I spent the last two days at the Mobile 2.0 Conference - Silicon Valley. This is my fourth Mobile 2.0 and like the previous three was a first rate event with great speakers and panels and attendance by a global group of mobile entrepreneurs and developers. It was ably organized and run with great attention to detail by Greg Gorman, Mike Rowehl and Peter Vesterbacka with off-site support from Rudy De Waele and Daniel Appelquist .
The event was two days packed with great discussions and presentations covering everything from the future of mobile to the nuts and bolts of mobile app and web development. Rather than trying to cover the whole conference I'm going to do a couple of posts, each concentrating on one of the highlights for me. I'm going to start with something near and dear to every mobile publisher, monetization.
There was a lot of talk about how to make money in mobile at conference. While application developers can charge for apps and upgrades, mobile web publishers must rely heavily on advertising for revenue. Unless you are targeting enterprise verticals, it's almost impossible to get users to for pay for Web content, mobile or otherwise. That means mobile Web sites are generally ad supported. As a mobile publisher I found the “What Have We Learned in Mobile Advertising? " panel particularly interesting. It was moderated by Skyfire's Raj Singh and featured panelists from Admob, AT&T, Yahoo, Velti and Smaato.
It's never been exactly easy to survive on advertising revenue and the recession has made matters worse. I personally witnessed a huge drop in ad revenue from my mobile sites starting in the middle of 2008.
Things seem to be slowly getting better. My site's CPM (advertiser speak for "cost per thousand impressions" - which translates to earnings per thousand for publishers) has been creeping up this year. The upturn was confirmed by Admob's Ali Diab who said that the ad network has seen CPM rising across its publisher base recently.
The panelists were in agreement that mobile advertising spend will continue to rise as the media becomes more mainstream. However publishers and advertising networks need to do a better job of educating publishers and agencies about mobile advertising's reach and opportunities.
While CPM might be up I'm sure that every publisher would love to earn more. The panel several suggestions.
- Yahoo's David Katz stressed the importance of having quality content to attract the higher paying big brand advertisers.
- The panel though that there is a lot of unrealized potential in new formats like video ads, behavioral targeting, local advertising and pay per call. Admob offers video ads on some platforms with good results according to Diab. Behavioral targeting isn't generally available to the average publisher yet but is used heavily by Yahoo on its own mobile properties. AT&T offers pay to call and according to the operator's Michael Rubin it is seeing very high click through rates with pay per call particularly with ads for local merchants.
- Publishers should test and experiment to find what works best for their content and markets. Admob's Diab suggested running A/B tests comparing different ad formats. Ragnar Kuse from Smaato recommended that publishers experiment with various ad networks to find what works best for them. Using a mediation layer like Smaato can help too, particularly for publishers who are already dealing with several networks. Mediation layers attempt to achieve high fill rates and maximize revenue by brokering between publishers and multiple ad networks. For each request the mediation layer's engine attempts to find the network that will deliver the highest paying ad.
Interesting stuff. I've done a bit of testing on might sites and found that image ads tend to work better than text ads. A single ad at the top of the page performs nearly as well as multiple ads and I suspect it's more palatable to users. I tried five different ad networks and found that they all suffer from low fill rate at times and in certain geographical areas. I've created my own crude mediation layer that tries one network and checks to see if it actually delivered an ad and tries another network if it didn't. I think I'm going to try Smaato and see if its results can beat my homegrown solution.