So you are sitting in a meeting and your boss asks you, “Tell me, what are the chances you think this strategy will work? How much revenue do you expect to generate from this project?”
Wouldn’t it be great if you could simply respond:
“We have analyzed this, and based on our datasets, our team tells me that we have a 90% chance that this project will generate in excess of $1,000,000 in new revenue within the next 9 months.”
Does that sound like a pipe dream to you? It did to me too, until we discovered predictive analytics.
Predictive analytics is the running of statistical modeling based on many pieces of data from multiple sources. The statistical modeling will then tell us what marketing strategies might work, what are the chances they will work, and point out opportunities to us that we never knew existed. We are doing this now at Fathom Online Marketing and it works so well that we have given this service the nickname “Project Gold Rush.”
To give you a more tangible example, we do ‘pay only for performance’ email marketing for a client. They compensate us on a cost-per-lead basis, so we really want to generate as many leads for them as we possibly can. We took all the data from their database, such as: first name, last name, email address, phone number, date of opt-in, time of opt-in, original interest, SalesForce.com data collected from their sales team and we then merged it with our data sources. These data sources include U.S. Census Bureau data, email campaign performance (who opens what emails and clicks on what links and fills out what forms), other proprietary databases we have created and have access to, and then we take keywords and ad copy performance from pay-per-click campaigns for that client. Next, we dump all of this data into massive databases and update them in real time. Lastly, we then run statistical models that ask (among other things):
- Which of our customer segments are most likely to turn into a sale?
- What messages did they respond to?
- What keywords are their hot buttons?
- At what time is someone in segment X ready to make a purchase?
Our models then tell us where to focus our marketing efforts and they do so with amazing accuracy. This can result in us sending fewer emails for our clients, yet making them five times more money from their email campaigns. In addition, these models can help inform our pay per click, search engine optimization, display, website conversion, and social media strategies for our clients. The end result is that we predict the future — and the future is: we will make you more money.
Want to know the coolest part? We give everything I have described above to many of our performance-based email marketing clients for free.
So do you want to predict the future and make a lot of success out of that power?