More and more brands are convinced that they must take ownership of their customer data. The burgeoning DMP business is a testament to the idea that marketers want to create and control customer profiles, leverage insights across their marketing programs, and keep their data safe. There's also growing awareness of the need to collect and measure activity across mobile screens in addition to PCs, and to combine mobile behavior data with PC-based data to develop comprehensive customer profiles. This primer outlines the challenges of collecting, measuring, and combining mobile data along with tips on how to make it happen at your organization.
But first: Why you should care
A few years ago, U.S. marketers might safely postpone addressing the mobile data "gap" because relatively little consumer time was spent in apps. But things have changed in a very big way -- and a lot more quickly than most people expected.
In 2014, comScore reported that time spent on mobile devices accounted for a full 60 percent of U.S. connected time. Even more interesting is that 52 percent of total connected time takes place in apps rather than on the web. This is significant because you can't collect data in apps the same way you do on the PC web. Net, without collecting mobile data, means that you're missing out on a massive proportion of the signals consumers are sending as to what they care about. That's why mobile data matters. This is particularly true if you have an app and use it to drive m-commerce.
Step 1: Getting mobile data
When most of us hear "customer data," we think first of third-party cookies, the little workhorses of that have done so much to help us begin to understand consumer digital behavior on the PC web. While there are a growing set of challenges to the veracity of third-party cookie data (cookie-blocking, deletion, etc.), cookies remain fairly effective on the PC web.
But mobile -- and especially mobile apps -- are an entirely different animal, and require very different data collection methodologies. There are a few mobile measurement and attribution options here -- many of which use SDKs that companies incorporate in their apps. Data from in-app measurement tools traditionally has been used to measure the effectiveness of marketing efforts.
The SDKs associate customer events with a "device ID" that is a (semi) permanent unique identifier for each smartphone. The common device identifier helps address the problem that apps generally can't -- or don't -- share information between one another. Without the ID, it would be extremely difficult to get a macro view of in-app behavior.
The good thing about device IDs is that they tend to be much longer lived than third-party cookies, so it is often easier to associate mobile activity with a single user. The challenging thing about them is that in order to be able to collect data on lots of devices, you need a lot of SDK installs. Only a small number of companies have that kind of device penetration.
Step 2: Measuring mobile activity
To measure customer behavior in apps, you must define the sorts of activity that are relevant to your business. We call consumer actions in apps "events." An event might be a product search, browsing activities, adding an item to a cart, beginning the purchase process, finishing a purchase, referring a friend, etc. Once you define events, the SDK tracks consumer actions by device ID, under the assumption that a phone is almost always used by a single person.
Mobile web activity works differently. Third-party cookies can capture some mobile web activity. But Apple's Safari browser blocks third-party cookies by default. SDKs are far better data collectors in mobile.
Step 3: Combining PC and mobile data
Once you have your cookie data from the PC web, and your mobile data from an app measurement company, you need to bring them together into anonymized profiles that represent a single individual across devices. To understand this fully, all behavior data needs to be combined into a single profile.
To combine PC and mobile data, we use a "device graph." A device graph infers that two or more devices belong to the same person based upon signals like household Wi-Fi IPs, login information, etc. These are called probabilistic matches. The limiting factor for most graphs is the number of mobile devices they track. Companies that can "see" more devices can naturally make more matches. The matching is generally performed by "data on-boarders."
Most brands have many other sources of first party behavioral data that can be anonymized and aggregated into these customer profiles. Examples include CRM email interactions, purchase records, and website visitation. Including these data sources naturally further improves your customer view.
Step 4: Putting your data to work
Having profiles isn't an end -- it's a means to an end. What you really want is to leverage these profiles to empower audience development and analysis that you can export and utilize across your marketing mix.
Insights from these united profiles can be leveraged across many of your marketing platforms, powering tactical programs like PC and mobile retargeting, push notifications, triggered emails, site personalization, and larger analytics projects. Most marketing tactics can be made better with audiences and data built on these omni-channel profiles.
What you can do at your company
Taking action at your company is a process of discovery and education. Whether or not your company has begun the process of uniting its PC web data, these five broad steps should help get you started:
- Find out how/if your company is collecting mobile customer data -- both on the mobile web and in apps. If you are, then you have a head start. If not, you will need to identify means of collecting the data.
- Understand the data ("events") that are being measured. Much of app measurement focuses primarily on app installs. This information is useful for evaluating the relative effectiveness of media vendors, but not at revealing what customers are actually doing in-app. Make sure that your measurement tool can and is collecting data on all of the event types that are relevant to your business.
- Once your company has a way to collect the data, you must ensure it is being processed and retained in a DMP and a set of anonymized customer profiles. Be careful here, as many companies have DMPs that aren't equipped to collect or combine mobile web, mobile app, and PC data. Remember that a PC-only DMP is only seeing on average 40 percent of customer activity. A PC web and mobile web DMP is only seeing 48 percent. You need to make sure in-app activity data is in the mix. It is also immensely valuable to have other types of first party data combined in these profiles as well, for an even more complete customer view.
- When all of these data types are being collected and retained, you need to ensure your company is combining them into omni-channel profiles. While it is useful to leverage mobile data for mobile targeting and optimization, and PC data for PC web efforts, it's far more effective to leverage a combined omni-channel profile and insights across the entire marketing mix. If you aren't, you'll need to task your tech team with identifying an on-boarder.
- Once you have all the data together, get creative with how you leverage the profiles across your marketing mix. Retargeting and "look-alike" marketing are the gimmes here, but imagine how a complete view of a customer can, for example, drive a precision push notification effort, or an individualized email program that responds to customer actions, or a marketing automation program that orchestrates a series of touch points to drive customer action.
- In conclusion, the "360 profile" is more than a marketing holy grail. It is (or soon will be) essential to drive optimal marketing effectiveness. While uniting PC and mobile data isn't a doddle, it is now possible -- and with more than half of connected time taking place on phones and in apps, it's a need-to-have, not a nice-to-have.
