We did this campaign last month for Haworth Media / Reelz Channel, taking content from their site (previews to over 10,000 movie trailers, interviews, etc.) and showing the right one in the ad based on a weighted average of performance, context, semantics, and behavior. We were able to catch the engine at work in a particularly cool semantic matching instance on Access Hollywood:
Note that the inventory was sourced through our ad exchange partners and each impression was scanned in real time for these types of semantic (and other) cues.
Fitting that a conference on social media strategies for Travel should be nothing less than transcribed via Twitter
It was invigorating to say the least. Our co-founder and COO Jon Aizen gave a great talk on Semantic (Dynamic) Advertising and how it can make Display relevant again.
The reactions were equally invigorating. Travel is our first love, and we have the results to back it up. Seeing the room light up when we showed them they could actually show people the right Travel offers based on the content they were browsing never ceases to make me smile.
Great post today by Sara Lacy on how the old-world model of display 1.0 is letting us down in spite of the huge capabilities of the web today. Yes, we should be doing better things, and that’s just where semantic advertising, dynamic advertising, and other great breakthroughs come in. Use intent people, please!
For the last 19 years, the prevailing metaphor of the web was pages. Self-contained content and feature silos, addressable using a URL. In the last 10 years people have been promoting the concept of the Semantic Web as a web of data, effectively a rather large, distributed, semantic database.
The interesting thing to note is that while both these visions were conceived and promoted by the same person – the honorable Sir Tim Berners Lee – one of them was an immense success, probably one of the most important invention of the 20th century, up there with nitrogen fixation (without which we’d starve) and modern antibiotic, while the other a (up to epsilon) total failure. What’s also interesting is that it is totally obvious to any observer that a wide scale Semantic Web is infinitely more powerful and valuable than a wide scale web of linked documents. So this looks like a classical example of a very high potential barrier forcing us to stay at a local minima, rather than moving to the absolute minima. This barrier is mainly caused by the lack of linkage between personal incentive and global incentive in this system, and given that the web, in any form, is a highly distributed system with infinite amounts of personal, small value contributions, without an alignment of interests between personal value and global value, the semantic web vision will never be realized. So an interesting question to ask is, can we dig a tunnel in the potential barrier? and if so, how would such a tunnel look like? what will the web be like along the way from its current form to the full vision of the semantic web? Here’s one cool hint:
The web will become more and more semantic as people generate more and more tangible value (e.g. money/traffic) from opening up their content and investing in building open systems. At Dapper we believe one of the best vehicles for this transformation is online advertising, and specifically display advertising. Advertisers are incentevized by the promise of improved CTR, conversion or brand engagement. So if opening up their content will mean better performance and ROI, advertisers will go ahead and do that. Let’s look at why this is in fact so.
A good friend, Ashu Garg has coined the term ‘Display 3.0′, referring to the fact that display advertising has gone through two and now arriving at the third revolution in its short life. To summarize the view, which I subscribe to, Display 1.0 was all about just putting the ad there. So the poster child companies of that era were companies like Doubleclick and Atlas who provided and continue to provide the basic ad serving capabilities we all need to get the machine going. Display 2.0 was all about market efficiency through optimization and targeting – the poster child of display 2.0 is probably the Right Media exchange, but also ad networks like Revenue Science and Tacoda. Display 3.0 is about moving from optimizing the buy and sell process to optimizing the effectiveness and relevance of the ad and landing page experience. It’s about understanding what the user may want and need, and then catering to his intent, not at the broadcast, dancing monkey kind of way, but in the very granular offer level. That is where the semantic web really starts to shine. Let’s understand why.
In an ideal world, as I browse the web, ads will bring highly relevant and rich offers. This could be a Berlin hotel room offer when I’m reading about bars to visit in Berlin, a fully stacked grocery basket when I’m reading a recipe, or a personalized mortgage quote when I’m reviewing a house for sale. In the last decade, we have come to expect sites to be dynamic, be responsive and be attentive to our needs, and they have become. When we go to Amazon, each of us is presented with a unique experience, that was tailored specifically for him. Same is true when I’m reading feeds on Google Reader, or looking at the weather forecast in My Yahoo!. However, to the most part, the web of 2009 is accompanied by an advertising experience that is situated way back in the 90’s. The main reason for it, if we attempt to close the cycle, is that those granular, rich offers are locked deep behind websites forms and interfaces, not open to be used in the context of an ad. In the semantic web vision, offers will ‘go out’ and ‘find’ the users that are interested in them, because they’ll ‘understand’ the content on which they are presented. In fact, it will be hard to distinguish between the main source of content and other commercial content that will be mashed-up to it in a cohesive experience where both parts of it provide the user with relevant value.
The nice thing about what we’re doing here in Dapper is that we bring this ideal future vision to our customers today. All of the examples I presented above are in fact ads that we’ve created for our customers in the past, and this is just the beginning. Our ability to harvest all of the offers that exist inside the deep web and behind advertisers’ closed doors, and match it intelligently to the hundreds of millions of different intents and tasks people are using the web for will transform the effectiveness and performance of display advertising, and will present a resounding incentive for everyone to move into a more structured, meaningful and semantic web. In 2009 display advertising is fighting for its life, and we’re geared up to be its white knight
