Four Dimensions of Value from Measurement and Analytics
When I describe to someone how and where analytics delivers value, I break it down into four different areas. They’re each distinct, but they are also interrelated. A Venn diagram isn’t the perfect representation, but it’s as close as I can get: Earlier this year, I wrote about the three-legged stool of effective analytics: Plan, Measure, Analyze. The value areas covered in this post can be linked to that process, but this post is about the why, while that post was about the how.
Properly conducted measurement adds value long before a single data point is captured. The process of identifying KPIs and targets is a fantastic tool for identifying when the appearance of alignment among the stakeholders hides an actual misalignment beneath the surface. “We are all in agreement that we should be investing in social media,” may be a true statement, but it lacks the specificity and clarity to ensure that the “all” who are in agreement are truly on the same page as to the goals and objectives for that investment. Collaboratively establishing KPIs and targets may require some uncomfortable and difficult discussions, but it’s a worthwhile exercise, because it forces the stakeholders to articulate and agree on quantifiable measures of success. For any of our client engagements, we spend time up front really nailing down what success looks like from a hard data perspective for this very reason. As a team begins to execute an initiative, being able to hold up a concise set of measures and targets helps everyone, regardless of their role, focus their efforts. And, of course, Alignment is a foundation for Performance Measurement.
The value of performance measurement is twofold:
- During the execution of an initiative, it clearly identifies whether the initiative is delivering the intended results or not. It separates the metrics that matter from the metrics that do not (or the metrics that may be needed for deeper analysis, but which are not direct measures of performance). It signifies both when changes must be made to fix a problem, and it complements Optimization efforts by being the judge as to whether a change is delivering improved results.
- Performance Measurement also quantifies the results and the degree to which an initiative added value to the business. It is a key tool in driving Internal Learning by answering the questions: “Did this work? Should we do something like this again? How well were we able to project the final results before we started the work?”
Performance Measurement is a foundational component of a solid analytics process, but it’s Optimization and Learning that really start to deliver incremental business value.
Optimization is all about continuous improvement (when things are going well) and addressing identified issues (when KPIs are not hitting their targets). Obviously, it is linked to Performance Measurement, as described above, but it’s an analytics value area unto itself. Optimization includes A/B and multivariate testing, certainly, but it also includes straight-up analysis of historical data. In the case of social media, where A/B testing is often not possible and historical data may not be sufficiently available, optimization can be driven by focused experimentation. This is a broad area indeed! But, while reporting squirrels can operate with at least some success when it comes to Performance Measurement, they will fail miserably when it comes to delivering Optimization value, as this is an area that requires curiousity, creativity, and rigor rather than rote report repetition. Optimization is a “during the on-going execution of the initiative” value area, which is quite different (but, again, related) to Internal Learning.
While Optimization is focused on tuning the current process, Internal Learning is about identifying truths (which may change over time), best practices, and, “For the love of Pete, let’s not make the mistake of doing that again!” tactics. It pulls together the value from all three of the other analytics value areas in a more deliberative, forward-looking fashion. This is why it sits at the nexxus of the other three areas in the diagram at the beginning of this post. While, on the one hand, Learning seems like a, “No, duh!” thing to do, it actually can be challenging to do effectively:
- Every initiative is different, so it can be tricky to tease out information that can be applied going forward from information that would only be useful if Doc Brown appeared with his Delorean
- Capturing this sort of information is, ideally, managed through some sort of formal knowledge management process or program, and such programs are quite rare (consultancies excluded)
- Even with a beautifully executed Performance Management process that demonstrates that an initiative had suboptimal results, it is still very tempting to start a subsequent initiative based on the skeleton of a previous one. Meaning, it can be very difficult to break the, “that’s how we’ve always done it” barrier to change (remember how long it took to get us to stop putting insanely long registration forms on our sites?)
Despite these challenges, it is absolutely worth finding ways to ensure that ongoing learning is part of the analytics program:
- As part of the Performance Measurement post mortem for a project, formally ask (and document), what aspects, specifically, of the initiative’s results contain broader truths that can be carried forward.
- As part of the Alignment exercise for any new initiative, consciously ask, “What have we done in the past that is relevant, and what did we learn that should be applied here?” (Ideally, this occurs simply by tapping into an exquisite knowledge management platform, but, in the real world, it requires reviewing the results of past projects and even reaching out and talking to people who were involved with those projects)
- When Optimization work is successfully performed, do more than simply make the appropriate change for the current initiative — capture what change was made and why in a format that can be easily referenced in the future
This is a tough area that is often assumed to be something that just automatically occurs. To a certain extent, it does, but only at an individual level: I’m going to learn from every project I work on, and I will apply that learning to subsequent projects that I work on. But, the experience of “I” has no value to the guy who sits 10′ away if he is currently working on a project where my past experiences could be of use if he doesn’t: 1) know I’ve had those experiences, or 2) have a centralized mechanism or process for leveraging that knowledge.
What do you say when someone asks you, “How does analytics add value?” Do you focus on one or more of the areas above, or do you approach the question from an entirely different perspective? I’d love to hear!