Ashu's Musings

While TV networks have begun to experiment with posting shows on-line, they continue to struggle with both the limited audience size (relative to TV) and monetization. I would argue that both issues are in part a reflection of the fact that that the medium is nascent and in part a reflection of the poor execution by the networks.  Lets look at both issues.

1. Audience Size: According to comScore Media Metrix, ABC.com had 8.5 million unique visitors (across its entire lineup of shows) for the whole month of February, while NBC.com had 7.9 million. By comparison, a single episode of CBS's "CSI" recently took in more than 20 million TV viewers in one night. So may be reach is an issue. Or is it ?
   1. Its been less than a year since the networks began to release content on the web. By comparison, how many many people watched network TV in the first year ? As a reference point, TV broadcasts started around 1939 and it was only in 1948 that the number of households with a TV set crossed the 1 M mark. But just 5 years later, in 1953, there were more than 30 M households watching TV.
   2. Around 140 M people (in the US) watched video on the internet and spent over 3 hours/month doing so (comScore MediaMetrix data for October 2007). Another survey indicates that heavy users (who account for 20% of all users in the US) spend over 14 hours/month watching video on-line (see http://news.softpedia.com/news/Are-You-an-Online-Video-Heavy-Viewer-78807.shtml for more details). So maybe the issue is not the size of the audience available, but more the fact that the networks have not been able to attract audience to their content. What does that say about the future of the networks ?
   3. Personally, I think the real issue is the networks have just not put enough content on the web - many popular shows are not available on-line, and for the shows that are on-line, very few episodes are available. Most viewer comments on hulu.com lament the lack of content, "why is there only 3 episodes out of 14 seasons?????" was a recent post on the ER page on hulu.com.  Networks still seem to apply the "programmed experience" model to the web whereby they can dribble out content on their time-line, but user expectation have changed.
2. Monetization of On-Line Video: While the networks are currently able to charge 40-100% higher (on a CPM basis), the size of the audience coupled with the fact that there are fewer ads per episode make the total numbers small. Here again, I would argue that the networks only have themselves to blame. Their current ad serving model attempts to replicate thee TV experience on-line vs. trying to take advantage of the power of the web. So, what could they do differently ?
   1. Improving relevancy: My experience with abc.com, hulu.com, fancast.com and cbs.com is that the ads I see are almost entirely irrelevant (to me). And to compound this, often, the same ad is repeated 5 times in through out the episode. Does that make any sense on the web ?
   2. Testing multiple ad formats: When will networks realize the the old "interrupt and repeat" advertising model is dying. What about trying over-lays, 3-D ads, "hotspots" (see http://www.asterpix.com/ for an example) and other new ad formats. While some blog post have recently argued that pre-rolls work well for professionally produced content but not for UGC (see http://blogs.mediapost.com/video_insider/?p=168), I think that multiple ad formats can be effective in professionally produced content - in fact, I think that hotspots that tag the stars and what they wear can be the next generation of product placement revenues for studios, and a way to increase consumer engagement at the same time.

The old adage is that 50% of all advertising is wasted. In a world of fragmentation (in media and consumer behavior), I would argue that 80 or 90% of all advertising is wasted. On the web, that does not need to be the case. Precision targeting coupled with innovation around the ad formats could result in CPMs that are 5-10 X of TV rates. And if the networks actually posted all the episodes of most shows, then the networks might even get the 100 M + people (in the US)  who watch video on-line but do not visit network web-sites to give them a try! What do you think ?

In earlier posts I wrote about how the next wave of innovation in On-Line Advertising will be around improving ad effectiveness, what I call, the Display 3.0 wave. In response to that post(http://ashugarg.spaces.live.com/blog/cns!F19091A9C9230F9C!178.entry), I got several questions/comments about the Display 2.0 wave (largely focused on improving market effectiveness) - these questions focused on whether I was implying that the Display 2.0 wave was over and how I would reconcile my comments against the fact that so many large established media companies (e.g. Forbes) are setting up ad networks and the recent wave of yield optimization companies like Rubicon Projects.

Top line, while I think that the Display 2.0 wave still has legs, the easy pickings are gone and the opportunities to build new venture backed start-ups are limited going forward.  Let me address both questions in some detail:

1. Traditional media companies are setting up on-line ad networks: In recent months, Conde Nast, Martha Stewart, Forbes and several other traditional media companies are setting up ad networks to aggregate and re-sell inventory from blogs and niche sites. In the short term, this is probably a viable strategy. The traditional media sites have advertiser relationships, large expensive sales forces and still enjoy a brand premium. They can leverage these assets to capture arbitrage opportunities. Over time, these opportunities will disappear as advertisers become more comfortable with buying audiences via larger more automated networks, and realize that they are paying a brand premium to Forbes (as an example) with minimal value add. The question is how long can these traditional media houses ride this gravy train ? Despite being a passionate advocate of automated media buying, the cynic in me thinks that this train has a long way to go. After all, which media buyer would be fired for buying Forbes, even if what they are really buying is "400 financial blogs" over which Forbes exerts no editorial control ? The IT folks will remember the old adage, "no one got fired for buying IBM"
2. What does the recent wave of Yield Optimization Players like Rubicon Projects and Pubmatic mean for the Display 3.0 thesis ? Both companies provide publishers with free yield management and optimization services and thereby improve market efficiency, ostensibly a Display 2.0 idea. But the big idea here is different - they also collect all the click stream data (for all users across their publisher network) and therefore best positioned to improve targeting and therefore advertising effectiveness. In fact, given that they give away the yield management capabilities in return for unfettered use of the data, one could argue that is actually their core business. Therefore, IMHO, these are both really the first generation of Display 3.0 companies.  In fact, the biggest validation for this thesis (and for the market opportunity for both companies) is that Google recently launched its free Ad Manager (see this article for more details http://seekingalpha.com/article/68652-google-ad-manager-it-s-bigger-than-it-looks) which provides the same service in return for similar data and for "first chair". Off-course, that does raise the question about  having to compete with Google ? But then, would publishers want to  provide Google with access to all their data ?    

Bottom line, Display 2.0 is not dead - in fact, ad exchanges like Right Media are just getting off the ground. Its just that if you are looking to start something new, you are probably better off jumping on the Display 3.0 bandwagon! What do you think ?

In some ways, day 2 was an extension of day 1 with many more discussions around on-line video advertising and ad targeting. In addition, however, day 2 was much more focused on the agency and advertiser perspective. Panelists spoke about the challenges that agencies face as they transform to thrive in the new world and how brand advertisers are responding to the media fragmentation and the control that consumers now exercise over their media consumption.  In my opinion, there were 3 additional over-arching themes discussed today:

1. The "Momentum" effect
2. The shift in power between media and creative agencies
3. The increasing focus on metrics and measurement

Each of these is an umbrella for several several interesting ideas

1. The "Momentum" Effect. In traditional media, brand advertisers kept control on the story and obtained reach and engagement through mass media. In digital media, especially with young consumers, that model no longer works. Consumers control what media and brands they engage with, and advertisers need to start conversations and then let their consumers continue the dialog. In so doing, consumers often evolve the story (which is challenging for brands) but also often create a tremendous viral effect. An example of this is a recent Adidas campaign on Myspace where the Adidas brand page only got 690,000 page views but the Adidas brand "built momentum: through the 21 M+  exposures it got thorough links that consumers inserted into their Myspace pages or forwarded to their friends. In short, the most successful brands will start conversations and then "build momentum" by leveraging communities of consumers that in effect become brand ambassadors, but without any constraints on hat they can say or do. 
2. Shifts in the power base within agencies are underway. Media agencies are positioning themselves at the center of the ecosystem by owning the communication strategy and the overall metrics/measurement. Media folks would probably argue that media fragmentation coupled with the increasing importance of digital media has resulted in the communications strategy becoming the starting point for the advertising process, and also its most crucial step. Furthermore, they feel that they are best positioned to drive the communications strategy and to manage the overall budget/client engagement because they drive measurement and analytics and can execute against changes in media mix (and over time possibly even the creative) on a day to day basis. As part of this transition, media firms are setting up creative teams in house, merging with their digital arms and in a few cases making technology acquisitions (e.g. WPP's GroupM bought 24/7). This is a pretty radical change when you consider that just 15 years ago, creative ruled and media folks played second fiddle. The question is what does this mean for the on-line folks ? I think its an opportunity as media firms are more natural partners and we could possibly play a role in accelerating this transition.  What do you think ?
3. Metrics and Measurement. Almost every speaker talked about metrics and measurement. The discussions on this front centered around:
   • The need for new metrics to track brand engagement on-line. Almost everyone talked about how CTRs were a poor metric for brands, but the lack of anything else meant that it was the de-facto standard and that in many cases, its use was harmful. In addition, most folks talked about the need to be able to provide analogues to GRPs, while only a few admitted that GRPs were not much better than CTRs and the industry required new metrics and models that link advertising to business goals (like market share). In this context, MSFT talked about its new Brand Engagement model (that provides attribution data across digital media types for consumer actions) and there were discussions about several start-ups trying to create measurement models for brand advertisers - Dynamic Logic, Visible Measures, Quantcast and Vizu. 
   • The need for dashboards and reporting tools that integrated data from disparate sources and automated the grunt work that agencies have armies of people doing. At a minimum, agencies/advertisers want to be able to integrate their disparate digital campaigns (including affiliate programs) in a single dashboard with automatic feeds and drill down reporting. Ideally, they would want to integrate traditional media into the same system. Both Google and MSFT expect to offer media dashboards (for advertisers/agencies) that integrate digital media over the next year, but the vision of a single cockpit across media types is a long way away.

To summarize, while it was a smaller regional version of OMMA NY, the Hollywood event brought together all the key stakeholders  in the digital ecosystem - creatives, media types, advertisers, publishers, ad technology vendors, digital studios, MSFT and GOOG. The notable exception was YHOO, but then, they probably have bigger fish to fry!

I am in LA for OMMA Global in Hollywood for the first time, and it was a different experience from OMMA NY or AdTech. For one, it was much smaller. Also, the profile of the participants was very different - more creative and media types, less geeks !! Even the discussions had a different tone - no-one mentioned "machine learning" in the entire day!

Yet, there were the same intense discussions in the corridors and many great panels and keynotes. Through all of those, I took away 3 over-arching themes on day 1

1. On-line Video is finally here and will change both traditional media as well as drive the next wave of growth of digital advertising
2. Improving digital advertising effectiveness is on the agenda for most industry participants
3. Folks are concerned about the recession and its impact on the overall on-line advertising market and on start-up activity

The rest of this post talks each of these.

1. On-Line Video. Given that we were in Hollywood, it was somewhat predictable that most panelists talked about on-line video content and what it meant for old media, or what some folks in Hollywood would call "traditional" media.  In addition, though, there were discussions about the opportunity to create digital content studios and about how video advertising was at an inflexion point
   • 2007 was the year that the networks and other traditional media content developers and distributors finally responded to the threat of the web. And with some success. The streaming quality on abc.com is impressive while both hulu and fancast are a pleasant surprise, in terms of the content sharing and community features that they offer. Yet what remains unchanged is that the web is breaking down the business model for traditional media, especially the networks. Networks make their money by bank rolling large numbers of expensive pilots (i.e. playing the role of venture capitalists for content developers) and leveraging their extensive distribution to aggregate massive audiences for the 1 or 2 that succeed. With web video, both pillars are being eroded. Digital studios like National Banana are creating pilots for less than 1/100 of the cost (often less than $ 5000) and leveraging Youtube or Myspace or video widgets on Facebook to aggregate audiences. Given that environment, how likely is that the networks will enjoy the same market share on-line that have currently have on TV ? Also, what are the chances that the networks will be able to sustain their high cost content development model on the web ? And what does that mean for their future ?
   • In 2007, advertising spend on on-line video lagged the shift in consumer behavior, but that is about to change. Already in the last 3-4 months, the pace of pilots have accelerated - one panelist said, "I got just 1 RFP in Q4 and I have already received 12 this year") Video advertising is at an inflexion point and over the next 5 years could grow faster than search has grown over the last 5 years. Sounds hard to believe. Well, here is what is driving the growth of video advertising
     • As Waikit (founder of Scanscout) put it, "Web video is a cousin to TV while search was a completely new animal for advertisers". Another panelist told me about how many agencies are funding their video advertising pilots from their TV budgets and not their on-line advertising budgets.  Once advertisers come to terms with the fact that web video offers brand advertising opportunities that are analogous to TV, then what is currently a snowball will become an avalanche.
     • The lack of standard formats, measurement systems, and integration with existing media tools (e.g. DFA)  have hindered the adoption of video advertising. In addition, advertisers look for scale first and then targeting while video ad networks offered targeting but not scale. All this is changing. The IAB is driving standardization of ad formats while companies like Visible Measures are creating measurement systems. Also, emerging video ad networks like Adap.tv, YuMe and Scanscout are starting to get scale with ad products that are analogous to TV and make it easier for media buyers to buy on-line video advertising.
     • Brand advertisers ideally want control over the context in which their brands are presented. The idea of their brands being surrounded by user generated content is scary for most product managers. This is changing though as evidenced by the brands that are advertising on Myspace and Facebook. Advertisers are willing to settle for some sort of guarantee that the context does not hurt their brand (by association). These concerns are starting to be addressed through a combination of technologies that automatically classify video streams and filtering technologies that  separate "inappropriate" videos. In addition, the growth of the mid-tail and head content is creating opportunities for even the most conservative advertiser. 
2. (Digital) Advertising Effectiveness. Targeting seemed to be a popular topic of discussion today. Several panelists spoke about behavioral targeting and re-targeting. There was even a panel on using ISP data for targeting with panelists from Frontporch and NebuAd. In the corridors, there were discussions about how "data" will win and about how the new Google Ad Manager product offers publishers free ad serving in return for click stream data. What was missing though was a more holistic discussion about ad effectiveness - about creative optimization and how dynamic creative construction could significantly improve ad effectiveness or about offer optimization or landing page optimization. Maybe, its because this was OMMA LA and we will hear more about this at AdTech San Francisco. Or maybe, its still early days and the trend needs to become mainstream before its talked about at these conferences. Well anyway, for more about this topic, read my earlier posts on Display 3.0
3. The recession...and what it means for us. Given that the markets were in free fall in the morning, it wasn't surprising that concerns about the recession were on everybody's mind.
   • There was a general sense that the funding environment has changed and start-ups should make their cash last for 12-18 months (some added "at least"). Several people were predicting that venture funding for digital media will slow down in H2 ? what do you think ?
   • There were also lots of discussions about the impact of the recession on on-line advertising ? Will growth slow down ? Will the market contract ? Or will growth in fact accelerate as advertisers shift more money to cost effective customer acquisition channels (i.e. on-line) ?

In all, OMMA Hollywood got off to a good start. Come back to my blog for more about OMMA later this week.

In an earlier post, I teed up the idea of Display 3.0 and how the next wave of innovation in display advertising will focus on improving its effectiveness. In this post, I will begin to flesh out the Display 3.0 concept.

It is widely accepted that click through rates for display advertising are abysmally low, less than 0.25 % in aggregate. Recent research indicates that even these numbers might be over-stated: in a recent study (http://www.redherring.com/Home/23756), 6 % of users account for 50% of all clicks and 68 % did not click on a single ad. Clearly click through rates present an incomplete picture and there is brand value associated with placing the right ad in front of the right person at the right time and in the right context. But studies indicate that most users ignore the display ads that they are exposed to, especially static banner ads. Yet, other studies indicate that even if a person does not perceptibly notice a display ad, repeated exposure leads to a positive perception of the brand (see http://adverlab.blogspot.com/2007/09/study-banners-work-even-when-overlooked.html).

So, other than concluding that there is a need to over-haul the metrics associated with display advertising (which is a topic for another day), where does this leave us ?  I think that there is an immense opportunity to improve the effectiveness of display advertising and that doing so has the potential to significantly accelerate the growth of the industry by accelerating the shift of both brand and direct response $$ to display advertising and possibly even opening up the SMB market for display (though that will take much more than the stuff I discuss below and is also a topic for another day).

So, what is Display 3.0 all about ? I believe that its about solving 3 core problems:

1. Figuring out what Offer to show when an opportunity to serve an ad presents itself (I will refer to this as targeting)
2. Determing what Creative (assuming there are multiple Creatives associated with each offer) to serve to meet the campaign goals (Dynamic Creative Construction)
3. Deciding what Landing Page to send a user to should s/he click on the ad (Landing Page Optimization)

None of these are new problems and companies like Tacoda and Revenue Sciences pioneered a while ago. What's new though is the opportunity to make targeting decisions based on a combination of  behavioral, demographic, contextual, geo. and site (or url) related attributes. And to combine hundreds of data sources (including off-line data) to create user profiles that are then aggregated into 1000+ segments or possibly even to generate thousands of targetable attributes that can be used to generate custom segments on the fly. Also, in the new world, rule based segmentation will give way to predictive modeling and machine learning based segmentation. Turn is a good example of a start-up focused on driving improvements in targeting using machine learning while companies like NebuAdd and AdZilla are providing rich profile data that is on par with the data that Google collects via its tool bar (though their foot print is still tiny at this point - in the hundreds of thousands for both companies )

Traditionally, Creatives are manually created by the agency with huge pride and "skill" associated with developing the "right" creative. Over the last 2-3 years, multivariate testing has become popular and agencies develop 100s of creatives that are tested and multiple creatives are often used within a campaign - for example a different creative for each target segment. Yet this is still an extension of  of the print world and represents a mental model that screams "we know what you want". Also, most display ads continue to static in nature, a mis-fit, given the interactive nature of the web. In the Display 3.0 world, creatives will be dynamically generated (ideally in real time) and optimized (using machine learning) to ensure that the most effective creative is served when an ad call is made. Also, the creatives will be more akin to micro-sites allowing the user to engage with the ad (and even transact) without clicking on it. Finally, I believe that the distinction betwene ads and content will blur (the web's equivalent of infomercials) a a result of dynamic creative construction. Afterall., if Aggregate Knowledge's Pique network shows you a series of ties that complement the shirt that you are paying for on the gap.com site, is that ad or relevent content ?

Companies like Offermatica (recently acquired by Omniture) pioneered the notion of landing page optimization. Lead generation firms consider this to be a core part of their tool set. Yet there is an opportunity to further innovate on this front, with dynamic landing page generation, machine learning based optimization (leveraging the insight gained via targeting, and taking advantage of dynamic nature of the ad unit). Furthermore, closed loop analysis will enable conversion path optimization, which for the purpose of this discussion I will club with landing page optimization.

In summary, the Display 3.0 wave of innovation will dramatically improve display ad effectiveness via targeting, dynamic ad construction and landing page optimization. I will discuss each of these in more detail in subsequent posts.

Tell me what you think ?

Display Advertising has grown to be a $ 9.3 B industry in 2007 (WW), and is expected to grow to almost $ 18 B by 2011. This growth has been driven by both the explosion in web usage as well significant innovation by the display service providers. This innovation includes enabling standardized ad placements, consistent metrics and reporting as well as systems and processes to enable efficient buying and selling of ad inventory.

Yet, the display advertising business faces significant challenges, especially in terms of demonstrating its effectiveness. Users appear to ignore most display advertising: click through rates for displays ads are less than 0.5%. In fact, there is anecdotal evidence to indicate that users automatically block out the right column where most ads are placed. A leading software company tested a new version of their ERP application where they had designed the UI to look like a browser interface with a host of contextual links in the right column. When they tested the product for usability, they were surprised to find that not one of 500 testers used any of the functionality in the right column. Bewildered, they conducted focus groups to understand why. And users told them that they were habituated to ignore the right column because they associated that with ads!

In this posting, I will briefly propose a framework to discuss the evolution of display advertising and begin to make the case for accelerated innovation to improve the effectiveness of display advertising.

The first wave of innovation, Display 1.0,  in display advertising established the foundations for a new medium. This included defining what would count as an impression, agreeing on standards (ad format, size, etc.) and building the systems that would allow large numbers of advertisements to be placed, served, tracked and paid for. Companies like DoubleClick, aQuantive (Atlas), 247 Real Media and comScore emerged out of the wave of innovation.

The second wave of innovation in display advertising, Display 2.0, was focused on improving the overall market efficiency, and were in some ways an attempt to "catch up" with Google's auction based contextual advertising juggernaut. As part of the Display 2.0 wave, companies aggregated inventory (especially across mid and long tail publishers), improved pricing transparency, establishing risk sharing models (such as CPA pricing) and created markets to buy and sell remnant inventory. While there has been much activity on this front in the last 3-4 years, including M&A by the GYM (Google, Yahoo, Microsoft) gang, its still early days,  and several new and exciting Display 2.0 start-ups  have been funded in the last 6 months.

The third wave of innovation in display advertising, Display 3.0 will, I believe, focus on improving its effectiveness. There are many levers that can be pulled to improve display advertising ranging from dynamic creative construction to improved targeting and better landing page optimization. The common theme across all of these is the attempt to apply algorithmic approaches to improve the effectiveness of a medium that has been historically dominated by the "creative types".

In subsequent posts, I will flesh out this framework and talk about the implications for the various stakeholders in the display advertising ecosystem - advertisers, agencies, ad networks/other intermediaries, and publishers. 

I would also love to hear your comments about this thesis !!