Predictive Scoring Interview with Tony Yang of Mintigo

Today I had the pleasure of interviewing Tony Yang, VP of Demand Generation for Mintigo, a predictive marketing company. While Account Based Marketing has stolen the show the past year, predictive has become more and more a reality at hundreds of companies. And, predictive tools are sending the message they do ABM too, building more complete pictures of Accounts more likely to convert to revenue.

In this interview, I chat with Tony about when Predictive is a good buy and how ABM and Predictive work together. Enjoy.

Josh: In the Marketing Technology Maturity Model, I placed Predictive tools at Stage 6 – the very end of the 24-36 month build cycle for firms. Do you agree with this? Where do you see Predictive being the most useful for a company?

Tony: Predictive can be used across the Marketing Technology Maturity Model because some of our customers aren’t even fully using marketing automation yet. For example, we had customers forgo setting up a basic lead scoring model in marketing automation and jumping straight into using predictive for scoring. To get started, predictive models require a good set of data records to begin statistical analysis – we typically recommend that you model off of about 300 to 400 customers. For companies that are in the growth stage and don’t have hundreds of customers, we recommend that they look a little bit higher in the funnel and supplement the record set with late stage opportunities.

While companies that are earlier in the martech maturity curve can certainly benefit from predictive, in my opinion the marketers who are a bit further along and are employing multiple marketing channels and technologies can really see compounded value from predictive.  This is really driven off of the insights and predictive data from the models, because once you know what characteristics and buying signals make an account a better target than another – and if you get these data points into the rest of your systems – you can get the right message to the right audience through the right channels at the right time. This is automated personalization at its best.

Josh: Which firms are most ready for predictive tools? Which are not?

Tony: Most firms are looking for the proverbial “needle in a haystack” to find the “GlenGarry Leads.” If we look at the growth cycle of companies, there is a period where customers are few and product market fit isn’t necessarily attained. At an early stage firm, predictive has no power because it would be based on say, 5 records. There aren’t enough statistically significant inferences to say about those 5 records.  You might as well just call all of them to learn as much as you can about these accounts.

Since Predictive’s power is based on what we call the “Positive Set” of actual customer data when building and training your model, we prefer to see a company with about 400 closed-won records. As I mentioned prior, you can certainly supplement this list with late stage opportunities.

Basically, the predictive model will look for key common characteristics of those customers or late stage opportunities that will comprise your ideal customer profile. This profile, which we call your CustomerDNA™, is what predictive scores are based off of. The characteristics of your ideal customer profile as deemed important for your business by the predictive models are essentially derived from a variety of data including your standard firmographic and demographic data, technologies used, hiring patterns and trends, behavioral data, and purchase intent signals. 

Josh: How does Mintigo think about ABM+Predictive? What’s different about ABM+predictive vs. statistically correlated lead scoring?

Tony: Most firms start with a list like the Fortune 500 and say “let’s target the top 50 ecommerce companies on this list.” This is guessing because you don’t have any real data points to indicate whether or not they will buy.  And because ABM requires a lot of resources and time in order to succeed, you may be potentially wasting a lot on accounts that were never a good fit to buy your products in the first place.

Predictive marketing can not only help with ABM by identifying the best accounts, but it can also help discover “net new” accounts and leads, personalize your messaging to each account based on the account profiles from the predictive data, and enable your sales teams with these insights and buying intent signals. Thus, in the context of ABM, predictive isn’t simply about lead scoring anymore – it can help drive intelligent engagement with the right accounts who will most likely buy from you.

Summary

Predictive and ABM continue to merge to assist firms perform better targeting, instead of “best guessing.” While I see Predictive tools being used well later in a firm’s martech build cycle, Tony sees possibilities much earlier, even when firms must rely on late stage Opportunities, instead of customer data. Each firm is different and if your firm need enhanced targeting, you may want to explore Predictive products like Mintigo.