The Dangers of Publishing Networks for Predictive Analytics Tools
I spent the week in Miami at the Forrester B2B Conference and I got to see several of our competitors in the predictive analytics space talk about their solutions. I had recently discussed with the analyst at Forrester how all of the different definitions of account-based marketing are confusing the marketplace, and at this event I noticed there’s a similar effect happening in the predictive analytics sphere.
Most, if not all, of the predictive analytics tools that we see in the marketplace today are doing a really good job of acquiring multiple sources of internal data, normalizing that data, and appending an external data feed. More often than not, that external data feed is from a publisher network.
So the first question is: what is a publisher network?
A publisher is a large content organization that has access to dozens of web properties that service their network. Some of the big publishers may have more than a million subscribers that receive e-mails linking them back to this content. Data is collected about the subscribers’ interactions with the in-network web properties, and that data is funneled into the algorithms of our predictive competitors in order to generate a predictive score. This data is then coupled with third-party data, CRM data, and marketing automation data to make for a fairly complete picture of a prospect organization’s intent.
Taking it a step further
There’s a lot of noise in the B2B predictive analytics market right now, and so a lot of people who are getting their feet wet with these tools tend to ask me how MRP’s solution stacks up against our competitors. The truth is that the main goal of our predictive platform is similar to the goal of our competitors – to see buyer intent. I do think we’ve managed to add a layer of nuance, though, because we don’t use those publisher networks. We’ve found them to be way too narrow to feed the algorithm customization capabilities that we strive to deliver to our clients (you can read more about this in the first post of our series). Publisher sites produce content that represents far less than 1% of the World Wide Web, and even then, there are IP resolution issues that most of these predictive providers have. The end result is a lead score that’s very minimally influenced by pre-purchase research.
Our external data feed, on the other hand, is comprised of 93% of the sites available on the Web – not just the walled garden of a publisher network. We are able to ingest this volume of data on a daily basis because of the technology of our parent company, First Derivatives. FD is the majority-owner of KX Systems. KX Systems is the fastest time-series database in the world. For years, FD has leveraged KX’s incredible processing power to build streaming analytics solutions for the largest banks, stock exchanges, and government agencies in the world. MRP is now using that exact same stack of technology to manage the ingestion and analysis of a massive amount of external data.
We then append CRM data and marketing automation data and third-party firmographic data to round out our view of intent. But our predictive score is primarily based on the pre-purchase research of the organizations we are targeting as opposed to the internal data of our clients.
We feel this gives us a much richer view of the true intent of the organization.