Stay marketing and tech-savvy. Get the latest in martech - subscribe to MarTech Today.
Tips For Selecting The Right Predictive Intelligence Solution
With a growing number of predictive intelligence options, columnist Mary Wallace explains how to choose the solution that’s right for your organization.
Predictive analytics continue to intrigue me. Having looked under the cover of what they can do, I’m convinced that marketing organizations will soon have no choice: They must employ a predictive intelligence solution or be overtaken by competitors who do.
“In today’s cross-channel digital world, prospects are self-educating. Seventy percent of the buyer’s journey takes place before a prospect hits your website or fills out a lead gen form,” Amanda Kahlow, CEO of 6sense, told me. “To succeed, you have to find buyers early before they find your competitors by tapping into the buying signals and intent data from the prospects you don’t know about.”
One only needs to review the Chief MarTec Marketing Technology Landscape to see the ever-growing number of options in the predictive intelligence space. Finding the solution that works best for you can be daunting.
Fear not — here are a few steps for selecting the predictive intelligence solution that’s right for your organization:
Step 1: Clearly Define Your Business Needs
What business impact do you expect to realize? What are you looking for (e.g., real time/batch scoring, contact level/company scoring)? How will the information be used (e.g., for customer retention, for marketing to focus on the right leads, for sales to better drive client acquisition)?
Include all stakeholders in this effort to ensure the entire team’s commitment to success.
Step 2: Know Who Will Drive Implementation
Understand who will implement and run the system and what their technical acumen is. Will you use external consultants or leverage in-house resources?
If the team is technical, a more technical solution is an option. Will you be able to deploy the solution in-house, or is a cloud-based solution preferable?
Step 3: Review Functional Capabilities
Check the functional capabilities of the options against your requirements list developed in Step 1. For example:
• Do you need both individual leads and companies scored? Do you need a 360-degree view of the lead?
• Do you need a two-pronged score for explicit (demographic) and implicit (digital body language) data?
• Does the data being collected by the predictive intelligence solution meet your breadth, depth and diversity requirements (variety, volume and velocity)? Note: If you’re a B2B company, variety will have the biggest impact because a wider range of intent can be leveraged.
• Do you need the predictive intelligence solution’s model to learn from itself — almost like artificial intelligence? Do you need a static model, or should it be variable with a sliding window for analyzing the data?
• Do you need the predictive intelligence solution to connect contact-level data with anonymous visitors from your website?
• At what level of granularity do you need the solution to predict interest (e.g., product, need or company)?
• Do you need a solution that will scale as your business grows?
Step 4: Examine Your Technical Needs
Dig into the technical environment of the options that meet your business requirements to make sure they can do what you need them to do. For example:
• Will they fit into your existing marketing stack? Do they have open APIs? Can they communicate with your CRM, website data, and marketing automation solution using standard connectors?
Customization of integration between systems in the marketing stack should be avoided. In fact, most times, this is the cause for a less-than-successful implementation.
• Is the modeling function flexible enough to meet your business needs? Is the solution purpose-built to be a predictive intelligence tool, or has it morphed into that solution? Do the options have the technical infrastructure to accomplish the necessary functions (e.g., relational database vs. Big Data store, linear modeling vs. network modeling)?
• Is the solution able to connect known to unknown leads? If it connects the known to the unknown, does it do so using an email address or is it more flexible, incorporating fuzzy lookups or other methodologies?
• Does the solution have the capacity to process a high volume of data? Does it assimilate intent data across a wide network?
With the decision made, the implementation process will begin. As with any new solution, make sure to communicate the benefits and value to all who will interact with it. Take time to convey the status of the project. And most importantly, score an early win!
AgilOne CMO Dominique Levin gave me this example of success with predictive analytics:
“When Shaklee, a natural nutrition company, started to use predictive intelligence, they found a number of distinct clusters, including a weight-loss cluster that had been under-communicated to. With this information, Shaklee was able to send additional personalized communications to this unique cluster, resulting in a 3x increase in revenue.”
Similar results are possible for a wide range of companies that use predictive analytics.
Whether you’re a B2B or B2C marketing organization, predictive analytics allow marketers to move beyond gut feelings about customers and their preferences, and make data-driven decisions that can advance personalization, foster brand loyalty, and improve business results and sales.
Opinions expressed in this article are those of the guest author and not necessarily MarTech Today. Staff authors are listed here.