seeing the "no ad" message, so I'm counting that as a job done.

Dev notesWhy the rewarded ad eCPM says $76

The dashboard shows an AdMob eCPM (revenue per 1,000 impressions) of $76.98. That looks wild, but it comes from having only 42 impressions. A couple of high-paying ads in a tiny sample push the per-1,000 figure way up, and rewarded ads in Korea pay more than banners anyway. Once there are a few thousand impressions it will come down, so I'm not using this number to forecast anything.

AdMob revenue is also an estimate. It gets finalized after invalid traffic is removed, and AdMob only pays out once you pass a payment threshold. Three dollars is a long way from that.

October 30: one player, three purchases

The first purchase came on day 8, October 30. One player bought three things that day:

• Manager Perks, ₩12,900
• 1-8 Staff Pack, ₩3,900
• Starter Support, ₩2,900

That's ₩19,700. Compared with the shop screen below, they took everything near the top except the Legend recruit offer (₩9,900). On the 31st a different player bought Manager Perks. On November 1 another player bought Manager Perks and the 1-8 Staff Pack (₩16,800).

The 1-8 Staff Pack is a deal that appears for 72 hours after you clear stage 1-7. Stage 1-8 is a Dark stage, so the pack lets you pick one Light hero and levels them to 20. Two people bought it in ten days. It's an offer that shows up right before the wall, and it worked better than I expected.

All three buyers bought the right to skip ads

So: three paying players, and all three bought Manager Perks. ₩38,700 of the ₩49,400, 78%, came from that one product.

Manager Perks is a permanent pack. You get the rewards from all nine ad spots without watching the ads, and 3x battle speed with no limit. In other words, the thing that earned the most in these ten days was the product that removes ads. Ads: $3. The pack that turns them off: ₩38,700. Put side by side, it's a little funny.

This product used to be called "Rewards without ads". As I wrote in part 10, during the October 12 shop rework I renamed it Manager Perks, put "3x battle speed, unlimited" on the first line and moved it to the top of the shop. Clearing the story takes hundreds of replays of the same stages, so I figured faster battles would matter more to players than skipping ads. The detail sheet also shows numbers based on your own save, like "up to 13 ad rewards a day, instantly" and "72 normal stages left". Did the rename sell it? I can't tell. I never launched with the old name, so there's nothing to compare against. What I can tell you is that a ₩12,900 pack sold three times.

Here's the shop funnel from GA events (this includes the Android test builds):

Of the 24 who opened the shop, 18 left without tapping a single product. But almost everyone who looked at a product went on to the payment sheet. The weakest link right now is the shop's first screen.

Dev notesWhere the App Store report and my estimate disagree

The dashboard's "estimated net" of ₩38,173 is my own estimate: store price minus 10% VAT, minus a 15% commission. Apple's sales report lags a day. Through October 31 it shows 4 units and ₩28,155 in developer proceeds, with one Manager Perks at ₩11,141. That's more than my estimate of about ₩9,968. I'll check which one is right at the first payout. Differences like this are why the dashboard keeps my server ledger and Apple's report as separate panels.

42 downloads, 9% day-1 retention

Only posting the good numbers would be cheating, so here are the bad ones. Day-1 retention is 9.1%, about 7 of 77 players. Day-7 retention still reads 0%, although not many first-week players have been around for seven days yet.

Fewer than half of the players who started the tutorial finished it, and 10 cleared a chapter. Those who stay play for a while: average engagement is 20 minutes 43 seconds. At least two of the three buyers had cleared 1-7, since that's when the 1-8 Staff Pack appears.

Of the 42 downloads, 37 were from Korea, 2 from Canada, and 1 each from the US, Australia and Malaysia.

Which rewarded ads people actually watched

Every ad in this game is a rewarded ad. No interstitials, no banners. You only see one if you tap the button for the reward. Here's how many were watched to the end in ten days, by placement (GA counts, Android test builds included):

That's 63 in total, plus 9 closed before the end. The Hire ad was watched by 15 players, roughly one in two. If there's a "one more" next to the button that gets you new staff, it's hard not to press it. Business gold went the other way: three players watched it nine times. The people who use it keep using it.

GA's 63 is higher than AdMob's 42 impressions because GA includes the Android test builds. The Android app in AdMob is still marked "review needed", so it doesn't earn anything yet.

The night of the first purchase, I worked on ads

On the night of October 30 I made two ad commits.

First, I split the ad events. Until then, "no ad was available" was being counted together with "the player cancelled". On days with no fill, that makes it look like players hate ads. Now the game logs three moments separately: when the button becomes available, when it's pressed, and when it fails because there's no ad.

Second, menus like Hire, Forge and Battle now show a badge when an ad is ready to watch. You used to have to open the menu to find the ad button; now you can see it from the lobby. By GA counts, the most-watched ads so far are Hire (18 views) and Repel gold (18 views). Last, I added daily ad boxes that open at 3, 5 and 8 ads watched, with gold, a Soulstone or Reinforcement Stone of your choice, and Mana Gems.

Players with Manager Perks fill these boxes without watching anything. If you paid to skip ads, you should still get the ad rewards.

Closing out the first ten days

Ten days earned ₩49,400 in purchases and $3.23 in ads. Measured against the hours I put in, that isn't an hourly wage. Still, a column that sat at zero for a week finally got a bar on day 8. Three strangers paid ₩12,900 each for my game. Of all the numbers in this post, that 3 is my favorite.

I plan to do the next tally around November 23, one month after launch, and I hope Android will be out by then. Before that I'm fixing the tutorial, where more than half of players drop off. That's the next post.

If you want to see where this game started, part 1 is the first prototype. The look back at the 39 days to launch is part 13.