What if you could predict – within a day or two of launching it – what the long-term results of your social media campaign would be? What if you could use that prediction to direct and improve the campaign on the fly? From sentiment analysis to predictive ROI, digital advertising agency Organic gives clients social media measurements that matter.
When I first read about Organic’s use of SAS for predictive social media analytics in MIT Tech Review magazine, I knew immediately that I wanted to talk to the agency’s analytics team to learn more. Recently, Jonathan Prantner, Manager of Statistics for Organic, took some time to explain how statistical methods like sentiment analysis and velocity and acceleration calculations inform his clients’ digital marketing decisions.
A lot of traditional marketers’ heads are still spinning over this new era of social media engagement. I’m wondering – as a digital marketing agency – was social media marketing and measurement a natural transition for you?
From a marketing point of view, social media is the next logical extension for us. Organic has a long history around marketing mix work with direct-response mediums.
Social media offers an outlet for branding messages and brand experiences that don’t have an immediate impact on purchases. Social media measurement is tricky in much the same way that tracking how TV advertisements relate to sales has always been tricky. Historically, measuring awareness-building activities and tracking how they affect your brand has relied a lot on surveys, which allow you to gauge those changes in awareness. But they’re generally very slow-moving and one-off processes.
However, social media allows for much more reflective and reflexive measures. Using responses to measure your impact helps you calculate a real-time impact to see how marketing efforts are affecting brand awareness and consideration.
What concerns do your clients have about adding social media to their marketing mix?
People are definitely more excited about it, but measurement and ROI are big issues right now with clients. Clients are treating social media much the same way that traditional media was treated in the past: “It’s great. It’s interesting. But we don’t’ know what value it provides. We know we need to do it – but we don’t know what it means.”
Has your focus on digital measurement given you an edge over other agencies?
I definitely think it has. Since we have a laser focus on the interactive piece, that’s where we put effort on the way we tell our story. As a digital agency that seamlessly blends creative and predictive intelligence we have carved out a space for ourselves. It makes the conversation a little more equal when sitting at the table with other marketing partners.
Tell me about the velocity and acceleration model that was mentioned in MIT Tech Review.
That model is something that’s fairly straight forward. It looks at the rates of change in how your social media imprint is growing. At the point when you examine results and you’re able to see spikes in your increase, you look at the cumulative social imprint and look at first and second derivatives. When we see the spike, we’re able to relate the height of that spike to the ceiling that you’ll eventually see from that spike further on down the line.
Take a Facebook ad. If you get a spike in new fans or likes the first day, we’re able to judge the long-lasting impact and predict how it will trail off in the future. When doing PR activities like events, we can use the same model to judge the success of that event not just based on that day but on the long-term impact of the event.
For any campaign, based on what the overall goals are, you can use these measures to course-correct.
What other ways are you using SAS to measure social media campaigns?
Another method we’ve developed with SAS is to use interactive and social media sentiment measurement as a proxy for consideration. If you’re able to measure your imprint and your Web traffic, you’re able to capture that for computation as well through social sites. We then distill that down into an overall consideration score.
Essentially, clients can see how they’re gaining or losing compared to competitors. If your competitor had a market increase this month, we can see which social metrics are driving that.
Can you provide an example of how this works?
With a homebuilder client, we were able to link online consideration scores to the leads and inquiries that were generated by them. We were able to find that relationship and helped compare this with competitor sites, and estimated what lead volume competitors were receiving at same time.
Changes in customer habits helped inform the client’s marketing plans. One of the main things we learned is that one of their competitors had artificially propped up their scores in a way that wasn’t sustaining because it was driven by sweepstakes. This measurement helped influence the discussions about sweepstakes and the usefulness of those programs.
Jason Harper, my boss and our VP of Marketing Intelligence, will be presenting this case study at an upcoming CMA conference.
Given all the different scenarios in a digital advertising agency, how important is the flexibility that you get from SAS?
The great thing about SAS is that it’s so powerful and has such a broad offering. You’re able to have all these tools at your fingertips to develop distinct models for each part of the process. So, even just trying to capture the spread of a message, you may use one tool to deal with multiple social media and interactive touch points and then another for removing correlations between those touch points and looking at causes. Then, you can use a different area of SAS to develop models for ROI and another for forecasting models, yet it’s all self-contained. Once you build your data sets, you can slice it in so many different ways without having to switch from one software package to another.
The other great thing is all the SAS user blogs and the wide reach of SAS, not only through SAS user groups, but the ability to sit in presentations from different industries and learn from what they’re doing. Sometimes I hear how the Census Bureaus or medical research companies are using SAS, and that outside perspective just fits perfectly for how we’re looking at customer life cycles in a different way.
That’s one of the great things about working for something as broad as the consumer life cycle. Everything from Google to market basket analysis gives you a starting point. You just need code and a different way of looking at something then what you had originally thought.
