Have you seen the new Everstring Audience Platform? I had a chance to catch up with JJ Kardwell, President Everstring, at the end of last week to get a briefing. if you haven’t heard of this yet, I suggest you take a look at their video product tour. It’s well done and worth checking out.
The short story of my personal takeaway…
- Sweet integration with marketing automation – built-in answer to the common questions like, “what do I do now?”
- Love the user input – this isn’t a “black box” platform, users control the flow of the insights discovery
- Clever positioning – for sales and marketing, a solid ligament to help bind sales and marketing efforts
- Unique – capabilities and approach are new, and appear to answer what are normally very difficult questions
- Powerful use cases – load up your competitors recent wins to find their path to success
- The Ironman vs Android point is covered at the end of this blog entry:-)
First, I need to be straight forward. The predictive analytics or marketing folks have a real tendency to prod one of my pet peeves… “predictive” is an adjective, not a noun. It describes something like marketing or analytics. It’s not a person, place or thing. Sigh. Got that off my chest.
JJ’s description of the new offering stayed with me as… an on-the-fly web based tool where sales and marketing folks can find high value audiences. He called the resulting output “Predictive Segmentation”. Many of the steps felt familiar, like popular data purchasing websites. The thing that really caught me was the ability to ask slightly different but potentially really interesting questions. For example, what if you have a target audience revenue size but are really interested in companies of a certain size that are also growing quickly? Answering THAT question gets hard very quickly. This is an example he used to demonstrate the underlying power of their platform to manage non-linear ideas that don’t fit with typical regression models. The beauty is that it chewed through these examples pretty quickly and gave some interesting and actionable insight.
The experience reminded me of the difference between using a card catalog at the library vs using Google. For those who don’t know what a card catalog is, you can, well, Google it. For those who already get it, it was like being able to view the entire library from the dimension you choose.
It’s my sense is that they are using the term “Predictive Segmentation” to describe the blending of two capabilities:
- Look-alike modeling – Users need to seed the model with some answers and some examples. From there, the statistical model responds with options that fit your input. The wider the aperture of inputs the wider the response offered. If your business card says something about sales or marketing, chances are you’ve already tried this with one or more of the company index data firms. Doing this longhand can be very time consuming and wildly limiting. It’s not easy to ask questions outside of basic geographic and firmographic dimensions.
- Segmentation – Segmentation is as old as direct mail. Having said that, it’s a strong element of what enables programs in that channel to produce incredible ROI, even in the face of CPMs that can range above $10,000. Can you imagine that? Fact… catalog marketing, for the right industries, produces unbelievable ROI and has cost structures like that. My point is that effective segmentation is quite important and with the right insights can revolutionize marketing programs.
Back to the case at hand, while some of the underlying concepts are not necessarily newsworthy, I believe they have added some secret sauce to produce something that can provide immediate value. To boot, it’s easy to deploy, provides value to a wide range of users, and has built-in connectors to transition insights to action.
Now, back to the title question about Ironman vs Android. The folks at Everstring used this to describe their approach, and I thought it was incredibly clever… they want to empower sales and marketing professionals to do superhuman feats of marketing and not create a robot to try to do it for them.