Before Considering Predictive Analytics, Make Sure You Have the Right Source Data to Make it Effective
If you don’t have the right data inputs, you will not get the outcomes you desire
Data sourcing has never been easier in B2B tech marketing. And yet what remains critical is finding effective ways to leverage that data effectively to drive more productive action. This is especially important when you’re building out an account-based marketing (ABM) strategy. You need to define the right plays to make sure data will be used to its fullest potential.
With the rise of predictive analytics solutions, there’s a hope that we can take advantage of machine learning and AI to make sure we’re registering and scoring prospects effectively. Not so fast. While predictive analytics is important for marketing – it can point you at suspects – it can’t paint clear pictures of individual prospects on its own. For that, we need detailed behavioral data and actionable purchase intent insights.
The ‘False Negative/False Positive’ Problem – Unidentified leads from a modeled account or leads that actually aren’t any good
Your predictive analytics solution can incorporate data on things like website visits, which of your white papers your inbound prospects have downloaded and which webinars they’ve signed up for. However, if you don’t feed it the right data, there’s significant demand you’ll miss – both from accounts and names you already have in your database, and especially about prospects you don’t have behavioral information on at all. You’ll still experience plenty of ‘false negatives’ because predictive itself doesn’t help you with insights like:
- Competitors prospects are engaged with
- Visits to relevant external content
- Technologies installed
- Related topical interests
- Additional active prospects on the buying team
- And more
Since any predictive analytics solution is only as good as what you feed it, when you can provide only historical and internally-available signals, the lead scores it generates will naturally be skewed towards companies that already prefer you and therefore interact with your lead gen relatively frequently.
Likewise, you have plenty of people who will visit your site, respond to your outbound and the like – they consume lots of content — but the reality is that many of them have no intention of ever buying your products. You could essentially blacklist these folks from your lead management process, but if you don’t do that and especially if they come from an account you’ve scored high with your model, your processes are going to mistakenly treat them as leads – these are false positives.
The false positive issue gets more and more expensive the further it moves down the funnel without being stopped. It kills ABM effectiveness, and most importantly, it hurts your internal reputation, driving a wedge of distrust between sales and marketing.
None of this necessarily means that predictive analytics done right is failing marketers. It simply means that we have to be super conscious of the effect of good data inputs on real business outcomes. To make predictive modeling work well, we simply need to do a really good job of complementing CRM and marketing automation data with actionable purchase intent insights that come from external behavioral data.
Leveraging (not just predicting) active buyer behavior with purchase intent insight
Moving beyond the false negatives issue that comes fromover-reliance on CRM data alone requires a feed that provides you a deeper understanding of your prospects and the buying behavior of the account as a whole.
If you have access to basic data like email addresses, industry information, company revenue, and content downloads, you can compile a large pool of prospects that might be a good fit for your business.
This is where actionable purchase intent insights come into play. Purchase intent insights go beyond just which content a prospect downloaded from you. The most effective insights come from external behavioral data—the actions prospects take that you typically can’t see. This is exactly the type of insight that TechTarget provides to enterprise technology marketing and sales teams. We monitor the behaviors your prospects are exhibiting across our network of over 140 enterprise technology sites, analyze them, and make them available via our Priority Engine purchase intent insight portal. Priority Engine provides direct access to buyer purchase intent information in your market, delivering rich fuel to power your marketing systems and activation efforts.
When you narrow your prospect pool from those who might be a good fit to those that actually exist in your market, you can drive more effective decisions. Predictive analytics is a part of this process, but you have to take a hard look at your data sources if you want to move the needle.
If you want to learn more about improving B2B marketing performance with externally-sourced intent data, check out this webinar with TechTarget and SiriusDecisions, Beyond Predictive: What You Can’t See (in Your Own Data) Can Really Hurt.