John Nardone is CEO of [x+1], an online buying platform and data company.
AdExchanger.com: What trends is [x+1] seeing from its clients in 2009? Can you describe momentum for [x+1] this year? Revenues, deal size, vertical strengths, product interests, etc.?
JN: [x+1] has been extremely fortunate in 2009 to see its client list, revenue and average deal size all grow quarter on quarter from 2008. After the first quarter, we announced 81% year over year growth. We’ll be in the same ballpark for Q2.
Results have been based on the strong performance and critical insights we’ve been able to deliver to clients. The insights pay major dividends in terms of client retention and our ability to win additional share of budget. As our clients respond to the global economic recession, we’re seeing them move more money online and specifically into accountable, performance based campaigns. This has been great for us and our outlook for the remainder of 2009 looks very strong.
What third-party platform and data providers is [x+1] using today and why? What is the ultimate goal as you aggregate partners?
Our core platform is entirely our own. We have our own optimization engine…we don’t rely on yield manager. We have our own data centers and distribution network…we don’t rely on AppNexus. We have our own decision and ad server infrastructure. And we’ve built this platform to be open and extensible. This allows us to work with any ad server on the market as well as any of the myriad data providers in the industry.
The fact that our optimization engine (POE) can accept any kind of data is particularly central to our strategy. It seems like there are new data providers launching weekly and we’re committed to helping our agency and advertiser clients test and utilize these new data sources. Importantly, we have no designs on being a data broker. We are simply enabling as many data providers as we can on the platform so that our clients have choices for the data they can use.
To this end, We’ve also done a ton of work to ensure that we can leverage the data transfers and APIs that the major ad servers have published. We’re aware that our clients have invested both time and money to integrate these technologies into their process and workflow and feel that it is critical to be able to integrate with them as seamlessly as possible.
How has [x+1] evolved in the past few years? And, how have shifts in the marketplace affected [x+1] strategy?
Originally [x+1] (formerly Poindexter Systems) was focused on optimizing external media campaigns. Unfortunately, in early 2000 the market wasn’t ready for our audience-centric optimization because media buyers purchased content sites as a proxy for audience targeting and there was no mechanism to only purchase impressions for individual audiences. In a sense, we couldn’t drive the improvements we knew were possible because context and audience could not easily be separated. We adapted our solution for website and landing page optimization where targeting and optimizing message delivery to specific audiences could be shown to have a huge impact on results.
In 2008 we returned to external media optimization in a major way, utilizing the infrastructure and decisioning engine that we had built and perfected over the years. We started purchasing inventory through networks and exchanges as a means of executing our predictive targeting. We are now expanding our media solution to tap into other media channels and are preparing to offer it to agencies and advertisers as a self-service platform.
In a sense the recent changes in the media market have allowed us to really unlock the power of POETM, our Predictive Optimization Engine. Not only has it always been the center of our company, it is also the only proven engine in the marketplace.
Are you seeing more brand or DR marketers who are interested in your products? Where do you see this going?
Direct response marketers have been our primary customers but we are absolutely committed to expanding our products to support brand marketers. We recently launched a couple of new products aimed at the needs of brand marketers: total campaign reach and frequency analysis and pre-campaign audience modeling, which are both getting a lot of interest. The ability to do brand-based analysis in terms of reach and frequency and audience profiling of both exchange and non-exchange buys is very appealing to brand marketers who are used to managing delivery metrics. In addition, we are working on a portfolio management product so that brand marketers like Kraft and P&G can buy online media “up front” and intelligently decision which brands within their portfolio’s get the impressions that are best for each. We should have this live by the end of the year.
We are also seeing more dollars that are neither pure DR nor pure brand, but are hybrids. I think this is a reflection of an overall trend for accountability. Even brand campaigns are using objective measures, and the most enlightened clients are using multiple metrics as barometers of brand effectiveness. The car companies do this extremely well.
What is your view on ad exchanges? Benefits? Any traps?
We believe that we’re in the infancy of something that is going to become very, very important to the media ecosystem. The ad exchanges will become a platform that carries a significant portion of overall display spend. That said, I don’t think direct publisher buys are in any danger of going away. Rather, smart planners will use targeting tools and analytics to combine and optimize an overall plan of which exchanges are an important part.
The biggest trap that I’m warning clients about is solving for the problems of today instead of planning for and solving the problems we’re going to see in the future. I use the Wayne Gretzky analogy with my clients all the time: Stop trying to optimize to the state of business today. Watch where the puck is going and make sure you are planning for where it is going to be a year or two from now. Our business is in a period of very rapid innovation. The current state will be very short lived.
On a more tactical level, we’re seeing the usual traps associated with the exchanges, such as privacy, inappropriate content and sales channel conflict that need to be addressed. That being said, we believe that technology, smart policies and process improvements will ultimately address these shortcomings.
Is there a perfect model for online attribution and can it be achieved?
No, there is no perfect model for online attribution. That being said, we work with a lot of direct marketers and in many cases we get visibility on their display media, search and website sites and we know that the current, prevailing last click / last impression attribution methodology needs to be improved ASAP.
We provide our clients with tools for understanding the interaction of online channels and are beginning to help them figure out a more accurate attribution methodology. We do that for clients today because we have a very strong competency in analytics overall, including regression and mix based analytics.
But there is no perfect answer. We believe there is likely to be a different model for every client we work with.
Is RTB (real-time bidding) and demand-side optimization an important development? How will [x+1] respond to this new feature of the exchange model? In that it's based on a just-in-time/spot market, are predictive algorithms relevant?
Yes, RTB is an extremely important development.
Predictive algorithms are more relevant than ever with real-time bidding and demand-side optimization. In fact it’s the only way to go. If you don’t have a predictive algorithm in the new real-time bidding environment, how are you going to do things like optimal bid pricing and delivery forecasting? You can’t have media buyers eyeballing this kind analysis if you hope to drive the best results.
For us the term ‘predictive modeling’ is all about understanding whom is likely to convert. In the real-time bidding environment you’re trying to answer the question ‘What impressions should I buy and how much should I be willing to pay for them?’ and you can’t really answer these questions without predictive models.
If you were a publisher, how would you be preparing to take advantage of the increasingly data-driven world of online media?
We believe that publishers need to embrace the fact that they are not only content creators but also data providers. Publishers need to realize that they have a wealth of data inside their own house and they need to make sense of that if they’re ever going to get the maximum value for their audience.
Representing the interests of the advertisers, I’d like to see publishers be more creative with the custom packages that they make available to advertisers, both in terms of creative format as well as how they bundle audiences. For example, Yahoo’s search retargeting product is very powerful and effective. An example of a custom media package we’d like to see is session-based placement. A lot of premium publishers could totally change the game for themselves by selling the first 3-5 pages of a consumer’s session, so there is an opportunity to tell story as the user moves through content. We’re hearing from our clients that they would like access to these types of programs.
It would seem that predictive targeting would have some application to the publisher side - perhaps predictive yield? Does [x+1] have any aspirations on the publisher side?
Great question. We get asked by publishers all the time to use our algorithms for yield optimization and the answer we give is no. We are 100% focused on the buy side so we avoid any potential sell side conflict of interests and our clients can have a greater degree of confidence that we’re going to get them the best results.
A general question for you: What will media buying agencies need to do to remain relevant in the future?
Media buying agencies need to realize that how they buy their target audience is changing. Media buyers need to learn and adapt classic direct marketing competencies and integrate them into their view of what media buying online is all about. That starts with buying audiences rather than buying placement. That also means that they need to acquire a new technology platforms that allow them to manipulate and enable audience data in a way that they never had to think about before. We’ve been speaking to a ton of media buying agencies and feel that they’re really starting to get it and focusing on acquiring these new skills and technologies which is great.