Why Multiplicity In Analytics Matters
Lately I’ve been writing this column with a very “stream-of-consciousness” style and approach. I often choose my topic for the week from real-life experiences I’ve had interfacing with numerous clients, both internal and external to my agency. And this week’s no different.
Several conversations occurred this week on the range and types of analytics tools required to assess the impact of digital marketing programs, search included. One theme emerged that made me think of multiplicity in analytics data sources: the elusive “single voice of truth.”
Looking to a single data source as the definitive word on all things measurement and analytics may be a convenient practice, but it’s incredibly shortsighted. Multiplicity in analytics is required in order to best understand the dynamics of your marketplace.
Multiplicity in analytics isn’t a new concept or idea, either. Avinash Kaushik introduced the concept to many of us in the industry in 2007 with his now famous blog post on the subject. The topic of multiplicity would also become a key theme of his book, “Web Analytics 2.0.” Essentially, multiplicity points to the complexity of the Web and determines that in order to truly understand how consumers are interfacing with brands, marketers must account for multiple data sources that provide insights into various aspects of the customer experience.
For example, a website analytics tool like Google Analytics is perfect for on-site clickstream analysis, but isn’t able to deliver meaningful voice of customer insights like an iPerceptions or ForeSee can.
But here is where idealism clashes with reality. Most organizations fail to implement a framework of analytics multiplicity, and it’s usually for the same select reasons. Other than the obvious budgetary limitations an organization may be faced with, there is typically pressure and/or incentive to seek a single source of truth for all communications data. This dynamic exists because:
1. The boss doesn’t get it, and it’s easy to retreat to the most simplistic, and easily explainable.
2. Analytics packages often measure in fundamentally different ways. Parsing through those differences is an arduous task.
3. Introducing a new analytics data source could contradict existing beliefs/trends. Let’s not rock the boat.
For these reasons, many will want to collapse all date into a single view, one that is familiar and non-threatening.
I should point out, too, that I don’t want to discount the good job that many of the enterprise analytics players are doing. Adobe/Omniture in particular has done a good job at understanding the needs of the modern Web marketer, while building a framework to house a “single source of truth.” But even then many data sources are stitched together from 3rd party API integrations. Multiple core technologies are still required for that complete view.
But we know one thing for sure: The digital landscape continues to evolve, and it’s going to continue to grow more complex. Layer in the even more complex conversation around multichannel analytics and attribution, and this topic becomes decidedly more complex.
But rather than retreat from that complexity, we have to dive in and make sense of it all. Ignorance is not bliss. We can’t pretend these broader dynamics are not at play. It’s time we look up from our Google Analytics dashboards and see just how rich the data is around us.