Are You Being Served? We’re talking Data-Driven Programmatic Advertising with Omar Abdala, Chief Data Scientist at Lotame

Anyone who has done even rudimentary research on AdTech would know that it is a fairly complex ecosystem. In fact, the ecosystem is as complex as the process itself. We have demand - and supply-side platforms (DSP), DMPs, ad and data exchanges and networks, ad servers, media agencies and trading desks, data and analytics vendors, and of course, bot farms, click farms and other sundry undesirables. Wondering what the data-driven programmatic ecosystem looks like? Here are the key players explained in an MTA-exclusive interactive graphic. 

But first, let’s backtrack to a time before data-driven everything. Back in the day, programmatic began as a way to purchase leftover inventory often at throwaway prices. Back in the day when reach was the key metric for advertising. To make it technically possible, programmatic buying was born. “It would simply be unfeasible to run a campaign off of slices of inventory from a variety of publishers if each slice had to be purchased directly. With programmatic platforms, advertisers and publishers automated and streamlined the buying and selling of media” says Omar.

But given the nature of digital marketing, soon the focus of display shifted from just reach to ‘who was being reached and where’. Today, programmatic technology has matured to the point where ad exchanges are able to enable real-time bidding for premium inventory – something earlier left to the direct buying process.

So today, data quality and scale are both crucial to the success of data-driven programmatic. But recent moves by industry giants like P&G; are also putting the focus back on 3 other elements that go along with higher data and process complexity: transparency, flexibility, and personalization.

With the challenge of bots, click farms, ad fraud and ad blockers facing the industry, there is a huge demand for the same rigor in data quality that B2B marketers have become used to with inbound and ABM type lead generation tactics. “80% of today’s marketers say audience data is critical to their digital advertising efforts, and another 53% have increased their annual spend on data-driven advertising. As the importance of audience data continues to grow for media buyers and sellers alike, concerns over quality and transparency have exploded. And it’s understandable. Everything from unreliable data collection and volume incentives among data providers, to shared PCs across consumers, have fueled quality concerns among advertisers”.

But not every marketer has actionable data. In fact, data has been the Achilles heel of many a brand. While even the most inefficient marketer would be generating tons of first-party data just with day-to-day ops (website, email and other touchpoints), not everyone knows or even wants to know what to do with it. Would it be possible that we’d see a day when working with a third-party DMP with increasingly high-quality data and sophisticated data management engines would eliminate the need for any kind of inhouse data at all? In other words, would we see a day in programmatic when the third-party data would be so good we wouldn’t need first party data?

Perhaps this sounds a ridiculous question to Omar - a data scientist- who may not relate to the aversion some marketers have for scientific, data-backed media buying operations. "Any company not collecting and activating first-party data from their properties is leaving money on the table. The most useful data for any enterprise is that created by a user’s direct interactions with their own properties, campaigns, or resources. Third-party data can augment the view of a user, but first-party data is primary. Taking control of this first-party data should be a priority for any company who wants to monetize, learn from and use this asset to scale valuable audiences, deepen engagement with consumers, and improve ROI across marketing efforts". Fine, so there is no escaping the data and finding ways to leverage first-party data in ever more strategic ways. But the reality is, we still see companies fail with their data efforts every day.

Brands collect a lot of data, but using it intelligently is another matter altogether. And by the time they get around to figuring out what to do with the data, it’s redundant.  We see that every day - not just as marketers, but as consumers too! Every time you get a senseless retargeted ad, every time you get irrelevant emails from a vendor you just spent $1000 on, you feel cheated as a customer!

I ask Omar why that still happens, given all the intel and smart adtech at our disposal? What was the most common oversight he sees marketers making with their data when it comes to programmatic? "The biggest mistake that some advertisers make with first-party data is assuming that it’s enough for a campaign. It’s not. Third-party data is also necessary. It gives you access to more data points to use for expanding your target audience beyond your profiles, or filling in gaps of information about your current audience”. I was talking about giving up first-party data altogether, but apparently, there are marketers that feel the need to get by on first-party data alone!

So, how can a data-driven approach help with the added complexity of cross-device targeting? Just like too much martech happens in silos, too much adtech is also executed in silos. Advertisers are all seeking a better understanding of this cross-device challenge right now. It’s not just the device though – it’s the message and the delivery format too. Display, native, search, content and social marketing need to all come together in a way that seems perfectly natural to the customer. According to Omar, "most marketers aren’t executing successful cross-device or cross-platform campaigns right now – and most publishers aren’t able to deliver the audience guarantees to support them. It’s one of the biggest challenges facing digital advertising today, and it’s a challenge that requires sophisticated technology to address”.

Given the difficulty of executing cross-device programmatic for a seamless advertising experience on one hand, and the need to streamline display (outbound) with search, content, social media marketing and other inbound tactics, what are the chances that adtech and martech would fuse into one in the near future? Will we see a new set of ‘hybrid vendors’ that will specialize in connecting the two emerge sometime soon? (Side note: I’ve already been predicting the emergence of ‘stack rationalizers’ as a new breed of martech specialists for months now!). Omar feels that the convergence is an irreversible trend but more than any external vendors rising to the challenge, it is enterprises that need to take the need to fuse the data seriously, even suggesting a ‘hybrid department’ that enables the right data sets to be commingled (you heard it here first!) for that seamless multi-touchpoint experience across the buying journey.

Also Read:  Top 12 Programmatic Advertising Trends for 2020 and Beyond

Data-driven marketing - including data-driven programmatic advertising is growing, and it's not going to stop growing. Unfortunately, to make the most of it, marketers are going to have to deal with the data question like their life depended on it. Brands run by teams that are serious about data are going to be the brands that people will choose – even though they won’t know that someone already knew what their choice would be. That is the power of data-driven marketing.

Here are the key players in the AdTech ecosystem explained in an MTA-exclusive interactive graphic. 

Editor’s Note: If you weren’t impressed enough with Omar’s insightful comments in my feature, check out his bio!

Omar is Chief Data Scientist at Lotame. He is the author of more than 15 technical papers and patents on statistical modeling, inference, ad network optimization, targeting, and inventory management. Omar holds a B.S. and M.Eng. in Computer Science and Electrical Engineering from MIT and is a Ph.D. candidate on leave at the School of Engineering and Applied Sciences at Harvard University.