Uncovering aha in reporting: aggregate data vs microanalysis

One of the cardinal rules with any marketing plan is to have measurement in place before you start—a key performance indicator (KPI) or set of them to guage success. But…what if you set the wrong KPI?

Imagine the following example. You are attempting to analyze a brand’s marketing and media effectiveness. The brand is running media in certain parts of the country and you expect to see some kind of measurable, positive impact on the business. However, in your analysis and reporting, you realize the data is not telling you the true story. The KPIs don’t add up to a meaningful narrative of media effectiveness or lack thereof. This can happen when looking at aggregate data — the flaw of averages, as the cliché goes. Then what?

You dig deeper, that’s what. You look at individual store data within that aggregate data and find that, indeed, no two stores are created equal. However, when identifying good locations and struggling locations, what if your original KPIs regarding media effectiveness are so general that you still can’t correlate the media down to individual locations? Instead, you find that after all that, it’s still an inconclusive report?

In cases like this, lack of insight is an insight itself. That probably sounds a little too philosophical, so what that means is to use your inconclusion to build new hypotheses:

  • Stores that have a certain square footage are going to perform the same way
  • Stores that have a certain type of ownership are going to perform differently than others
  • The ROI or benefit of the media for one type of store is better than another type of store

At this point, you would use your new questions and ideas to run new reporting models to identify what’s working, specific to each hypothesis. Look at your media broken out by TV, radio, video, social and see which media channel works specifically with each hypothesis.

From this, you can start to identify which stores are seeing a specific value or positive ROI from specific media channels — or not seeing that value or ROI — and identify areas where maybe some of your media has been misspent.

Though this microanalysis, you can start to see some much more meaningful things than were seen in the aggregate, given the added level of detail. Maybe you uncover 100 stores that benefit from media, and 40 that don’t. Use that information to inform your media spend allocation.

As a CMO, you might start to understand the profile of a store that will benefit from marketing efforts more than others. The more stores you identify that fit your profile, the better you can benefit those locations with marketing. Conversely, when you identify a location or type of store that isn’t benefitting from your media, you can have a discussion about what you could do to help those locations fit the model store template.

This kind of approach can have the benefit of changing how you look at your stores, and your media spends and campaign plans. This approach can call into question your current reporting approach. You need to have a certain level of trust in whomever it is you’re working with on the data side to know that they’re going to look at things the right way, and understand what success looks like for your business. The introduction of additional, detailed data can be a lot to take in. Where do you start? It requires a combination of technology and business analysis. You need a powerful data analytics system (like the Callahan Intelligence Platform) to be able to assimilate and process data in a granular way to allow for micro-analysis. And you need the right humans working with the data to recognize that, often, aggregate data doesn’t tell as clear a story as micro-analysis.