By: Cesar Brea
With estimates that as much as 50 percent of digital advertising is now being bought programmatically, the trend is now clearly at its flood tide. But to maximize advertisers’ chances of making the most of the current, it pays to have a keen sense of where enabling capabilities stand. Reach too conservatively, and you’ll miss opportunities to get full competitive advantages from your efforts. Reach too far, and you may subsidize someone else’s experiment but wait beyond your own firm’s patience for returns.
Programmatic capabilities – indeed the definition of the term itself – are shifting incredibly quickly, yet some homework suggested an opportunity for an early 2015 guide to requirements for advertisers evaluating current Demand-Side Platforms or looking for new ones. So, I reached out to some friends in the business to get a sense of what they see as table stakes, differentiating and important-but-still-early capabilities. Here’s what I learned:
One obvious requirement for a DSP historically has been access to a large inventory to bid on. But today, most major DSPs have all rounded up the usual suspects. A greater concern if you’re working with more than one is to avoid overlap, so you don’t end up bidding against yourself. Razorfish Senior Account Director Ryan Clune suggests a different and better consideration is to ask, “What unique inventory do you have access to?” One example of this trend is the emergence of private exchanges offering “first looks” at inventory, or enhanced viewability.
Similarly, with the proliferation of public and affordable commercial data sources for enhancing targeting, it gets harder and harder for programmatic intermediaries to claim that what they have is unique and even more importantly, actually provides an edge.
One way some are compensating for this is to use creative data sources to achieve greater influence on upper-funnel behavior that indirectly influences conversion. Tom Flanagan, director of strategy at DataXu, described how his firm worked with Columbia Sportswear and its agency BSSP to use weather data to shape ad selection and placement, noting a double benefit of increased lift with this new variable, as well as more efficient buys available through the lower prices on upper-funnel impressions.
The most significant data-related frontier today is cross-device user tracking. A good solution for this confers significant advantages to managing experiences and realizing media efficiencies. There are two ways to do this. One is deterministic, meaning that you have an ID of some sort that you can use to match a user across devices. Large players like Facebook and Google have recently started to roll out solutions that rely on their massive bases of almost always logged-in users: “If cookie X from browser A matches to username Y, and cookie Z from browser B matches to username Y, then browser A and browser B are being used by the same person.” (You knew that transitive property thing from third grade would come in handy someday!)
Others like DataXu have taken a more open approach that adds high-quality probabilistic matching signals, like matching device IDs and the demographics of properties visited within a household, and leaves the door open for deterministic partnerships to be included as well. This provides the advertiser scale and flexibility to track regardless of the availability of any individual platform’s key. Importantly, this enables data transparency – a clear view into what data is being used where to track – and gives marketers the ability to quickly adjust to privacy concerns around different types of data signals, whether those concerns stem from consumers, the marketer or regulators. Regardless, it’s still early days for these capabilities, and so the most important thing is to make sure that a vendor’s approach is a good fit for the targets you’re trying to reach and the situation (legal and otherwise) you are in. A good question to ask here is, “How high is my match rate for the audience I’m trying to reach, and how do I audit that?”
The need for speed – for faster and faster bidding platforms that can turn inside competitors cycles – to cherry-pick desired inventory at fractionally better prices is slowing. With viewability rates on open exchanges running only 35 to 40 percent and fraud a major problem, speed is being supplanted by a focus on assessing the quality of the inventory being bid on. Raj Beri, COO of Ozone Media notes that even mid-market customers are now more interested in transparency of what they’re buying.
Unprompted, Raj, Tom and Ryan all pointed out that serving the right content can be as important as buying the right cookie. Flexibility in media types – video of course, rich and/or interactive media, but also increasingly programmatic native – that can be served helps you leverage one production unit across a range of placements, which keeps your creative budgets manageable. Also, the ability to include or integrate solutions for programmatic creative, or alternatively, dynamic creative optimization so you can field ads efficiently and objectively is critical too. (Here’s a good article explaining the difference.)
Tom Flanagan also observes that there’s still lots of value to be added by closing the link between analytics and planning. “The field still looks mostly backward and extrapolates from that to what media to buy next.” There’s an opportunity to be more predictive – considering economic, seasonal, and as we’ve seen, even weather factors – about how markets will evolve, and advertisers should ask about this linkage more often.
Your RFP Itself
Perhaps more important than any factors, given the complexity of the field and its fast pace of change, is being clear about the business problem you’re trying to solve, and that translates to your programmatic execution. Ryan Clune notes, “A lot of times, programmatic equals direct response. But are we trying to get visits? Transactions? New or repeat customers? Retention, cross-sell or upsell? Also, the key to connecting a business goal to a media objective is to define something we can actually measure. Yes, we want new customers – but can we actually identify them as part of managing the buy?”
Tom Flanagan echoes this thinking and suggests challenging vendors to think beyond what they have to do and what they could do, something he feels most would be very receptive to. “RFPs we see are often too focused on a rigid checklist with clients asking, ‘Do you have this? Do you have that?’ rather than providing good context in terms of the business challenge they’re trying to solve and engaging us in figuring out how to get to a good solution.” Tom also highlighted that extra value can sometimes be squeezed by coordinating programmatic efforts across product categories that may share customers. “When you buy cookie X, try product A, but if that doesn’t work, try product B. A programmatic Center of Excellence in your company can help you exploit these opportunities.”
Listening to all this great advice from accomplished practitioners on the front lines reminded me of a paradox any good set of requirements should reflect: Your success with programmatic will be fueled of course by technology, but it will quickly become less about the technology you start with and more about the people behind it. Specifically, it’s a reminder to emphasize working with people that you feel listen to you and that you learn a lot from. When it comes to ROI, tools come and go, but productive relationships compound!
Cesar Brea is author of Marketing & Sales Analytics (Pearson FT Press, July 2014,) and founder and managing partner of Force Five Partners, LLC (forcefivepartners.com), a marketing analytics consulting firm serving leading brands in multiple industries.
With more than 25 years of marketing experience, Cesar has worked as an executive, advisor and entrepreneur across nearly two dozen industries. Prior to starting Force Five Partners in 2008, he headed the digital media and marketing practice at The Monitor Group; led sales and marketing at Razorfish (the world’s leading digital advertising agency); helped build two Internet software start-ups (ArsDigita Corporation and Contact Networks); and led engagements with high-tech clients for Bain & Company.
A frequent writer on marketing and analytics strategy, he published his first book, Pragmalytics: Practical Approaches to Marketing Analytics in the Digital Age, in 2013 and blogs at octavianworld.org.