• by Kenneth Harlan , Op-Ed Contributor, May 10, 2016

In the mobile advertising industry, data is the Holy Grail – and for good reason. According to a McKinsey study, improved marketing analytics could up ROI anywhere from 10% to 20%, or $100 billion-$200 billion per year, out of a $1 trillion overall budget.

Data-hungry marketers are clamoring to get on board with the latest craze – data matching – to connect the dots between data sets for a more in-depth understanding of campaigns across apps and the mobile web.

But it is no easy feat.

With more brands looking to test the data matching waters, it’s time we acknowledge the reasons data matching can be so difficult and address what we can do to improve the process to better campaigns.

What is Data Matching?

The process of data matching is similar to that of a chef creating a new dish. Everyone knows there are certain foods that go together, but a chef often uses obscure foods and experiment with flavors to see how they complement one another. Similarly, media buyers and advertisers are on the lookout for new data sets that can blend with existing first-party data and tap into a new narrative, providing additional insight into consumers.

Data matching is a process of trial and error. Every company organizes data differently, due to the unique styles of “handwriting,” or code, each database has been written with.

During the process, all parties must anonymize data in order to adhere to privacy regulations, making it even more challenging to find commonalities.  As a result, not all companies’ data will mesh well, but it’s nearly impossible to predict until put to the test.  

The Challenges of Finding a Match

It’s important to address the sheer amount of connected devices in the world today. In 1992, there were 1 million devices connected to the Internet. In comparison, today, there are more connected devices than people – roughly 18.2 billion -- by 2020, the Internet of Things (IOT) is estimated to receive data from 50.1 billion devices.

One of the most frequent issues with data matching is scale. A general rule of thumb: compare apples to apples. A common mistake is to blindly match data sets without considering size and the overall value the match data will provide.

For example, matching data sets can be difficult if you use granular first-party data and try to match it with more generalized or modeled data from a third party. The reverse can also hold true. That said; resist the urge to bring too many data sources into the mix. Tapping into too many databases can backfire and result in unusable match rates.

Data matching is a manual process; as such, there are often challenges with the time it takes to make a match. Before embarking on a data-matching quest, it’s critical to have a realistic timeline. It can take several weeks to ensure data and file formats are aligned.

Tackling Data Matching

Unfortunately, there’s no foolproof way to skip the trial and error phase of data matching; however, there are strategies to employ for a smoother experience.

There is plenty of publicly available data APIs out there, but when courting a third party, you’ll first want to establish whether they’re willing to license their data. If so, a good approach is to first outline your parameters and get a better understanding of the potential partner’s framework by asking questions like, what data they track, how frequently they refresh it, sources, etc.

Word to the wise for mobile: Typically, if their data doesn’t tie back to device IDs, your match rates will ultimately suffer. That is if you’re able to match it at all.

Once you establish the data is a good fit, you’ll want to see where they generate their information, whether it be from apps, POS or Web sites, and compile that data into Data Matching Technology (DMT) to use for your own benefit. Dig into the data and look for any trends, consumer profiles or information that can serve to augment or test theories. From there, let the matching process begin!

Are You Ready to Match?

If you’re unsure if your company or client is ready to take the leap into data matching, here are a few important questions to ask:

Do you or your client have a DMP?

Does your client currently have an app or a CRM database?

Do you currently purchase media programmatically?

Are specific data points like geo or weather important to your targeting strategy?

Do you already leverage first and/or third party data in your ad strategy?

If the answer is yes to these, it’s time to start looking into data matching. If not, the prospect of data matching might be just what the organization needs to get the ball rolling.