Bret Gundersen is the Group Product Manager for the Adobe Analytics data management team. He’s been involved in nearly every major release of Adobe Analytics for the last 7 years.
While at the Adobe Summit in London, I had the opportunity to interview him on the Product Management of Adobe Analytics.
Nicolas Malo: What are the hot enhancement requests from Adobe Analytics customers?
Bret Gundersen: The most common requests I hear from customers tend to be in these areas:
#1 : cross-device visitor identification. Every company has the need for it right now with the explosion of tablet and mobile device usage.
#2 : connecting non-web data. Most companies are looking for an easier way to connect data from CRM systems and internal data warehouses to web data. This work is typically done by data scientists, but the demand for data scientists is about five times the supply.
#3 : statistical analysis : companies want Adobe to tell them what is happening that matters and what is going to happen for key metrics. This requires regression models, audience clustering, correlation with non-web data (for instance weather).
Nicolas Malo: Do you have any plans to enhance the attribution capabilities in Adobe Analytics?
Bret Gundersen: Of course, this is right near the top of our list! While we’re looking to add pre-defined models, we prefer to allow statistics to determine the best model. Most models allow a marketer or analyst to influence the outcome based on a human assumption. While that is important, especially with incomplete data, we’d prefer to provide ways for the data to speak for itself.
We are considering ways to include view-throughs and ad impressions as part of the attribution models, not just click-throughs.
Our vision is to move toward customer analytics. This requires that we recognize the same user across different devices as well as bridging that user profile with all marketing activities, whether online or not.
The ability to merge data across devices is what we call « visitor stitching ». Once visitor stitching is available, attribution is the next highest priority.
Nicolas Malo: What are you plans for cross-device identification or visitor stitching?
Bret Gundersen: We already allow a form of « Cross Device Visitor Identification » by allow companies to set the s.visitorID variable. Logged-in visits with the same s.visitorID value can be matched together, but I don’t recommend doing this in a primary report suite. Rather, it’s currently best to set the visitorID in a second report suite, one that contains a copy of the server calls used in the primary report suite.
As we consider what visitor stitching should allow, we see the need to include historical data in the stitched view. Most visits are anonymous before logging-in, sometimes for months. The campaigns that drive people to a site should be given credit for post-login conversion. This is exactly the problemmy team is working to solve.
Matching pre-login visits requires careful privacy considerations. Depending on the industry, marketing teams need the ability match data just back to the start of the visit, or farther back to when the visitor was first seen. Privacy is the first consideration with this feature and will be built into it from the beginning.
Nicolas Malo: There are still some limitations for processing rules (products variable unavailable, no support for regular expressions, etc…). Do you have any plans to remove them ?
Bret Gundersen: We have recently added some enhancements to processing rules :
- the limit has been upgraded from 50 to 100
- individual rules can be appended to other report suites (instead of replacing all rules)
- the « ELSE » condition is now supported (IF this THEN that ELSE something else)
We will be releasing sub-conditions next month (for example, triggering an event on the search results page, but only if there are no results for an internal search).
We are also considering adding products variable support, simple regular expression-like functionality, routing to other report suites, and a « do not set this event » feature.
Nicolas Malo: Do you have any plans to remove the limitations for props (75 max), eVars (75 max) and events (100 max)?
Bret Gundersen: We are working on it but we do not have an ETA yet. We are also working on ways to incorporate more offline data into Analytics, such as tying attributes at the visitor level. These projects will help reduce the need for extra props and eVars.
These projects are an example of what companies are going to see more of from Adobe. Our goals are to help our customers fully understand the effectiveness of their marketing activities, which means the analysis will need to account for much more than just web data to be complete.
Nicolas Malo: Thank you very much Bret for this interview!
ADDITIONAL INTERVIEWS:
