Friday, March 11, 2011

The ABCs of DMPs

Special thanks to iMediaConnection for publishing this first!

• DMPs deliver marketing performance benefits by helping companies learn from the totality of marketing information they collect and purchase
• Brands deserve to collect and keep all of the information they pay to buy, collect, and base action upon. DMPs bring this all together
• A DMP can take a general target, like women aged 18-34, and subdivide it into groupings that help you plan and execute marketing efforts more effectively

From the beginning digital marketing data have been both a potential boon and a missed opportunity for the vast majority of marketers. The ability to track and analyze virtually every aspect of marketing communications brings with it a very real challenge to do more than react to bits and pieces of that information.

In our changed world we have the ability to learn about our customers and prospects constantly. But this relentless flow of information -- from different channels and through different tools -- needs to be gathered, combined, and analyzed in a timely manner in order to capitalize on its value.

A new category of marketing services companies is emerging to help answer all these questions: data management platforms (DMPs). Oh, I just detected audible groaning. Another category of middlemen? Call Kawaja to update the slide (no need, I'm sure he's way ahead of you). I feel your pain, but I also believe that this set of marketing services and companies may really make a difference in our businesses.

In my view, two recent-ish news items reflect the growing importance. The first was the emergence of Red Aril, a start-up led by Jim Soss and Kira Makagon, two well known advertising/ technology veterans. Red Aril is a DMP with a platform created through over 150 man-years of development. The second key item is the recent purchase of DMP Demdex by Adobe. This will add dynamic online ad targeting capabilities to the Adobe Online Marketing Suite.

Both are signs of the likely growth potential in this arena.

The problem: Too many data buckets
The customer relationship management (CRM) team uses Prizm to analyze and segment your hand raisers. The media team buys a variety of third-party data sources to target ads through the exchanges. The market research group commissions important studies that parse and segment customers and provide valuable lifestyle and psychographic insight. And the web team uses Omniture and Quantcast to understand whose visiting and how their demographics and lifestyles impact pages visited, time spent, and purchases. And the social media folks gather learning from activity on Facebook pages and across other online venues.

Sound familiar?

Naturally, marketers have (or should have) some knowledge of all these efforts. But does the data come together? It really really should.

Mark Silva, founder and EVP, emerging platforms for Real Branding, put it this way:
"Think of it as a spectrum that runs from collecting data, to understanding behavior, to identifying and leveraging real insights that answer the question of why people do what they do. In my view, you need to triangulate at least three data sets in order to understand behavior: advertising performance, CRM metrics, and social analytics. Understanding behavior isn't insight per se, but it's a critical step in identifying genuine insights."

Without connecting the dots, too many questions remain either unanswered. Peter Platt, president of PSquared Digital, provided this example:

"Too often, digital efforts are looked at in a silo unto themselves and we miss the real impact that our advertising efforts are having. A great example of this just happened the other day. I was reviewing a law firm client's web analytics and we saw a huge spike in traffic from search last September. Turns out the hero in this effort wasn't our search marketing program but rather a heavy TV flight during that time period. In this example we had visibility into the overall marketing effort, but if we hadn't there might have been erroneous decisions made about how to allocate resources."

Peter's effort clearly helped his team find the underlying truth. But why isn't all our learning leveraged fully? Because, in addition to it being in different places, it's difficult to rationalize and standardize data sets, and then examine that massive data bank for real insights.

Enter the DMP.

The value of data aggregation
Duh. If you have a tool that can handle it, more relevant data is better. And quite frankly, if we are spending time and money aggregating irrelevant data, well then... I'll let Tamara Bousquet, SVP-media director of MEA Digital, say it for me:

"My sole focus is to stay ahead of this ever-changing landscape and deliver actionable results for our clients; squeeze every cent of value out of our client's budget. We find a key challenge for most clients is accurate attribution for different marketing tactics and sales analytics. My team's analytic approach and deep knowledge of the rich first party information and resources help us understand exactly how all tactics, online and offline, impact our clients business and how to best deliver a positive result."

Tamara is not alone in leading her team to do this. But the arrival of tools that can do this on a more granular and comprehensive basis can make the process easier, and potentially more effective.

How DMPs work

So what are these things, anyway? In their current incarnations, DMPs deliver marketing performance benefits by helping companies learn from the totality of marketing information they collect and purchase. They represent the antithesis of data silos.

At its core, a great DMP needs to do four things:

•Aggregate data sources: DMPs are designed to take disparate data sets and combine them into a single, actionable data set. We all know that different tools and platforms gather and collect info in different ways. A DSP partner will set up your instance to take into account the sources and differences of your data sets so that information comes together constantly and consistently, with few errors.

•It is essential to know if the platform you select is capable of parsing the information you already have and shows evidence it is planning (or already work with) many more data set flavors. Because one thing we all know is that what we use today may not be what we are using tomorrow. I'm not talking about all of your company's data (that's IBM's job, or Oracle's). Rather, DMPs focus on marketing relevant information.

•Give you information ownership: Brands deserve to collect and keep all of the information they pay to buy, collect, and base action upon. DMPs bring it all together so that more insights are possible from the combined totality of information.

•Analyze and model: Once the data come together, DMPs offer the means to derive critical information from the data, and work to segment your audience into groups that may warrant tailored marketing efforts. On a blockhead-simple level, a DMP can take a general target, like women aged 18-34, and subdivide it into groupings that help you plan and execute marketing efforts more effectively. This could be based on important demographic criteria, more esoteric psychographic/sociographic ways, or ways you haven't even considered yet. It may also identify heretofore overlooked populations that may be prime opportunities.

•Drive action: The DMP helps to refine and sharpen ad targeting approaches and purchase media more precisely. For example, a DMP could empower better purchase decisions on the ad exchanges, and continue to collect and refine the learning for greater future precision. This is the "immediate value" DMPs can provide.
Jim Soss, CEO of Red Aril, describes the value proposition of his DMP thusly:

"Red Aril's DMP was designed explicitly for real-time channels, the ability to leverage all data, and the integration with a broader marketing database strategy. Our clients see the proof every day -- data drives relevancy, relevancy drives results."

A key part of the aggregation service is a standardization of taxonomy. Indeed, it is a critical part of successfully merging data. Scalability is also a critical consideration. A large brand could be experiencing and recording billions of interactions a month across its marketing efforts. The DMP can only be successful if it is able to store, process, and act upon what could easily become an avalanche of data points.

Some of you are thinking that other types of marketing service providers offer some of these benefits.

You're right. They do.

DSPs, for example, are collecting information in real time and using it to dynamically optimize campaigns and programs and enable users to integrate first-party data in real time. Some DSPs are encouraging users to run all their buys through their platforms to provide a more comprehensive audience view.

For example, MediaMath (disclosure, a CSF client) has put a major focus on providing many of the services that are traditionally the turf of DMPs. Their approach is to empower all buying with first and third party data, not just inventory bought on the exchanges. The principle of empowerment through data naturally has big benefits for pub direct. And most brands do not live by exchanges alone. I think it's natural to expect that all the brands we associate with the DSP sector to move in this direction over time.

The most technologically sophisticated ad networks also work with first-party data, when you choose to provide it. For any business that uses data to define and refine what it is buying or working for you, the race is on to do more with more. As we all know, convergence is the middle name of our industry.

DMP is about driving action from the totality of your marketing-relevant consumer information. You know better than I do whether you are already doing that using a solution that describes itself by whatever name. If not, then the value of the DMP may be significant for you.

In my view, the ideal company for a DMP has data-intensive marketing practices and tactics, and is sophisticated enough that it is focused on incremental marketing improvement. What I mean by that last bit is that a DMP is great for a company that knows it's doing a lot of things right, and is now looking to drive improvements on their good general direction.

But back to initials. My suggestion is, don't get caught up in the monikers. Rather, consider whether you think bringing all the data together is likely to provide enough of a business benefit to justify the time and money required to do so. Do you have a strong CRM database? Are you really collecting great information on the site? Are you dealing with multiple media vendors collecting and purchasing data on your behalf separately? If the answer is yes, a DMP may well make sense for you.

Who will adopt these first? DR brands probably come to mind, because the initial focus for actionability will be in ad targeting. But it would be a big mistake for so called brand marketers to discount the value of more complete view of the consumer.

And consider this: data aggregation and modeling is not the same is genuine consumer insight. Says Mark Silva,

"In our business, first customers often have disproportionate influence on how the category plays out. DMPs need to be aware that while their first customers may be DR marketers looking for the magic data bullets that drive incremental sales improvement, behavioral information is not insight. Ultimately their success in the market will be significantly determined by the extent to which they offer the tools and opportunities for brand marketers and agencies to discover genuine behavioral insights. They're good at getting to the 'what' -- but they need to also empower us to get to the 'why'."

I like that as a concluding thought. It makes sense to get that data together and empower your marketing with it. But don't leave it at that. It's not their job to do our thinking for us. We need data to find the answer, and the most effective marketing is going to come from both actionable data and a heckuvalot of noggin' scratchin' to discover the seeds of overall brand relevance.

1 comment:

  1. Insightful analysis Jim. Deeper than most I read regarding DMPs.



Because people have been abusing the comment platform to place phony links to deceptive sites, I am now moderating all comments. If your comment is legit and contains a relevant link, it will be published.