How does a Data Management Platform work for marketers, anyway?

995 0

By Fouad Bedran, co-founder of OOX 

For the uninitiated, Data Management Platforms or DMPs are specialized ad technologies that allow companies to collect user data for the purpose of potentially serving the same users online advertisements that are more relevant to them.

The objective is ultimately to improve the performance of advertisements in the sense that targeted users will see and interact better with ads they deem more relevant to them.

It sounds all altruistic, right? Not so. Over many years of online ad practice, advertisers used to rely on clicks as a broad measurement of campaign performance. With time, clicks have dwindled due to a variety of reasons leading to a drop in measurable interaction between web users and online advertisements.

One of the industry’s answers to this situation has been, among other efforts, the exploitation of the user data which is currently largely harvested thanks to the famous (inedible) cookie; A sort of a mini-document that sits deep in your web browser and records your user behavior.

Improving performance

How are DMPs improving the online ad performance? Well the DMPs are simply exploiting the data amassed by the cookies to get to know users better. Once a user profile is being formed e.g. XYZ is a user who is planning to book a ticket to Barcelona in October.

We know that because the target user has visited travel websites over the last 4 weeks and has requested prices for economy class tickets from Dubai to Barcelona. Additionally, the same user XYZ has been identified as a Male who is based in Dubai, UAE, aged between 34-40. We also know that because XYZ has registered his details on one of the travel websites.

If you are still not clear on how this info will improve the ad performance, here is a practical example based on an imaginary case: Emirates Airlines plans to fill its seats on outbound flights from Dubai to Barcelona, they coin a special autumn offer with a super deal on economy class.

The airline briefs its digital outfit who in turn uses a DMP to run the ads online. Consequently, users who fit the target profile, rather than seeing ads for ‘car deals’ and ‘water park adventures’, end-up seeing the Emirates Airline ads promoting the Barcelona special autumn offer. The target user likes the offer, clicks on the banner and buys the ticket!

Infinite variations

There are practically infinite variations to the creation of target user profiles, as many variations as there are variables collected by the DMP technology. It is also important to understand that DMPs must closely cooperate with Publishers and Websites. For example, a Publisher like Google can collect a vast amount of data related to a single user by allowing cross communication between its platforms: Gmail, YouTube, Google Search, Google Display Network/Google AdSense etc…

To illustrate the journey, let’s take me as an example. I use my Gmail account to correspond with my son’s pre-school school teacher. I also use the same Gmail account to correspond with my son’s potential new school since he is due to begin school in September.

My son borrows my iphone from time to time to watch kid’s videos on YouTube Kids. Loads of data collected about me already. An algorithm that sits at Google somewhere in the cloud has already created a profile about me, it could be interpreted as follows: “User 1234567 is located in Dubai, United Arab Emirates. He has a kid who is a preschooler and his male parent is looking to place him in a new school, foundation stage.”

The data will be made available to advertisers, if they so choose to target me based on my profile. The same goes for Facebook and most of the international, regional and local publishers.

Not only publishers and tech giants track users, businesses also do! If you visit for instance, there are at least 20 ad trackers loaded onto the website for ad profiling purposes! Here is a shortlist of the ad technologies that have appeared on Emirates website: AppNexus, Criteo, DoubleClick, Facebook Custom Audience, Rubicon, PubMatic and Media Math.

To any technology there are advantages and disadvantages. Let’s start with the advantages (DISCLAIMER: To make it simple, I have covered points that are based on my own professional exposure with no bias intended. In other words, the list is not exhaustive).


Improves Audience Targeting

It is a feat for advertisers today to be able to tap into a customizable core audience. That was in the realm of science fiction when David Ogilvy and Leo Burnett were running their ad agencies.


DMPs gather data on a global scale. This presents an advantage for global advertisers with centralized ad tech hubs. This will allow London office for example to manage digital ad placements across the whole EMEA region without picking up the phone once to ask their local agency for help. Imagine the cost savings!

Levels playing field

Imagine a world where a small airline could target the users of the largest market player with the help of the largest market player! (Yes I said that, that’s not a typo). Well if Emirates is selling data to DMPs, any competitor could attempt to tap into this audience and lure clients in their direction. The DMPs will most likely not help in that task, it is up to the advertiser or his digital outfit to figure it out. The clues are out there, know how to find them. Hint, ask for travel data to destinations dominated by the leader.


Increases cost

Obviously, all the partners who have contributed to the collection of the data are asking for a cut – advertising is a business after all. But then again it is up to advertisers to weight the investment vs. return at the end of the day. Net, technology should not be used simply because it has become trendy – always crunch the numbers. It might work for your neighbor, but not for you.

Adds more middle-men to the ad sales equation

Have you looked at the number of middle-men added to the programmatic sales path? Before it was simpler Advertiser > Agency > Publisher – Agency being the middle-man. Today, it is Advertiser > Agency > Agency Trading Desk > Demand Side Platform > DMP > Ad Network (Optional) > Supply Side Platform > Publisher. Of course, each new player taking a share of the sale. Do the math.

Exploited by fraudsters

DMPs cannot differentiate if their data has been used to target real humans vs. bad bots. The technology is not yet there. While your DMP provider might be selling you impressions or clicks in good faith, the impressions could simply appear as a pixel = not seen, or they could appear to bots = not real people. Mind you, the DMP will still make its commission on the sale if the fraud has not been detected.

Lacks proof of performance

Due to the novelty of the technology, it will take time for current users of DMPs to build compelling evidence that clearly showcases the direct performance of this technology – with all other marketing contributions considered. Until then, data should be used wisely with clear cause to effect equation. Google has explained it so well in AdWords, if it takes you 4 clicks to make a $4 sale, don’t pay more than $1 per click.

The article appeared in the November 2017 issue of Gulf Marketing Review.