infoTECH Feature

January 31, 2017

Data Management Platforms: Nailing the Implementation

By Special Guest
Sunil Rao, Digital Analytics and Audience Management, Merkle

So now that you’ve acquired a data management platform (DMP), it’s now time for you to set it up. A DMP implementation is a very iterative process, so you’ll need to prepare in advance. Some components of the implementation may be complete before others, and that is perfectly normal. A successful implementation will be revisited and tweaked to ensure long-term relevance of the tool.

Have a Plan

Before you begin, it is strongly recommended that you define five to ten key use cases of how you would initially want to use the DMP. Prioritize the use cases based on your intended impact on business goals, which will help drive where you plan to collect and where you can use that data. If use cases are not the preferred approach, a secondary recommendation is to focus on a handful of products or lines of business. By narrowing the focus, this will ensure the full team is aligned to what you’re trying to achieve. A few quick wins will have the remaining product teams lining up for their turn to be onboarded into the DMP.

Once the DMP project has a focus, a project plan should be created to organize the cross-functional teams that will be involved in the implementation. The solution to success in any implementation is frequent communication; key stakeholders should meet at least once a week to discuss overall project status, review any issues/risks, and identify any ad hoc meetings required.

Data Collection: Online & Offline Work Streams

As mentioned in the previous section, the first major phase of any DMP project is data collection. This breaks down into two main subsections: online data and offline data. Online data being that which is collected from digital channels such as website, landing pages, email, and media. Offline data is data stored in your current CRM system that you want to push into the DMP to further enhance your identification and understanding of your customers.

These two work streams are often handled by different groups and involve not only IT, but also members of the marketing and analytics teams. There are a variety of data types and key uses.  

• Online data refers to the majority of data collected via tag (News - Alert)/pixel fires. However, certain considerations should be made based on each channel:

• Site/landing pages: Page visits help build a digital persona about the users visiting your online properties. In addition to capturing page visits, web pages often have other data that will help fuel the DMP. Although the DMP cannot collect any PII (name, address, email), it can collect product interests, shopping cart items, application amounts, and a user’s account status as well as the user’s online CRM ID. All of these help build the necessary attributes to segment the user and target/suppress them for future campaigns. Most tag management systems (TMS) have an out-of-the-box integration with the key DMP vendors, which often helps with quickly implementing this work stream.

• Email: Since most companies have email campaigns directed to current customers, adding tags into their campaign emails is a great way to capture which customers are opening and clicking within emails. In addition, most email systems can include the online CRM ID of the user. The ability to capture this ID to the ID graph within the DMP is a big win, if available.

• Media: Placing a DMP pixel on digital media creative allows you to ingest impression data into the DMP to segment users based on media consumption. This work stream can be implemented at a later time to capture DMP/DSP syndication campaigns that are initiated.

• Search: Similar to media integration, paid search clicks can be captured within the DMP without much heavy lifting and help further inform prospecting and retargeting campaigns.

Build your Integrations

Once some (or both) of your data collection work streams are completed, it is time to start putting this data into action. Ideally, conversations on these integrations began during the project kickoff, and now you are close to testing and launching the tactics planned within your use cases. Below are key integrations for your DMP and some important considerations.

• Personalization: Leveraging the on boarded first-party data, you can now begin to personalize the experience and journey for your customers by pushing audiences to your site personalization/optimization platform.

• Syndication: As the volume of data collected in the DMP increases to a reasonable amount, these audiences/segments can be pushed/syndicated to various DSPs, ESPs (email), and DCOs (dynamic creative optimization) for targeting/suppression as a function of various digital ad campaigns.

Where to start

An initial DMP implementation can take anywhere from two to four months. This is dependent upon the availability of resources, technology, and release schedules. To simplify the implementation below are some guiding principles:

1. Don’t do everything at once: Start with implementing in a couple of key areas, with a focus on your high-priority use cases. The setup should be driven by the business needs and stakeholders.

2. Phased approach: Crawl, walk, run. Start simple with collecting some data (both offline and online). As you learn how you are applying this data, start to layer in more data and more complex integrations.

3. Collect / Analyze / Use: Give the DMP a chance to collect enough data volume to build insightful learnings. Once the volume is there, it can be applied across the different integrations/campaigns you planned for.

4. Rinse and Repeat: The implementation process is not a one-time event. You will be continuously looking to add additional data to the DMP (both offline and online), whether due to the creation of a new landing page or because you are adding another product to the DMP.

The larger the DMP footprint, the more ROI you will achieve.

Sunil is a Senior Director in Merkle’s digital analytics practice. Sunil has more than eight years of experience in formulating marketing strategies through the use of customer-centric analytics and technology, for clients across all industry verticals. He has deep expertise in the areas of digital data and integration, data management platforms (DMP) and cross-channel attribution and measurement. He has a master’s degree in engineering and operations research from NC State University.

Edited by Stefania Viscusi

Subscribe to InfoTECH Spotlight eNews

InfoTECH Spotlight eNews delivers the latest news impacting technology in the IT industry each week. Sign up to receive FREE breaking news today!
FREE eNewsletter

infoTECH Whitepapers