Where will the next ad network breakthrough come from? September 24, 2009Posted by jeremyliew in ad networks, advertising.
I’ve noted in the past that the four core competences of ad networks are:
Better targeting has historically been one core area of competition for ad networks, especially those focused on direct response advertising. However, as Anand Rajaraman (co-founder of Kosmix, a Lightspeed portfolio company) points out, more data usually beats better algorithms. Andrew Chen recently noted that after three years of work, Netflix awarded its $1m prize to a combined team of experts for an algorithm that only improved targeting by 10.5%:
This means if you combine dozens of the best machine learning people in the world, some of the cleanest datasets, you get a measly 10.5% increase. Compare this to starting a new ad network where you end up with noisy datasets, lots of crappy traffic, and a small team looking at the problem – that’s not an easy path to disruptive change. In general, 10% is not a big enough number to counteract the other economic drivers in the ad market, which revolves around better deal terms, a larger selection of advertisers, better ad inventory, etc.
Note that this observation comes from a guy who was a co-founder of Revenue Science’s Ad Network business!
While I agree with Andrew in principle, I think that even a 10% edge in targeting can be enough to build a competitive advantage in the direct response world. Because competition for publishers is fierce, and publishers switch ad networks frequently in search of higher RPMs, a slight edge in targeting can lead to a slight edge in publisher payouts which can lead to an overwhelming win in volume.
Andrew thinks that breakout ad network performance will come from two of the other key competency areas:
I think disruptive change will come not from algorithms, but rather two other areas:
* Better ad inventory: New websites and mechanics emerge all the time, and who knows what happens when you put ads on them? It was clear, until they tried it, that with the right ads search can be >30% clickthrough rates or more, which is unheard of.
* Better data: The other big opportunity is in using specialized data to drive your algorithms – rather than basing everything off of domains, cookies, and ad impressions like everyone else, there may be ways to extend the targeting to unique datasets that no one has access to. This is what’s happening in the world of retargeting.
These are good thoughts, and well worth exploring. Better ad inventory can be difficult to defend in an age of exchanges like Right Media and the Doubleclick ad exchange. However, in some areas such as mobile, video, in-game advertising and client driven inventory, it is still possible.
Data is also improving. But because it is also becoming more of a commodity, the real question will be whether this data can be proprietary. If the proposed FTC rules on third party cookies for behavioral targeting take effect, it could give some of the big web properties access to their own proprietary targeting data that will give them advantages over third party networks. Taking offline data and using that for online targeting is also another possibility.
In the brand advertising world, I think that sales will be a real differentiator. The big brand budgets are just starting to move online. CPG, one of the core categories for brand advertising, is starting to shift online this year in a meaningful way. But brand advertisers need to be sold to the way that they want to buy. Not all online sales teams know how to do that. Facebook’s recent partnership with Nielsen to show brand lift means that now only four online media companies have the ability to show the impact of a campaigns effectiveness on brand metrics (Yahoo, AOL, Facebook and Brand.net). I expect more companies to start reporting these sorts of brand lift metrics as a matter of course if they want to take their share of brand advertising dollars as they move online.
Which new startups do you think have a breakthrough in any one of these areas of core competence?
Last month I raised some concerns that the government could make monetization even harder for online ad networks and publishers through limiting their ability to do behavioral targeting. The pressure to do so is rising as the NY Times reports:
On Tuesday, 10 major privacy groups plan to demand new privacy legislation from Congress regarding online behavioral tracking and ad targeting.
The roster of groups is a who’s who in consumer and privacy circles: Consumers Union, Electronic Frontier Foundation, Consumer Federation of America, Center for Digital Democracy, U.S. Public Interest Research Group, and others.
Among the things they’re asking for: No sensitive information (like health or financial information) should be used for behavioral tracking, no one under 18 should be behaviorally tracked, Web sites and ad networks shouldn’t be able to keep behavioral data for more than a day without getting an OK from the individual they’re tracking, and behavioral data can’t be used for discriminatory purposes.
While it is always hard to argue against privacy, the impact of this level of restriction would be enormous for companies relying on online advertising. Financial services and pharma/health are two of the leading categories for online advertising; the youth demographic is highly attractive to many advertisers, and limiting behavioral targeting to one day without an opt in severely restricts the usefulness of the data.
I’ve spoken to a number of people at venture backed ad networks, and it is clear to me that more needs to be done to organize feedback to the FTC and congress about the proposed rule changes and legislation.
Government could make monetization even harder for online ad networks and publishers August 13, 2009Posted by jeremyliew in ad networks.
In the last few years, behavioral targeting has gone from being an interesting experiment to core to the success of many ad networks. Many networks are using data collected from one site to improve the targeting of advertising on other sites. But in the past two months, both the FTC and some members of the House have discussed limiting the ability for online advertisers to do behavioral targeting.
Businessweek notes about the new FTC Chairman:
Leibowitz wants to terminate—or at least rein in—… delivering ads to individuals based on the Web pages they visit and searches they carry out. Appointed by President Barack Obama in February to run the country’s top consumer watchdog, Leibowitz has made so-called behavioral targeting a top priority.
Leibowitz is not content with advertising industry self regulation:
… Leibowitz hints that he’s growing impatient with marketers’ efforts. “It’s not clear that they’re moving far enough or fast enough, even though they’re making some progress,” Leibowitz says. He supports the controversial approach of making more of the targeted ads on the Internet “opt-in”—meaning they would require consent from Web users before collecting data—and is in talks with members of Congress intent on drafting legislation for online ads.
The member of Congress that he is talking to are members of the House of Representatives’ Subcommittee on Communications, Technology, and the Internet, who are also keen to limit behavioral targeting. Businessweek again:
Representative Rick Boucher (D-Va.), who chairs the House subcommittee on communications, technology, and the Internet, has stated publicly since February that he plans to draft legislation on targeting practices this year. He says sites should maintain plain-language privacy policies, visitors should be able to opt out of data collection, and any third-party companies working with publishers must obtain permission from Web users before acquiring or using their information.
This position has bi-partisan support. Notes Paid Content:
His counterpart on the committee, Rep. Joe Barton (DR-Texas), also said he appreciated the relevance of targeted ads, but he was dismayed at how much information is collected about him on the web without his knowledge. Barton: “I hit the delete button every week and erase the cookies on my computer. I’m always amazed at how much information is taken from me. I think I have the right to know what information websites are gathering about me and what they’re doing with it. And poll after poll shows that the public agrees with me.”
The current FTC guidelines on behavioral advertising call for the principle of “Transparency and Control”:
Every website where data is collected for behavioral advertising should provide a clear,
concise, consumer-friendly, and prominent statement that (1) data about consumers’ activities
online is being collected at the site for use in providing advertising about products and services
tailored to individual consumers’ interests, and (2) consumers can choose whether or not to have
their information collected for such purpose. The website should also provide consumers with a
clear, easy-to-use, and accessible method for exercising this option. Where the data collection
occurs outside the traditional website context, companies should develop alternative methods
of disclosure and consumer choice that meet the standards described above (i.e., clear,
prominent, easy-to-use, etc.)
This princple is reasonable. But as noted in italics in the quotes above, what Boucher and Leibowitz are talking about is to move beyond this principle and essentially establish an opt in process for third party cookies (which would include the cookies for all ad networks).
Taking away the ability to do behavioral targeting and retargeting would reduce overall industry eCPMs. This would impact both publishers and ad networks by reducing their revenue. It would make it harder for advertisers to target their customers, resulting in overall higher customer acquisition costs. And it may even lead to more advertising in general as publishers try to make up for lower eCPMs with more ad units, which will have an impact on user experience.
Companies who rely on contextual targeting and content adjacency to sell their advertising have little to worry about as they do not use much third party data today. Examples include well know brands (e.g. NYTimes.com) and sites with endemic advertisers (e.g. WedMB). The big portals (e.g. Yahoo) and search engines (e.g. Google) who see a high enough proportion of all web users to be able to use first party data for targeting will also have little to worry about. TBut many ad networks, and the publishers who rely on ad networks for a substantial proportion of their revenue, should be closely watching the positions of both the FTC and the Subcommittee on Communications, Technology, and the Internet.
More ad networks or less? April 6, 2009Posted by jeremyliew in ad networks, advertising.
A closer look at the distribution of ad spend by Razorfish clients reveals several trends, including:
– An increasing reliance on ROI and proven channels like search
– A continued shift of budget away from portals
- Renewed fragmentation in the ad network space
Despite the drive towards increased efficiency because of the recession, ad networks as a category saw only a slight increase in share year-over-year. One trend reversal we saw was in the concentration of spend amongst the top five ad networks dropping to 62% from 76% in 2007. A few things contributed to this change in direction. The first is a rise in spend outside the U.S. and the development of branded networks such as Forbes, Turner Entertainment and Fox Audience Network, and the move of many premium advertisers away from general networks. Additionally, the rise of specialty vertical networks like the community sites BuzzLogic, Six Apart, Lotame and BlogHer has further fragmented this category and put a refocus on testing the emergent opportunities.
But at the same time as Razorfish is seeing more ad network diversification, they are predicting:
4. Online ad networks will contract;open ad exchanges will expand
In 2009, the online ad network world will see both contraction and expansion:
• The traditional ad network world will contract as competition for declining ad dollars increases. There are simply too many broad networks competing for the same inventory and not telling a new story.
• At the same time, branded networks will expand. Large publishers (e.g. the Fox Audience Network and Turner Entertainment) will continue to take back control of their inventory and monetize it themselves, or they will work with fewer ad networks to ensure quality and maximize value.
• Expansion will also come in the form of Ad Exchanges like Right Media, DoubleClick and AdECN, which are newer open markets for online ad inventory that increase buying efficiency by delivering unprecedented transparency in the process. Development of this ecosystem will put further pressure on small and mid-tier ad
networks to survive. If Ad Exchanges are widely adopted, it could revolutionize how online media is bought and sold.
So which will it be, more ad networks or less? Most pundits are predicting less. However, I believe that there will be more. The fourth generation of ad networks are living in an environment where access to inventory is getting commoditized (through ad exchanges), data for targeting is getting commoditized (albeit slower, through companies like Lookery and Blue Kai), and targeting algorithms are turning out to be not as effective as previously thought (more data usually beats better algorithms). In this instance, sales execution becomes the key differentiator. And sales teams typically work best when they can focus on a set of accounts with a lot of commonality, whether demographic, industry, or geography. This means that it will be easier (not harder) for smart small teams of sales people to start their own targeted ad networks. We’re already seeing some of this as Razorfish notes above.
I think we’ll see more ad networks, not less.
What do fourth generation ad networks look like? February 25, 2009Posted by jeremyliew in ad networks, advertising.
There has been a proliferation of online ad networks over the last decade. There are three distinct generations of ad networks, and they have each excelled at a different part of the value chain:
First Generation: Controlling Inventory
The first generation of ad networks were built on their ability to aggregate and control inventory from a wide array of websites. They did a terrific job of building publisher relationships to be able to bundle together wide reach (even within a channel) and offer this as an efficient way for advertisers to buy ads. In some cases, this first generation of ad networks integrating themselves directly into their publisher sites by supplying their ad server or other elements of their advertising infrastructure. Most of the biggest ad networks are good at doing this.
Second Generation: More data
The next generation of ad networks came up with the innovation of third party cookies. They dropped pixels on their publishers pages in order to be able to track users across all of the sites in their network, and to start to target advertising based on recognizing a user when they showed up on different sites.
Third Generation: Better Targeting Algorithms
The third generation of ad networks pioneered behavioral targeting. Not only were they able to recognize a user across their network, but they could begin to predict which users had a greater propensity to click on a particular ad based on their past web surfing behavior.
Together, these three elements represent three of the four core competences of ad networks:
We’re starting to see a few changes in the market that are going to serious affect the relative importance of these factors. Firstly, the ad exchanges (Right media, doubleclick exchange etc) are rapidly commoditizing access to inventory. Networks with publisher relationships as a core competence may find that this is less of a competitive advantage going forward.
Secondly, a new generation of startups including Lookery, BlueKai and others are commoditizing data. Ad networks and advertisers can now buy fairly detailed demographic and behavioral data on users, and simply watch for those users to turn up on media that they control. They can even buy cheap impressions from the ad exchanges and enhance this with the data that they bought.
This places additional emphasis on the two other core competences of ad networks, targeting and sales.
Performance ad networks who have targeting as their core competence are going to be safe for a while. Performance advertisers don’t care how your “black box” targeting algorithm works, as long as you’re able to hit their CPA targets.
However, this is less true for brand advertisers. A “black box” approach to targeting brand advertising (unless there is a performance component to their campaign that they can measure) simply isn’t going to work. Advertisers won’t just trust your algorithm. As a result, the targeting that they are looking for is typically not algorithmically complex, but simply a repurposing of demographic or behavioral data (e.g. women 18-35, auto intenders). For brand ad networks, algorithms are not going to be a differentiator.
That leaves sales. It is somewhat obvious, but sales must always be the core competence of the fourth genearation of an ad networks.
Would love to hear readers thoughts.
Four flavors of ad targeting July 7, 2008Posted by jeremyliew in ad networks, advertising, targeting.
I recently had lunch with Iggy Fanlo, CEO of Adbrite. He is one of the most thoughtful people I know in the online ad business and I always enjoy our conversations. He related to me how he sees ad targeting falling into four flavors:
He noted that the first two of these, geographic and demographic, are black and white, and focused on the user. You are in one and only one location. You have one and only one gender, age or income. In each of these cases, the key is to gather a broad dataset with which to target. As a result, the largest sites and networks, with the largest datasets, will tend to be best at these flavors of targeting.
Contextual targeting is also black and white, but it is not user centric. Rather it is focused on the page. You are looking at a page that is about some topic. Search is the easiest case, where the user tells you what the page is about. Vertical ad networks with endemic advertisers are also pretty easy to contextually target because they only include sites within their desired topic. But the general case is much harder. Now the winner isn’t necessarily the one with the largest dataset of users, but rather the one with the best algorithm for figuring out what the page is about.
Behavioral targeting is not black and white, but rather shades of gray. Furthermore, it is both user centric AND page centric because behavioral targeting is the accumulated sum of historical contextual targeting. It is based not on what page you’re looking at now, but rather on what pages you’ve looked at in the past.
In this case there are advantages to both having more information about users past behavior, AND better algorithms. In it’s simplest form, retargeting, a web user who had visited Ford.com in the past will be shown Ford banner ads while on other sites. But ad fatigue limits the frequency with which one can retarget based on a single datapoint. Good behavioral targeting systems need good historical data as well as good algorithms to best manage the portfolio of advertising opportunities to a single user. Companies are using many different sources of historical data, including search history, looking for a user on the ad network, watching a user at the ISP level and even watching offline behavior.
AdBrite recently launched an Open Targeting Exchange where it will let any company with a targeting algorithm bid to be used to target ads across their network. It is a very interesting idea, and I’ll certainly be watching closely to see how it works for them.
Forecasting ad sales for web startups April 3, 2008Posted by jeremyliew in ad networks, advertising, business models, models, start-up, startup, startups.
Andrew Chen has a good post on how to forecast advertising for web startups:
The right way to model out inventory is a number of equations – I’ll pretend that a site has two types of inventory, their “brand” stuff and their “direct response” (aka remnant) inventory:
Brand revenue = # campaigns sold * average campaign size * brand CPM
Direct response revenue = (total impressions – brand impressions) * remnant CPM
Total revenue = Brand + remnant revenue
In an actual forecast, you could get a ton more detail in the brand revenues side, since what you really care about is the # of ad sales people you have, how many campaigns they’re selling per quarter, the size, etc. Again, think of this as an enterprise sell, and treat it as such.
Essentially, he suggests that brand advertising is a function of the size and efficiency of your direct ad sales force (and is demand constrained) while remnant advertising can go to networks and is supply constrained.
As Ed Sim notes about a direct ad sales force:
… many entrepreneurs underestimate the direct capital and management costs necessary to build such a team. In many ways, building a direct ad sales team is similar to building an enterprise sales team. These thoughts may seem quite basic to you but here they are nevertheless. First, don’t ramp up your sales team too quickly until you have a product to sell. That means if you don’t have scale or enough eyeballs you are better off using Google Adsense. If you don’t heed this advice you may quickly burn yourself out of business. Secondly, I know that many startups may not know what kind of ad units to sell but be careful of not having a standard product list or rate sheet when you go out to the market.
This advice can be difficult to follow in a new market where there are no standard product lists, which is why new forms of advertising are hard.
2008 Consumer Internet Predictions December 3, 2007Posted by jeremyliew in 2008, ad networks, advertising, casual games, Consumer internet, games, gaming, mmorpg, predictions, semantic web, social media, social networks, structure, user generated content, video.
Last year I made some predictions about the consumer internet in 2007 and they were at least directionally correct. So let me take a crack at 2008. Regular readers will not be surprised at some of my predictions as they are themes that I’ve been talking about for some time. Later in the week my colleagues will take a crack at predictions for Mobile, Infrastructure and Cleantech.
1. Social Media advertising, Online Video advertising and In-Game advertising start to become scalable.
Social media, online video and games are at early stages of development as advertising vehicles. Even more than the internet at large, a disproportionately small percentage of advertising dollars are being spent on these three media relative to time spent. Some people have even questioned if social media will be a media business at all, or online if video is a good way to monetize.
The slow start is because there are no standards yet in any of these media. If an advertiser wants to buy TV advertising across NBC, CBS, ABC and FOX, they can buy a common unit, the 30 second spot. If she wants to buy print advertising across Time, Fortune, Forbes, Newseek and Businessweek, she could similarly buy a common unit (e.g. a full page ad). But to buy across YouTube, Metacafe and Break, or across Myspace, Facebook and Bebo, or across GTA, Wild Tangent games and Pogo.com games, she needs to buy custom ad units in each property. This makes ad sales look more like business development – she is negotiating not just price, demographics and reach, but also what the actual units are. This is what makes new forms of advertising so hard. All three industries need ad unit standards to be able to scale. Otherwise they will be trapped by demands for customization.
This year, standards will start to emerge in each media. Some candidates for standards include (i) for social media; behavioral targeting, content targeting, demographic targeting or social ads, (ii) for online video; contextual targeting, overlays or pre-roll and (iii) for in game advertising; rich media or product placements. I don’t know which of these candidates will become standards, but I am confident that we will start to see growing support from both advertisers and publishers for the more successful units.
Ad networks will also gain share in each media, helping make the process of both buying and selling advertising easier.
Viewed through this lens, Facebook’s recent Beacon launch and subsequent adjustments are simply early moves towards figuring out what will be the native social media standard.
2. Structured web emerges.
The last couple of years have seen an explosion of user- generated content, across blogs, social networks, social media sites and user reviews. Previously, when most web content was created by editors, there was good structure and metadata around it. As most of the user- generated content has been unstructured, there has been an overall decrease in the level of structure, and hence a decrease in the ease with which people and computers can access and use this data.
But Meaning = Data + Structure. Search on user-generated sites has not been a great experience so far. This year we should start to see some point solutions emerge to help add structure to unstructured data, substantially improving the user experience. This will include both explicit (user-generated structure) and implicit (inferring structure from domain knowledge or user behavior) methods.
3. Games 2.0
Tens of millions of users are now using casual immersive worlds and playing MMOGs. These sites are some of the stickiest on the web, resulting in some of the highest levels of time spent per month online, and indicating that this is becoming a primary form of online communication for some users. Many of these users skew young, and if you believe that demographics is destiny, then you will expect this behavior to spread. The social aspects of these games is key to their popularity
Even more people are playing casual games online. These people often don’t have the ability to commit the time that MMOGs demand. They want to play with their friends, but instead of spending hours online together, they want to do it on their own schedule and in bite sized chunks.
These trends are likely to come together in asynchronous multiplayer games.
Other key drivers of growth for these products will include innovation in business models (free to play, ad- based and digital goods- based models) and channels (in- browser gaming, mobile, widgets).
Note – this post is cross posted to Venturebeat.
Is games 2.0 just around the corner? November 27, 2007Posted by jeremyliew in ad networks, business models, distribution, games, gaming.
Yesterday I put up a post claiming that Web 2.0 has been driven by variablization:
Variablized Development Costs
Variablized Content Costs
Variablized Marketing Costs
Variablized Distribution Costs
It struck me that many of the same dynamics are now starting to apply to the game industry as well.
Variablized Development Costs
While development costs haven’t become variable across the board, web based games are definitely becoming cheaper and easier to build. As new rich media technologies improve (Flash, Silverlight etc), development tools improve (Flex, Laszlo, etc), and reusable game engines become more widespread, it gets easier to for people to build games.
Variablized Content Costs
Just as with the web, user generated content allows for models with dramatically lower and more variable content creation costs. Gaia, Habbo, Second Life and other casual immersive worlds, while not really games, let users create content for each other, and in fact let users BE content for each other. PvP games such as Scrabulous and the Campaign Game are another way that “content” costs are variablized – instead of having to create more levels, a game becomes replayable because against live opponents each game plays differently.
Variablized Marketing Costs
Again, just as with the web, search marketing has created a completely variable channel for player acquisition. Web games and downloadable PC games benefit from this; console games still need to rely on more traditional marketing means.
Variablized Distribution Costs
The two mechanisms that have enabled variable/free distribution online are social network platforms and virality. We’re starting to see games taking advantage of both of these mechanisms, including Attack!, Warbook and Kings of Chaos.
Ad networks and contextual advertising have driven the variable monetization model for web 2.0 companies. We’re still very early in the game for in game advertising networks, but Google is making its first forays in in-game-advertising, and startups like Mochi and Neoedge are also taking up the challenge. Kongregate is building a destination for online casual games where they share ad revenue with independent game designers, a more centralized approach to making monetization variable for game designers.
The other interesting emerging direction for monetization has been free-to-play games with digital goods. Games like Three Ring‘s Puzzle Pirates and K2‘s Knight Online have demonstrated the viability of this model.
I think we’re going to see an explosion in gaming over the next few years comparable to the web 2.0 phenomena; I plan on exploring this topic further over the next few weeks.
Business models for apps and widgets November 16, 2007Posted by jeremyliew in ad networks, advertising, business models, facebook, myspace, open social, platforms, self espression, social media, social networks, user generated content, widgets.
This afternoon I spoke to the Stanford class on Creating Engaging Facebook Apps.
As I said at Web 2.0 expo, building big businesses online is hard work. While it isn’t hard to start an app company, especially as a single developer ($250k in revenue) or even to support a small team ($2.5m in revenue), it gets quite hard to scale revenues to $25m/yr.
Assuming a 5% daily active rate and 3 pageviews per visit, an app developer with a $0.50 RPM would need to get to 926m installs to get to $25m run rate. Compare that to the app with the most installs on Facebook – Slide’s Superwall which has around 20-21m installs. Clearly, broad reach app developers need to develop (i) multiple (ii) high engagement apps [ie higher active rates and pageviews/visit than these assumptions] (iii) across multiple social networks to be able to get close to this revenue target. (RPMs will likely be higher for companies with a direct sales force as well, so the target isn’t quite as high, but you get the point).
Under the same activity and pageview assumptions, an app developer with a $10 RPM would need 46m installs to get to $25m in revenue. Apps with endemic advertising opportunities can easily realize this level of RPM but will still need to be in multiple social networks to get to those levels of installs. It doesn’t make sense to limit your world to being a Facebook app. Social network platforms are avenues for distribution, and app developers should be taking advantage of all of them.
I also did a similar analysis for digital goods business models in the presentation. Here is a link: Stanford Facebook Class presentation