CMO

Creating the business case for a Data Management Platform

How to work out if you need a DMP in your martech stack

Like many categories of marketing technology today, Data Management Platforms (DMPs) are getting their fair share of buzz and hype. Likewise, there’s plenty of confusion about what a DMP does, and why you may or may not need one in your martech stack.

At a basic level, DMPs are repositories of online consumer behavioural data. They play a key role in building a qualified audience by enabling first, second and third-party data to be aggregated and matched against other sources.

With marketers increasingly looking to grow their use of owned data for customer gain, as well as capitalise on programmatic advertising and media activities, take-up of DMPs is, not surprisingly, rising.

In a survey of marketing and agency professionals in the US last year by AdAge and Neustar, more than half of respondents were found to already be using a DMP. In addition, just shy of 28 per cent planned to implement a platform in the coming 12 months.

A similar survey, conducted by ExchangeWire and Oracle, found key reasons include improved ROI around marketing and advertising activities (24 per cent), the ability the turn data into insight (19 per cent), first-party data integration (14 per cent) and reducing wastage in media buying (14 per cent).

Yet multiple research studies also indicate many of these systems are falling short of expectations. In AdAge’s report, for example, fewer than 25 per cent of respondents said they were satisfied with how their DMPs are handling channels.

So how does a DMP successfully fit into the modern martech stack and does your marketing function actually need one?

Why DMPs are important

A DMP provides a centralised platform of customer data, which can then be segmented in order for brands to target select audiences via other martech or adtech engines, such as a Demand-Side Platform (DSP). If you’re a publisher, a DMP is a way to better segment and position your own readers for the purposes of selling ad and content inventory.

Thanks to the store of anonymous audience information DMPs have access to, these platforms also often allow marketers to find audiences similar to their current customer set (look-a-like modelling) to buy and target digital advertising and messaging to.

In its most recent Forrester Wave: Data Management Platforms, Q4, 2015 report, Forrester highlighted Krux, Oracle’s DMP (powered by BlueKai), Adobe’s Audience Manager, Lotame, Neustar and KBM Group as key players in this space. There are also a number of adtech players with DSP propositions also offering DMP capabilities in some fashion, including Turn and MediaMath.

Deloitte Digital partner, Steve Hallam, positions DMP as a cornerstone, back-end technology for achieving consistency in cross-channel interactions with a customer or prospect.

“There’s essentially a brain that sits in the middle of all your channels and has the message along with the particular segment or customer. Then you use the marketing technology proposition to deliver that message,” he says. “That’s becoming more important for cross-channel interactions and consistency. Someone who walks into a store, interacts in a mobile app, calls your contact centre or digital platform should have a consistent and cohesive experience.

“The DMP is what drives that: It deals with the complexity in aligning data from different channels, assists with cross-channel interactions, and allows the data as it comes in to build that feedback loop. So if there’s one interaction in one channel, the other channels know about it.”

Technically, achieving this is very complex but it’s very powerful when you get it right, Hallam says.

“That’s becoming real and we are seeing it become more prevalent in work we are doing,” he says. “As artificial intelligence and cognitive technologies hit their exponential curve, I think the next 1-2 years is going to be interesting in this space.”

Locally, it has largely been publishers, digital media players and entertainment sites that have embraced DMPs to improve audience segmentation and management. Earlier this year for example, Mamamia Womens Network rolled out a DMP to collate, segment and utilise audience engagement data in order to better understand the content interests and behaviours of its female readership.

Ticketek, meanwhile, is not only using a DMP to target specific audience groups, but pivot its business entirely, opening up new commercial opportunities to become a digital agency to event promoters and partners by better tapping these specific customer sets.

Foxtel is another player that has embraced the DMP model alongside new big data capabilities to improve media buying and audience acquisition.

According to Forrester, advancements in DMP technology in the past two years are helping cement the platform’s importance within the martech stack. For instance, the analyst firm says many DMP offerings are now more aligned to the needs of brands thanks to better integration with third-party vendors, and provide increasingly sophisticated data sharing capabilities between advertisers and publishers.

Building the case for a DMP

Founder and principal of Raab and Associates, David Raab, points out DMPs were originally created to store anonymised Web browser cookies for advertising networks; as such, they did not contain personal identifiers such as name, phone number or email address. More recently, several vendors have modified their DMPs to add this information, which is needed to support direct communications with known individuals.

“Similarly, DMPs originally kept a single profile record for each person or cookie. This could have detailed attributes but it’s not the same as a database with separate tables of transactions, promotions and responses,” he explains.

Raab’s advice for any marketer looking to build the value proposition for such investment is to dive into the detail of what these data management platforms actually do, rather than what they’re called.

“You have to be careful when looking at any particular DMP to understand what it really does,” he says. “As DMPs expand to more closely resemble true marketing databases – which, by definition, have personal identifiers and multiple tables – they make the cloud systems more powerful.

“Of course, what really matters is having that rich central database, not what you call it. The marketing clouds haven’t gotten very far in building those central databases; rather, they focus mostly on creating a shared customer ID to link the separate databases of their component systems.”

DT partner and executive technical director, Tom Beeby, says the idea of a shared and actionable audience across the marketing cloud is a hugely powerful idea, and it’s driving DMP uptake.

“Over time, our customers may have brought in a tracking tool on their site, brought in a DMP or media agency to do X or Y, or they’ve invested in Experian data. The problem is it all sits in pockets and it’s hard to get a consistent view, particularly in the context of data security and governance,” he says.

“Business need to know what they are collecting, where they are storing it, who has access to it, and what they’re going to use it for. A DMP ticks a lot of those boxes, and allows you to better data harness.

“We’ve never truly had a consistent voice across channels as marketers. By aggregating first, second and third-party data, we’ve done some great work in that space and I see DMPs as a transformational technology in that sense. There’s no easier business case to write than for a DMP.”

Beeby says many DT clients have large legacy media spends that could easily be improved through better utilisation of marketing technology.

“If you look at that with any level of scrutiny, and if there’s a way to use money better, that’s where the DMP comes into conversation,” he says. “But another interesting use case you don’t hear enough about is people increasingly thinking about how they want to be spoken to.

“Opt-out rates in a general audience can be quite significant from an email standpoint. DMPs can be a great way to take a message you hope to communicate through an email campaign, and push into onto your site or paid channel.”

Gartner analyst, Martin Kihn, agrees a DMP is vital for creating audiences for advertising purposes, as well as for aggregating information with third-party data vendors. However, he suggests that in their current form, DMPs are not necessarily critical if your marketing cloud is not positioned around digital advertising.

“For example, Salesforce doesn’t really need a DMP, it has a customer data platform, which is first-party data with a personal identifier,” he comments. “But that said, a DMP is essential for marketers wanting to do advertising outside of Facebook.”

Finding that single source of customer truth

Marketing technology commentator and editor of Chiefmartec, Scott Brinker, says marketers are increasingly striving to secure a single source of truth of customer data in order to gain the Holy Grail: A 360-degree view of the customer.

“But even if you're more modest in your expectations — a 180-degree view of the customer would be fantastic compared to what most companies have today — you still need to converge a system of record for customer and prospect data,” he comments.

DMPs are just one of a plethora of technology platforms options being touted to achieve this, Brinker says, and are commonly found in the martech stack of a company with a large digital advertising budget.

“The problem is that several independent solutions have grown up to address different pieces of that 360-degree circle,” he continues. “CRM is the most mature, but traditionally more focused on sales-oriented information. DMPs grew up mostly around anonymous identities to better target digital advertising. Marketing automation platforms have been great at tracking the intersection between customer data and campaigns, expanding profiles of known individuals based on email and Web touchpoints. Social relationship platforms have risen recently to better manage and mine social media touchpoints with audiences over time.”

Inside and outside of large vendor portfolios, competition for which of these data management solutions will dominate is heating up. However, Brinker points to two other possible scenarios.

“The first would be a federated solution, where all of the pieces continue to exist as distinct entities, since they serve different purposes, but work is made to synchronise them across common keys,” he says. “This is actually how a lot of best-of-breed marketing stacks work today.

“The second would be a new solution that underpins all of them, what has often been called a customer data platform [CDP]. The CDP would pull data from, and then push data to, all of these more specialised data systems.”

The idea of a federated data model is what IBM is looking to provide as an alternative to a DMP through its Universal Behaviour Exchange (UBX). Launched last year, the data connection hub allows users to bring together information stored in different applications within and outside of IBM platforms to power marketing activities.

In simple terms, UBX is different to a DMP because it doesn’t storing all the data, IBM Commerce VP of offering management and development, Kareem Yusuf, told CMO during the recent IBM Amplify event in the US. To date, the vendor has pre-integrated UBX with more than 100 different adtech and social platforms to achieve this.

“A DMP exists to store data and to deliver you that integrated value, you need to put all that data in that data mart,” Yusuf explained. “We looked at that strategy and thought deeply about it, but decided we didn’t want to go down the DMP path in that same way. It’s not holding all the data that matters, it’s the event and triggers people are really interested in.

“If you look at our relationship with MediaMath and that paid and owned connection, all marketers really want to know is that their ad was clicked on, and an identifier that associates with who clicked on it. Based upon someone doing that, and also coming to my website, I’m going to put that prospect in a specific audience segment for a specific treatment.

“What IBM is focused on is those events and driving correlation. In my mind, that is where the help was needed – not in storing all the data, but how you bring this value to bear.”

Whatever technology approach ends up dominating, whether it’s called DMP, customer data mart or something else entirely, what’s clear is that the concept of a data management hub that bridges the gap between martech and adtech is critical for marketers if brand experiences are ever to be consistent for consumers.

“As enterprises discover the ad tech power of data-driven decisioning, they, too, will embrace the opportunity to harness, learn from, and deploy data for better business results with email and Web content customisation, among other areas,” Forrester said in its Forrester Wave Data Management Platforms, Q4, 2015 report.

“With the DMP at the centre, the marketing cloud will be the hub of enterprise intelligence.”

Read more about how brands are using DMPs:

Follow CMO on Twitter: @CMOAustralia, take part in the CMO conversation on LinkedIn: CMO ANZ, join us on Facebook: https://www.facebook.com/CMOAustralia, or check us out on Google+: google.com/+CmoAu