Understanding your martech stack’s evolution makes working with data analysts more efficient.
At our recent MarTech Conference session, “It takes two to tango: How marketers and data analysts can excel together,” Arti Munshi and I talked about some high-level concepts marketers can understand to help them work efficiently with data analysts. A key part of this is understanding data’s limitations. One of those limitations is a martech stack’s evolution.
Different types of evolution can hinder what a data analyst can do with data collected over time. Here are a few examples.
Analytics platforms
It is not unheard of for a company to change analytics platforms within its stack. There are numerous options for each type of component and plenty of reasons why a company may change which components they use.
A common example is a web analytics platform. Two major players in this space are Adobe Analytics and Google Analytics. If your company swapped one out for the other, this could limit your company’s historical web data. Granted, your company may have stored data from both systems in a data lake, but that doesn’t mean the platforms gathered and organized the data similarly. This makes things difficult for an analyst to account for. If a marketer can understand this history, they can incorporate this evolution into their requests and expectations.
Don’t forget that other stack components can also affect data collection and processing. Switching those around will present similar issues.
The broader organization
Marketing departments don’t stand alone. They exist to promote a broader organization. The broader organization can also present issues that affect data collection.
For instance, a company may not have a data lake where data from systems throughout the organization are stored. Having data all in one place can make it easier to bring other data in for a broader perspective. For instance, leads, conversions and opportunities are not the only pieces of the puzzle. Customer happiness, touchpoints with customer service representatives and so on are all part of the bigger picture.
More robust and complete findings are possible if analysts can look beyond a customer’s interaction with marketing campaigns. However, that depends upon how easy it is to pull all that data together in the first place.
Regulations
It is no secret that privacy regulations have evolved. In the past, it was permissible here in the United States for web browsers to collect a lot of information about website visitors. There was a lot more allowance for sharing or selling that information to other parties — especially with third-party data. Because of this analysts could provide richer and more robust information. Federal and state regulations have clamped down on collecting that data, so this is no longer the case.
Further, privacy regulations also vary throughout the world. The European Union is well known for protecting the privacy of its citizens — far more so than other governmental entities. This shouldn’t surprise any marketer, but analysts are more constrained with EU data when compared to the U.S.
Device manufacturer influence
Tech companies certainly have to comply with regulations. However, sometimes they adhere to more stringent standards than what they’re governed by. For instance Apple has restricted the data companies can collect about its device users. It turned that privacy focus into an extensive marketing campaign. Regardless of how much the perception matches reality, marketers need to note that it is possible analysts won’t have the same data for Apple and Android users.
Customer self-identification
Sticking with devices, this point is a well-known limitation. Most people use multiple devices ranging from laptops, phones and consoles. It is not uncommon for someone to have multiple of each. For instance, they do a few benign personal things (like checking the news) on their work laptop and do that again later on a personal laptop and phone. Ideally, it would be nice to stitch that individual across their devices so the data would represent them as one person instead of multiple people using different devices.
One way to do that is to collect first-party data and look for individuals using multiple devices. Perhaps they have an account and log into your systems from multiple devices. That is certainly one way to accomplish such stitching. However, an organization may not always have accounts for people to log into. In some instances, that may not make sense for someone to open an account with an organization’s systems.
If your organization was able to introduce accounts for people to create and use, please keep in mind that looking back to times when that wasn’t available or widely used will hinder analysts’ ability to track people across devices over time.
Limitations, not impossibilities
As marketers, it is essential to understand such data limitations. Understanding them will help facilitate working with data analysts. Having said all of this, these are limitations — not impossibilities. This is when data analysts can shine. They’re well aware of the limitations, but as data specialists, they likely have far more tricks up their sleeves than a marketer is aware of. Hopefully, coming to the table with a high understanding of the terrain will facilitate some creative collaboration.
The post Data analytics: Your stack’s past and limitations appeared first on MarTech.
MarTechAbout The Author
Steve Petersen is a marketing technology manager at Zuora. He spent nearly 8.5 years at Western Governors University, holding many martech related roles with the last being marketing technology manager. Prior to WGU, he worked as a strategist at the Washington, DC digital shop The Brick Factory, where he worked closely with trade associations, non-profits, major brands, and advocacy campaigns. Petersen holds a Master of Information Management from the University of Maryland and a Bachelor of Arts in International Relations from Brigham Young University. He’s also a Certified ScrumMaster. Petersen lives in the Salt Lake City, UT area. Petersen represents his own views, not those of his current or former employers.
(11)
