They expect answers from their advanced attribution systems but are often disappointed in the results.
The analytics industry deserves blame for overpromising and under delivering. No single solution is equipped to answer every question that brands have, but they are often sold as such.
But they should bother. Brands should not let perfect become the enemy of good. While no single solution can offer a perfect analysis, a combination of analytical tools and processes can lead to real incremental improvement.
The Problem With Analytics
Marketers still can’t accurately measure everything in a single system for perfect attribution, optimization and predictive scenario planning. The reality is, marketers do not have 100 percent visibility to every marketing stimuli that consumers are exposed to and the subsequent actions taken (or not taken).
It’s a problem that exists on both the brand and partner sides. Brands often lack visibility into their own sales data in a way that’s required for analysis.
Low match rates, cookie deletion and walled gardens leave gaping holes in digital media data. Mass media exposure is projected. And that’s extremely problematic for decision-makers held responsible for their marketing budget’s ROI when relying on a single analytic system.
A Current Solution
It’s true — analytics is unlikely to deliver a perfect solution that will improve marketing ROI by 100 percent or more. But marketers can get better. The following combination of analytical tools can drive 10 percent, 20 percent, even 30 percent ROI improvement:
- Marketing Mix Model (MMM): The top-down view of marketing performance that informs total spend, optimizes channel level investments and enables predictive scenario planning. Used in CPG for decades, MMM is now applied across a wide range of verticals including retail, finance, automotive, telco and restaurant. These models are built with some sales measure as the dependent variable and are typically updated annually or quarterly.Think of MMM like you would your investment portfolio. Rarely is a portfolio perfect — it requires constant optimization to ensure it’s performing at peak levels.
- Multi-touch Attribution (MTA): The bottom-up analysis of consumer interactions across digitized touchpoints. Algorithmic in nature instead of last touch, these models provide a more data-driven assessment of performance at the sub-tactic level to enable in-campaign optimiations. It’s applicable in any vertical with digitized marketing execution and a subsequent digitized KPI. These analyses are often done on conversion events and are updated monthly or as often as the volume of data will allow.Think of MTA like an conductor, keeping the orchestra on rhythm and in-sequence so that the combination of instruments delivers a harmonious symphony.
- Within-channel Analysis: The custom analysis required to understand the nuances of a given channel in near real Keywords, audiences, recency and ad formats are a few of the many dimensions of data for this analysis that is primarily used to assess the relative performance between execution at the most granular level within a given channel.
While sales can be the primary KPI, often brands are optimizing to leading indicators such as viewability, completion rate or cost per site visitor, with the expectation that improvement against those metrics will lead to increased sales in the near future. These analyses are often visualized in dashboards for continuous learning and optimization identification on a weekly cadence.
Think of within-channel analysis like specialized medicine, requiring depth of expertise and continual trials to give patients optimal health in one very specific part of the body.
Even if it’s not perfect yet, these techniques can drive real business value when applied together. And that’s the goal. Don’t miss out on that because of an imperfect solution. Incremental improvement is worth the effort.