Is Predictive Modeling the Future of Decision Making?

by · May 28, 20136 comments

With all the time we spend talking about the best ways to measure, the best metrics to measure and how to measure, it’s important to occasionally raise our heads to look at what is on the horizon. Then we can rise above some of the minutia that fills our brains today and get a glimpse of where this is all headed. As technology allows us to process more data faster the next logical step is to start to actually predict the impact a few key decisions will have on sales, revenue and costs.

If you have an extra $30k in your budget, we all want to know where we should put it to get the most impact. Should we put it into more content marketing or should we increase our online ad spend? What if you could log onto a measurement dashboard and answer exactly that question? What if it even projected the difference in revenue based on which triggers you move? Would you want that?

I think most of us would absolutely say yes. The good news is that the technology to do this type of predictive analysis exists and recently I looked at a model that was built by a Social Media Explorer partner that made my mouth water. The bad news is that it’s not cheap. It requires phD level scientists and a ton of analysis on existing data to build the right algorithms, so the price tag is going to be in the high six-figures to low seven-figures. However, there is a market for this today. Companies who see a return in the millions with a simple move of a needle can justify that expense today. Obviously, we hope that the cost of the technology comes down and makes it approachable for more mid-sized companies as well. But this post isn’t really about the fact that the technology exists, in reality bigger companies have been trying to build this in-house for the last decade.

What I want to talk about is what you would do with the data. As executives and marketers we are always on the hunt for better data that will help us drive decisions. However, I wonder if in the face of the best data we have ever seen if we will actually listen to it. Think about that for a little while. If your predictive software predicts that you should decrease advertising spend and increase content marketing significantly would your executive team actually do it? Or what if it was the converse and told you to significantly decrease the spend in social media and content marketing and increase advertising spend to drive immediate revenue gains? Would your executive team do that?

Unfortunately, for too many companies they would say yes to the first scenario and no to the second. As executives and marketers we have to find a balance between short-term ROI and building for long-term gains. We think our decisions would be so much easier if we had reliable data that provided an absolute clear outcome for decision-making. But in fact, better data may make our decisions even more difficult.

There are two sides to this coin. If we don’t listen to the results from the predictive model what good is the predictive model to begin with? How can we possibly justify the effort and expense if we won’t trust the model to help us make the right decision? I can’t think of anything more ridiculous than finding the optimal way to spend that extra $30k and then ignoring it based on our gut.

On the flip side, could the predictive model lead us to making decisions that benefit us in the short-term, but hurt us in the long run? Arguably, there is no predictive model that could’ve predicted the impact social media would have on businesses. By its nature, predictive modeling is relying on past performance to predict future performance. It will always have a blind spot where there historical performance doesn’t exist.

Is our future success found somewhere in the success of yesterday? Or do we continue to rely on our instincts to find the golden nugget that a predictive model would easily miss?

What do you think? Where does predictive modeling fit into decision making? Will it be good for businesses? Could it prevent organizations from investing in new technologies and channels that take time to build ROI? How should we balance the need for short-term return against long-term gains? Leave a comment and let’s have a healthy debate on the role of predictive modeling versus gut instinct. 

Did you enjoy this blog post? If so, then why not:Leave Comment Below | Subscribe To This Blog | Sign Up For Our Newsletter |

About Nichole Kelly

Nichole Kelly

Nichole Kelly is the CEO of Social Media Explorer|SME Digital. She is also the author of How to Measure Social Media. Her team helps companies figure out where social media fits and then helps execute the recommended strategy across the “right” mix of social media channels. Do you want to rock the awesome with your digital marketing strategy? Contact Nichole

Other posts by

Comments & Reactions

Comments Policy

Comments on Social Media Explorer are open to anyone. However, I will remove any comment that is disrespectful and not in the spirit of intelligent discourse. You are welcome to leave links to content relevant to the conversation, but I reserve the right to remove it if I don't see the relevancy. Be nice, have fun. Fair?

  • Dara Khajavi

    This is an interesting point. I think predictive modeling is important, but I also don’t think that any businesses should rely on it. I am interested to see what other people think too.

  • Ingrid Korgemagi

    I think Predictive Modeling is critical in making strategic decisions in both Marketing and Product. In a past agency side life, I have relied upon PM (and that PhD in the Analytics team) for how to convert customers in a pay-for-performance model. Without PM, we would not have had the information we needed to reach a profitable ROI for our efforts. On a brand side, Predictive Modeling helped us make decisions whether to keep a LTO product offering, where and how it would be marketed in the long term for maximized profitability. IMHO, I think PM and customer data as a whole is woefully underutilized by both brands and the agencies that service them.

  • http://blog.clearcastdigitalmedia.com/ mchamberlin

    This is a fascinating conundrum you present and it lays bare the battle between emotion vs data. No one wants to be wrong but no one wants to be obsolete, either. The high six-figure spend you mention at the outset would seemingly tip the scales in favor of acting on those results, no matter how counterintuitive or against one’s gut they may be. But, like every Charles Schwab ad says, “Past performance is no indication of future results.”

    Short-term vs long-term can also be driven by investors and Wall St, too. We abandoned stretch goals in favor of “shareholder value” long ago.

    My sense is that predictive analytics, big data, data analysis (whatever you want to call it) is still rather new in our field and like anything new(ish), there are relatively few antecedents which one can use for comparison, further adding to that queasy feeling in the gut. Data is better than no data, but how much is too much? And when you work in social, aren’t you trying to make emotional connections with people?

    I realize this is a long winded, non-answer, but this is quite the sticky wicket and I wrestle with these sorts of issues every day. Thanks for a great post.

  • Pingback: Top Internet strategy, marketing and technology links for the week of June 2, 2013 - Sazbean

  • Pingback: Predictive Modeling: How Much is it Worth to You?

  • Pingback: How to Make the Business Case for Social Media Marketing « MindCorp | Newsfeed