Any data without context is just numbers. If your company has quarterly revenue of $100m that might sound good, but if you were forecasting $200m you’re in some trouble. The forecast provides the context to the absolute number.
When you’re looking at social data finding context is also important. If you know that 10,000 people are posting about your brand, how do you put that number in context? You can compare to competitors to look for a “share of voice” or compare over time to see how things are changing. So far, so good – but with Facebook topic data you can do so much more than just count your brand mentions. For example, you can get insights into how different demographic groups are engaging with your brand. So you might find that most of the posts about your brand are coming from men aged 25-34, but what does that mean? It could mean that this group is this most active on Facebook. It could mean that this group is the most interested in your industry. Or it could mean that your brand is really popular with this group.
So how do you know which? One answer is baselining. Baselining is not the latest extreme sport (I’m imagining a hastily erected zipline at the top of a major landmark), but rather a technique for understanding data in context. Baselining allows you to compare one set of results to another and find the outliers. In one analysis of ours we found that men aged 25-34 were indeed the demographic group posting most about Honda. However, we suspected that this might just be because that group was most interested in automotive subject matter in general, rather than having a particular affinity for that brand. So we compared the demographic breakdown of Honda engagements to a breakdown of engagements about the whole automotive topic and normalized the results. (When I say normalized I mean that we plotted different sized datasets on comparable scales). We found that Honda was actually seeing less engagement from the 25-34 year old males than the industry in general and that women aged between 45 and 64 were actually the ones engaging with Honda more than they were with the topic as a whole.
The upshot of this is that while the absolute figures suggest that millennial men are Honda’s most engaged demographic group, there is something about Honda’s products and/or their messaging that is resonating with a completely different group. This should be of immense importance when positioning, creating and targeting future marketing efforts
Baselining can be applied beyond demographic analysis to time series, geographic and temporal analyses to give the real view behind the numbers, leading to those all important actionable insights. For a full explanation of why, when and how to use baselining in PYLON for Facebook Topic Data, check out our recently published design pattern on the subject.