Jon Aizen is COO of Dapper, an online ad network.
Please share what trends Dapper is seeing on the client-side in 2009. Any vertical strengths, typical campaigns, deal size, etc.?
Trends are toward performance-only metrics. We're seeing eCPA as the meaningful metric for our customers, and we're embracing that by offering it as a pricing option right off the bat. Verticals we're growing in are travel, where the emphasis on real-time pricing and inventory is obviously there, and now, Financial Services.
How does Dapper target the purchase funnel for its clients? Are you in the business of generating interest (top of the funnel), or fulfilling intent (bottom - similar to search)?
Definitely more down-funnel, especially when we retarget with our remessaging product. We're able to actually show the exact products the consumer was looking at before he/she abandoned the advertiser's website. Beyond that, when we do a RON campaign with our ImpressionDNA technology that infers intent from every page load, we're matching that intent with an offer from the search engine of offers we've created. So it works a lot like search.
Dapper appears to be offering automated behavioral, geo, creative and contextual optimization simultaneously with ImpressionDNA. Where is the behavioral information stored? It would appear that you use a combination of cookie data and semantic/contextual scanning to optimize. What is real-time in the process?
Behavioral information is stored in a cookie as per usual. What's unusual about this is that we're combining (as you note) these data with contextual / semantic data but also with user location data and past performance. The real-time part is when we add all this up with the proper weights with each impression to then search the offers we have from our advertisers. All this occurs in real-time with every impression. An example: an advertiser could be an online travel agency, selling hotel rooms in every major city in the US. This ad needs a user location and a possible date range.
That's on the advertiser side. On the media side, we're buying a ton of uncategorized RON media from a lot of the networks and scanning each page for a match to this "DNA Profile:" location (via geo-IP), date. When an impression loads that has a valid profile to match the ticketing ad's inputs, we search the OTA site for the right hotel for this user and voila! For the impressions where we don't find a match, we either find one with one of our other advertisers or we simply sell it back to the ad exchange. And of course, performance is used as a feedback loop in this whole process.
With so much power on the side of the algorithm(s), is there anything left for the client to do once the campaign starts?
Watch the conversions come in. Usually the client gives us feedback to expand the buy if it's something they're excited about, but because everything is automatic, there isn't much work on their side except to watch the campaign along with us. Often times we will work closely with our customers to continue the process of innovation on the ad creative, including optimization techniques for which offers to bring in.
How does ImpressionDNA benefit the publisher?
It allows for ads on their site that actually look and behave like integrated site features more than ads. Thus, publishers gain content relevant to their users (think about a travel site with a built-in feature that guesses where you want to travel and knows your home city) that keeps these users on their site longer. In some cases, it means that they can sell Dapper Ads to their own advertisers at a premium, but obviously this is only the case with premium buys. In short, the ad becomes an extension of their content rather than a competitor to it.
Will real-time bidding (RTB) and demand-side optimization impact Dapper?
Yes, absolutely. It makes the whole system even more efficient.
Are ad exchanges a "good thing" for Dapper? Does Dapper want to be an exchange with its impression level technology?
Exchanges are a crucial part of our process. We're partnered with several, including RightMedia, OpenX, and AdBrite, and we utilize them to do the whole buying, categorizing, serving up the wheat and selling back the chaff that powers ImpressionDNA. We actually are far more interested in being an *offer* exchange rather than an ad exchange: we're building up our own search engine indexing our advertisers' millions of individual real-time offers and look to those offers as the commodity that we'll be marketing on the web in the long term.
Your site says that offers are matched to the "right consumer through targeting on pre-contexutalized Dapper media." Some might argue that when media isn't contextual, it stands out more and therefore performs better. What do you think?
Our data simply doesn't support that. Nor does it our mission: we aren't out to get performance by dissonance (with all due respect to that approach), we're trying to get performance by harmony. Search works because it matches intent to a relevant ad, and that's why it's so successful.
Dapper said recently that "conversion performance is the only significant metric that we’re being measured on lately." Is the CPM dead?
For DR, it is heading in that direction. You might get away with charging CPM to a client and feel like you're managing your risk pretty well. But, at some point, your contact and his or her boss are sitting in a room and looking at total conversions compared to total spend. The campaign became defined by CPA whether you like it or not, and if you're not lucky enough to have dropped a conversion pixel on their site, now you're not even in the room to see those data and defend yourself. In order to build long-term relationships with our clients, we should be aligning our pricing model to reflect the success metrics that actually matter to them and determine whether they call us back for the next campaign or not. CPA pricing also allows companies like ours, whose primary intent is to bring performance, to capitalize more fully on the upside of our value.
Thinking about your agency partners, how do you think the agency model will need to evolve in the future?
It's hard to make generalizations across such a large industry, and agencies unfortunately get such a bad rap from tech companies. They're the human element in all this that understands the customer and the marketing goals of the campaign, and like it or not they're the filter of all the new technologies, most of which *don't* work, that throw themselves at their clients on a daily basis. Some agencies take more risks than others; some are more innovative. One thing that I hope won't drive them too much in the future is fear. When new technologies that have had some time to pay their dues and really do prove their worth happen to enhance (or even replace) a process that agencies perform, I hope that agencies continue to run to embrace them and integrate them in their processes.
Follow Dapper (@dapper_net) and AdExchanger.com (@adexchanger) on Twitter.