Interview: AdMob CEO Reveals Stats & Plans For An "Ad Lab" With Apple; Provides Sure-Fire Cheat Sheet For Novice Publishers

[The Msearchgroove exclusive mobile advertising podcast series in close cooperation with VisionMobile continues with AdMob CEO & Founder Omar Hamoui. Special thanks to Paul Nash, creative director at fifty50, for designing the cheat sheet for mobile publishers based on input from AdMob’s best & brightest. Paul, together with his colleague matt Harper, is also driving our new design, and a long list of innovations in navigation and site usability that really rock!]

AdMob has gained some serious traction since it broke on the scene just two years ago, building up the scale and the clout to take on major league players including Google, Microsoft and Yahoo. In fact, AdMob effectively enables the world’s largest mobile advertising marketplace, having just reached its mega-milestone of delivering over 5 billion ads in the last 20 months one billion of those in the last month. (We covered it here and calculated AdMob generally serves more than 370 ads per second.) Since then the company has quietly and cleverly launched a new advertising unit focused on the iPhone. In plain text, AdMob ad servers will recognize iPhone users and serve up iPhone-specific advertisements, paving the way for the company’s network of 2,000+ publishers to monetize their iPhone traffic and develop iPhone-specific content.

And speaking of reach, AdMob has sewn up a slew of deals most notably a tie-up with CBS Mobile, which made AdMob one of its “gang of four” advertising-enabling companies to provide clients ad options ranging from text and banner ads to video interstitials. More recently, AdMob struck a deal with Contec Innovations to deliver ads via Contec’s BUZmob mobile publishing service. Of course, enabling mobile advertisers to place more ads with more publishers creates inventory for all and solves the bottleneck of having more ads to serve than sites to show them on

I caught up with Omar Hamoui, AdMob CEO & Founder, to talk about the stellar stats, the competitive landscape and pose the all-important question: When are you going to be acquired and by whom?! The answer: AdMob will stay an independent company for now, but would consider the option more seriously if and when it runs out of steam. (Not likely to happen any time soon since Omar told me off-line that he predicts “on a revenue basis [by next year] we’ll make 6x what we did this year. If I annualize our revenue last year, the year was probably 10x of what we did last year.”)

Listen to the podcast here. [23:51]

And a special feature for novice mobile publishers: AdMob agreed to share the secrets of site optimization with our readers. It’s free and fact-packed and you can download it here.

Numbers that matter: First, AdMob has hit a new high, serving up 1.5 billion ads per month, and AdMob does not deal in adult period. What are the stats beyond ads served? “Anywhere from 1 to 3 percent are clicking on the ads. There are ads that perform less than that, and there are ads that perform better than that. But I would say that’s a good middle-ground in terms of what click-through rates are.” Beyond that “anywhere from 5 to 10 percent engage with the advertiser or make a phone call In many cases, if the content is free or you’re signing up [users] for free services, then it’s higher than that.” AdMob also collects reams of information on usage, everything from the user’s geography to the device type. It’s great stuff, and Omar is mulling over how to dice and slice it for us all to peruse. In the meantime, here are a few surprises:

The U.S. accounts for 40 percent of AdMob’s network; South Africa is also upbeat on mobile advertising a phenomenon that likely benefits from flat rate data plans and aggressive mobile operator mobile data promotions
U.S. carriers that pack them in are Boost and Metro PCS – as Omar puts it: They “highly over-index in terms of the percent of users that we re seeing from them.”
40-60 percent of devices on the AdMob network can do mobile video and could therefore receive video ads (probably one reason delivering them is high on AdMob’s agenda )

Partners that pay off: Granted, AdMob’s model doesn’t give operators a cut of the action (or revenues), but Omar can imagine how the shift to off-portal could change all that. “Operators have a unique advantage over anybody. They have data, and that data for demographic information, as well as just user behavior, can be used by advertisers. Even if it s not uniquely identifiable, it can be used by an advertising company such as ourselves to better target our ads. We could increase our CPMs, we could increase our CPCs, and they would rightfully, or should rightfully, be able to charge us for that. If there was some sort of cross-operator standard for providing non-uniquely verifiable data to trusted partners, we would absolutely participate in that and we d be willing to provide a significant amount of the uptick in revenue to the parties that are providing the data. I think that s a really easy example of how we might be able to work with them.”

Mobile search that delivers:
AdMob bumps up against Yahoo so they are competition for ad dollars. But all mobile search companies could be a boost to AdMob it’s a mechanism that points people to new and interesting sites. Nothing specific at this time, but interest is high. “We don t really have any sort of explorations going at this point but there s certainly an opportunity there for us.” Content discovery is a different matter. AdMob will likely announce a tie-up with an on-device portal company in “the next 3 to 6 months.”

Inventory that counts: “We ve gotten to the point where there s a lot of traffic, a lot of page using and more than you can fill with brand advertisers at this point, but what s important to the brand advertisers is not just the volume but the actual content, and I would agree that if you re looking at sort of the high-end sites like the CNN s or ESPN s, or anything like that, then there is certainly not yet a viable amount of content for advertisers to spend a huge amount of money.” He continues: “It s not the fact that it s not available for them to run decent sized campaigns, but what everybody s interested in is just much more content and a lot more standardization so that they know that if they really want to turn up the dials and go up to a huge spend, they d be able to do that without encountering some sort of ceiling on inventory. I wouldn t say we re at the point where demand for advertisers for inventory is so high that their bidding price is up to crazy levels or anything like that. I think it s at a decent point where pricing is still reasonable and advertisers are getting most of what they want, although everybody would want it to be more, if possible.”

R&D that rocks (!): So what’s the deal behind the cryptic iPhone announcement? The new AdMob unit “definitely falls more into the R&D category for us.” He adds: “It also allows us to do some really interesting data analysis and we ll probably be talking about that more in the next month or two, but there s a little bit of an ‘AdLab’ going on with the iPhone right now for us and so it s more than just kind of the overall reach because there s a number of devices out there relative to everything else that s still small. It s not a huge revenue driving issue right now; it s more of a learning as well as a R&D issue for us.” The bottomline: Omar can’t go into detail right now but basically it’s about “learning things that are going to help our ads on all the other phones.” Count on Andreas and me to keep a close watch

Next week we continue the series with Marc Henri Magdelenat, Screentonic CEO, who will discuss the mix that makes for a successful mobile advertising campaign, the tie-up with Microsoft and the role of mobile advertising.

EXCLUSIVE: Where’s The "M" In CPM? Millennial Media VP Talks About Targeting, Reach & Mobile Search

Continuing with our mobile advertising podcast series, we take a look at Millennial Media, the cross-platform mobile advertising company that covers all the bases and then some. I’m thinking here of its broad offer that includes Millennial Motion (encompassing animation and allowing brands to combine audio, video and interactivity) and Decktrade, a performance-based auction service and marketplace for advertisers and publishers. The company has been on my radar since it raised a $6.3 million first round from Bessemer Venture Partners, Columbia Capital, and Acta Wireless in January. Another eye-opener was the partnership with CBS which we covered here. (The four-way tie-up also involves AdMob, Rhythm NewMedia and Third Screen Media.)

I caught up with Eric Eller, Senior Vice President of Products and Marketing, to talk about Millennial’s roadmap, as well as the roadblocks that have so far kept spending and conversions down to such modest levels. Sure, Informa Telecoms & Media forecasts worldwide spending on mobile advertising could reach more than $11 billion annually by 2011 but a lot has to happen first. Eric identified three things that have to change fast.

Reach More people have to use the mobile Internet. “It s about forty million uniques in the world today, versus 180 to 190 million online, so we re still 20 to 25% of online reach and that has to increase.”

Standardization “The mobile experience varies greatly from carrier to carrier, from platform to platform, and even though companies like ours address that layer for the brand and agencies, there still is a huge difference in the consumer s experience.”

Metrics – “The more data that we can get, the better .And more effective programs can be developed to really achieve the targeting and the results that the brands want.”

Listen to the podcast
here. [20:42]

The dream team: It’s all about CPM but without the proper measurement it’s an empty victory. Against this backdrop, it’s the measurement companies so the likes of M:Metrics and Telephia that can fill the gap and deliver the analytics that can move the market forward. “Clicks are a good metric but, you know, what people in the online space have found is that clicks aren t necessarily indicative of brand impact or of conversion response so we ll have to learn in the mobile space what the correlations are between those different metrics .But I think they re all measurable and they ll become more measurable over time.”

Beyond our reach: The ability to target ads is a few years off but that’s no reason to neglect undertaking the detailed research that is a prerequisite for successful campaigns. “We [as an industry] need to aggregate reach; we need to definitely partner with the companies that are going to ask those questions to that broad reach of people and start building that dataset and understanding what the audience makeup is of the different types of publisher inventory, of the different types of carrier subscribers, of the different types of people that have common behaviours. Then we ll get much closer to being able to offer the different levels of targeting that most advertisers want to buy.

The numbers game: Conversion is always a sticky issue, but Eric offered more detail than most. His insights: “We certainly have seen campaigns where the click through rate is 5 percent and then 50 percent of those convert to buy and that might be for a promotional type conversion right. So, if you need to sign up for a chance to win something, those conversion rates are always very high because the consumer in some cases is incentivized to get those low level conversions so that they can be marketed to in the future. I think it s reasonable to expect that s a relatively low value conversion and it s easy to obtain. A higher value conversion, having someone subscribe to a $10 a month subscription, is a little bit more difficult and then, kind of at the higher end kind of following at the mobile commerce side, if you want someone to download an $80 smartphone application, that s a bigger conversion yet. I think the range of click-through rates is anywhere from 0.5 percent to 5 or 6 percent, depending on how well targeted the campaign is to the audience and the conversion rates can vary from 2 to 10 percent, or up to 50 percent for a promotional type conversion.”

Coming together: Mobile search and mobile advertising are a perfect match, but Eric doesn’t just state the obvious; he gives us a blueprint for how this might work in practical terms. As he sees it, the industry is in a testing phase and search isn’t really about search, it’s about tracking. “Since we re crossing multiple advertising mediums already through applications, through SMS, the search space can be considered a different medium, a different way to reach an audience.” In this context, “we may have a service or solution that allows an agency to buy through us and reach audiences in applications, audiences on search traffic, so in some ways you might look at that as a search engine marketing company in the online space. But there might also be some more automated hooks that allow us to deliver a combined campaign across as multi-medium including search.” (BTW: Yes, I asked. No mobile search company is a suitor for Millennial yet and so far the company is unaffected by the run on mobile ad companies and inventory.)

Motion matters: It’s all about banner and text links now, but Millennial’s Motion product, which enables advertisers to deliver animation, is gaining traction. It’s all about engaging audiences and vector animation is rising up to be part of the marketing mix. “We’ve had a lot of positive feedback from agencies on the richness of the environment that can be achieved when a user clicks on a [graphic] and goes to a motion environment where it s very similar to a web-based animation . Advertising becomes entertainment that can be interactive menus, a virtual showroom, streaming video .It’s almost like having DVD extras [in the content bundle], so that s very exciting.”

What’s next?: Predictably, Eric skirted around the specifics, but he did outline the areas that figure most prominently in Millennial’s roadmap. The focus for the rest of this year is “growth and scale.” So we can expect to see more publisher partnerships like CBS. The other objective: “An aggressive role out of the Millennial Motion product because it s primarily geared towards application inventory. We ve had a number of developers that are kind of in the process with that today and expect to see that inventory growing rapidly towards the end of the year.”

Peggy Anne Salz

VisionMobile teams up with Peggy at Msearchgroove – the premiere thinking-space at the intersection of content and context – to present this in-depth series of exclusive audio interviews.

TOMORROW: The mobile advertising podcast series continues with Omar Hamoui, CEO of AdMob, a company that serves a whopping one billion ads per month. More on the company’s future plans and a fact-packed cheat sheet for mobile publishers – so watch this space!

i-mode: the beginning of the end

i-mode, once the mobile industry’s role model for data services has been dealt with three swift blows in July from O2, KPN and Telstra. The implications of these announcements have been much debated, some industry pundits arguing that i-mode is fading in Europe, some hinting at the end of the i-mode alliance and some arguing that i-mode isn’t dead yet.

Lack of economies of scale in handset development and purchasing has been the root cause for the gradual demise of DoCoMo’s i-mode global expansion plans. But even if this is the beginning of the end, I suspect DoCoMo has new expansion plans. Let’s take things from the top.

Made in Japan
i-mode is perhaps the most succesful data services model, a success which many operators have tried to mimic over the years. i-mode was launched in February 1999 by DoCoMo and reached more than 47 million subscribers (an impressive 90% of the subs base) as of March 2007, based on DoCoMo’s financial report (.PDF). Since 2002 i-mode has been gradually exported to 17 EMEA countries reaching 7.2 million subscribers according to DoCoMo’s website. The next table summarises the countries where i-mode has been deployed (source: DoCoMo).

i-mode global operators

The beginning of the end for the i-mode alliance
Despite the success of i-mode in Japan, the i-mode alliance has not done done as well as DoCoMo would have liked. While in Japan service penetration exceeds 90% of the subscriber base, in the remaining 17 countries service penetration is typically less than 5% (an average of 3% in 2005 according to i-mode business strategy). The only exception is Bougues Telecom who had 1.6 million subs or just under 20% penetration as the operator reported in September 2006.

In July three operators announced separately that they would be discontinuing their i-mode service, dealing a blow to the success of the i-mode alliance. KPN announced it would not source any more i-mode handsets. Australian operator Telstra said on its website it would stop the i-mode service on December 2007. O2 UK said it would phase out its i-mode service by mid-2009.

Interestingly, O2’s i-mode was reportedly used by only 250,000 subsribers in the UK, or less than 1.5% of the subs base. This is particularly bad for O2 UK since the operator had reported that i-mode users consume twice as much data per month as they do on comparable WAP services. Note that O2 in 2004 had abandoned plans to rollout i-mode services in Germany.

A year earlier, Russian operator MTS and Israeli operator Cellcom had also announced they would discontinue investments in i-mode, according to Informa (here and here – subscription required).

The failings of the i-mode alliance
The reasons that have led to the demise of the i-mode global expansion plans are several:

1. Firstly, the gamut of handsets supporting i-mode is extremely limited. In the case of O2 UK, according to David Nicholas, the head of communications for O2 in Europe, the operator offers only 12 phone models with i-mode but more than 240 with conventional Internet browsers. The case of Cosmote Greece is very similar; on the operator’s website there are currently 13 i-mode devices on offer (from LG, Motorola, Sagem, Samsung and Sony Ericsson) from a total of 203 devices available on contract (a tiny 6%).

These examples are the rule in i-mode alliance deployments. The reason for this lack of handsets is that i-mode operators have each produced different specifications for their handsets, breaking the economies of scale that could have been reaped (as the alliance had tried with a procurement agreement with LG according to Informa – sub required). Furthermore, i-mode handsets released in Japan carry more advanced software – DoJa 4.0 specification instead of the DoJa 2.5 used in i-mode handsets outside Japan – meaning that i-mode handsets cannot be simply exported. Perhaps more importantly, handset OEMs have resisted releasing handsets where they have no control over service delivery (and monetisation) – the reason for Nokia’s absense from the list of i-mode handset OEMs.

2. The development and release of an i-mode handset takes anywhere from 6-18 months according to a presentation by Bouygues Telecom in September 2006 at the MAPOS conference. Due to the small ordering quantities, handset OEMs cannot prioritise handsets which are i-mode variants, thereby creating an agonising delay in bringing i-mode handsets to market. The first weeks of a handset launch are absolutely critical to sales performance – handsets sell mostly when they are new and cool, after which sales rapidly decline – thereby hurting i-mode subscription growth. Lack of economies of scale is therefore to blaim again.

3. The rev share has deterred some operators (particularly US operators) from deploying i-mode services. O2 offered a 86/14 revenue split while Telstra offered an 85/15 split – both close to DoCoMo s 91/9 split in Japan, but much less interesting for operators compared to premium SMS revenue splits (typically 40/60 in favour of the operator).

4. The investments in the i-mode portal compete with operator investments in other service portals, including SMS, WAP, HTML and HTML transcoding portals. Given that i-mode content provider deals have been local and not pooled within the i-mode alliance, the lack of economies of scale are again to blaim. This is even more so given that supporting third parties with cHTML content development and non-WAP portals require operator investment.

i-mode 2.0 ?
The i-mode alliance is bound to fail eventually without economies of scale, particularly at a time when operator strategies are geared towards opening the mobile phone to the entire internet and not to a ‘sticky’ garden. So what are DoCoMo’s plans for global expansion now ?

In June 2006 DoCoMo established the LiMo foundation with co-founders Vodafone, Motorola, Samsung, NEC and Panasonic – since August more software and equipment vendors have joined, including LG. The purpose of LiMo is to define a complete handset software stack which will support advanced data services. The Linux-based stack will be co-owned and co-developed by LiMo members and will launch in phones towards the end of 2008 (or more likely early 2009).

I believe that in LiMo DoCoMo sees a new export strategy for mobile data services, based on a common data services enabling software. Given that Motorola and Samsung are on board, economies of scale will be easier to achieve in most markets. Moreover, the base software stack and over-the-air software management will allow easier development and launch of i-mode variant handsets by major OEMs.

Is this part of the plans for an i-mode 2.0 ? I believe so.

[update: Informa s Mobile Communications Europe magazine reported in early August that there were just 5.2 million i-mode subscribers outside Japan at the end of Q1 2007, more than five years after operators began launching in early 2002; these figures are 2m subscribers less than the figure reported by DoCoMo at its i-mode alliance website. Informa s figures put i-mode subs to just 3.5% of customers of operators offering the service.

Bouygues Telecom has by far been the main exception to the rule, having the highest penetration of customers with an i-mode handset (20.6% of 8.7 million users). However, Informa reports that only half of these subscribers (about 10% of the subs base) are active users of the i-mode service.]

– Andreas