Six things to do before you use predictive to drive ABM

Predictive marketing in B2B. It’s new (to most), it’s shiny, it’s exciting.

You’re super keen. When you explain it (in layman’s terms) to your boss and your boss’ boss, they’re excited. New accounts we don’t know about? Buyers in market right now? Three times the response rate? When can we start? What are you waiting for?

Whoa there. Stop. Think.

This could indeed change the world of account-based marketing as we know it. But did you ever hear the one about the sexy new technology that didn’t quite deliver on grossly inflated expectations? Yes, we know: it’s not funny anymore.

Six Things for ABM

Before you write the check, open up those APIs and start adjusting those filters, you’d be wise to take these six simple-ish steps to success.

  1. Get your first party data in order

If predictive marketing pinpoints an account as being in market, you need to be sure you can identify the contacts within your database associated with those accounts. That means combining data from relevant systems—CRM, MAP and ERP—too. Contacts may be mapped to any one or more of these and there should ideally be a common identifier across all systems. This common identifier is often an email address. We go into much more depth on this topic in this article.

  1. Develop relevant sales and marketing ‘plays’

You’ve identified a set of accounts that are hot. Now what? Don’t let this take you by surprise. If you have email addresses and other information for contacts at these accounts, you’ll need relevant late-stage content offers, promotions to push, vertical industry use cases to refer to, or a way to deliver demos or provision trial offers. You will also need to measure their engagement and continue to nurture them if they do not turn into closed-won deals.

  1. Train business development reps to go in cold

Yes, the account may be red hot, but the first person you speak to may be ice cold. Inside sales teams used to doing inbound lead qualification, or key account managers used to dealing with decision makers may not be ready to deal with a frosty first response. Are they equipped to engage an influencer who may never have heard of your company, or won’t acknowledge they have issue?

  1. Proactively manage consent

The people you’re going to be dealing with aren’t your usual “hand-raisers”—they may not be in your database, or have even visited your website. You don’t even have tacit permission to communicate. Job #1 in this era of the GDPR is to systemize the gathering and digital consent across all channels you plan to use: phone, email, advertising and social media so that you’re set up to continue the conversation long after the first call.

  1. Implement a rigorous testing methodology

Ensure you have established a proper methodology for designing experiments and analyzing both quantitative reports and qualitative feedback. Otherwise, it’s all just guesswork. Campaigns get implemented, they may work, but you don’t know why. Models never get refined or improved and you’re starting from scratch every time. Whether it’s a simple A/B test or something more robust, make sure to incorporate testing in your process.

  1. Set internal expectations

Predictive marketing is not a magic wand. It’s not a silver bullet or a gold mine either. Start drilling into the data and you may strike a sales-ready seam early, but you may also spend the first few months coming up dry.

Now that you’ve under-promised, you’re ready to start over-delivering. Just take the time to select the right tool (hint: some predictive marketing vendors are way better than others) and if you’re still lacking the confidence to go it alone, there are some great partners out there willing to ride along. Giddy-up!


Steven Elliott


Steven Elliott is Global Partner and VP, Digital Strategy and Marketing at MarketOne International. MarketOne is an integrated global demand generation agency. It creates and nurtures leads, retains and grows customers, and gives clients the confidence their marketing investments will produce quantifiable returns.