iAds performance – now and later

by greg on July 8, 2010

The Next Web writes that an iPhone developer made $1,372 from his first day running iAds. The developer was kind enough to post the following stats from his Apple dashboard:

  • $1,372.20 total revenue
  • $147.55 eCPM
  • 26,651 requests
  • 9,300 impressions
  • 34.90% fill rate
  • 11.80% CTR

From this we can glean quite a bit.

First, note that Apple’s using a rather favorable calculation of ‘eCPM’, basing it on impressions actually served, not ad requests made. Given that 17,351 requests (65.90% of the total) resulted in no ad shown, I’d say the actual eCPM should be $51.49, not $147.55. That’s still impressive.

Second, these numbers are consistent with a 60% revenue share to the developer and a $10.00 CPM + $2.00 CPC cost to the advertiser. 9,300 impressions x 11.80% CTR = 1,097 clicks = $2,194 in CPC charges to the advertiser. 9,300 impressions x $10.00 CPM = $93 in CPM charges to the advertiser. $2,194 + $93 = $2,287 gross. $2,287 x 60% to the developer = the reported $1,372.20. Apple’s charging exactly what they said they would, at least for this campaign.

Third, at >10% the clickthrough rate is freakishly high – iAds are clearly still a novelty for the userbase. If we assume the clickthrough rate drops to a still-high-but-more-standard 1.00% and also assume that pricing stays the same, that works out to an $18.00 eCPM – or, if you take the unfilled inventory into account, a $6.28 eCPM. Still pretty good for independent developers. Not so good for popular, large brands with dedicated sales teams – but excellent backfill for unsold remnant inventory, assuming you can keep it from sabatoging direct sales.

Fourth, at $6.00 CPM + $1.20 CPC to the developer, just a slight improvement in clickthrough rates drives substantial revenue. There’s going to be a lot of iAd units placed really close to the game controls in the near future. At those rates, I suspect click fraud is also going to become an issue.

Fifth, I’ve read that Apple has $60MM in sales commitments. If clickthrough rates are averaging (say) 10% at the moment, that works out to $210 in charges to the advertiser for every 1,000 impressions – $10 in CPM, $200 in CPC. Let’s round those charges down to $200 to keep things simple. At that rate that $60MM buys 300MM impressions and 30MM clicks. How long is that $60MM going to last Apple? That entirely depends on the popularity of iAds – but note that in May 2010, AdMob did 2.75B impressions in the United States on iPhones and iPod Touches, and right now about a third of all traffic is on the iAd-compatible iOS 4.0. If the much-higher-paying iAd runs as many impressions in July as AdMob did back in May, that 10% CTR will burn through the $60MM halfway through the month. No wonder fill rates are low. Keeping up with demand and getting fill rates up is going to keep Apple’s sales team busy.

My biggest takeaway from these iAd numbers? Run, don’t walk, to get iAd-supported applications in the AppStore – it’s clearly ‘put the money in your pocket while the users and advertisers are still behaving irrationally’ time, and it’s not likely to last.

UPDATE: Naturally, right after I write this I see Kenneth’s post with much lower numbers. If the pricing’s consistent, his average of $10.00 – $15.00 eCPM (let’s call it $12.50) implies about a 1.00% CTR. In addition, his fill rate’s far lower – sub-10%, which brings iAd in line with every other ad network out there. I think he’s wrong about his numbers including Apple’s 40% cut, though, since I could only make the above developer’s payout line up with Apple’s announced pricing by adding the 40% to the developer’s figures.

UPDATE 2: Thinking more about the above. The developer with the >10% CTR apparently got it from a flashlight application. If the phone is pointed away from the user in order to illuminate something, of course there’s going to be some accidental clicks on an ad unit. (And incidentally, not so great performance for the advertiser.) So I wouldn’t be surprised if CTR was artificially high, and everyone, including me, is wasting cycles over a freak outlier. But if it’s Kenneth’s CTR that’s typical, the fill rate seems freakishly low – at 1% CTR they’d only be charging the advertisers $30 for 1,000 impressions and 10 clicks, so $60MM would pay for two billion impressions and 20MM clicks, certainly enough to meet more than 10% of the demand.

A few possible explanations for low fill rate: a) the ad campaigns are tightly targeted, so 90% of the audience just isn’t eligible, b) the ad units are tightly frequency-capped, and Kenneth’s app is used much more often per-user than the flashlight, or c) the $60MM in sales includes a lot of campaigns slated to start later, and the complicated ad creative just isn’t ready yet. I’d place my money on the last option – iAds might look great, but they don’t look easy to make. If that’s the case, look for fill rate to increase slightly over time as more ad campaigns come online, but for Apple to ultimately have a lot of trouble keeping up with demand as agencies struggle with creative creation.

{ 1 comment… read it below or add one }

matt July 8, 2010 at 2:36 pm

You are mostly right. Point 1 is wrong though, since the other networks I use all report eCPM based on impressions served, not requests. Your computation in Point 2 concurs with my iAd earnings to date.

This guy is the freakish outlier, as you say, and typical for a kid in a candy store, just publishing the numbers for more publicity. This is of course in violation of the iAd agreement.

Kenneth’s numbers are closer to reality of several days’ reporting, but you are correct about computing Apple’s share.

Leave a Comment

{ 1 trackback }

Previous post:

Next post: