June 15, 2021
5 mins read
How Multiple Predictive Models Are Game Changing for B2B Companies
5 mins left
Demandbase has been in the “Account Selection” business for over five years now. During that time, our predictive models have evolved past the simple selection of accounts. This evolution is evidenced within our Engagement Platform. We offer two very specific predictive models called Pipeline Predict and Qualification Score. These predictive models take billions of data signals to answer the age-old question: Who are the best accounts to go after right now and in the future?
We are now able to answer the next logistical question: Who are the best accounts to go after for a specific product or market?
Taking a concerted approach to evolve and build on our predictive analytics technology at Demandbase, we’ve released our latest innovation: multiple predictive models. Included in our Engagement Platform, our customers can now create a customized Pipeline Predict and Qualification Score for any product, business unit, or industry. Perhaps more importantly, you can create them without the need for professional services. In addition, we now offer the ability to define the engagement activities that should inform one model over another. This unique approach allows you to think big, start fast, and adapt quickly. Multiple predictive models offer higher quality and accuracy — for a complete understanding of your total addressable market and target accounts.
If your company offers a single product, then a single set of predictive models is often all you need to determine pipeline propensity and account fit. Accompanied by an engagement score, you will have complete visibility into a market. However, if your organization offers multiple products, some ambiguity exists when using a single score. That’s because the single predictive score loses its specificity when the data inputs represent an aggregation of all available data and behaviors across multiple product lines.
Chances are, it’s not enough to have a single predictive analytic model for your organization. The model needs to be correct so that it can be trusted across Marketing and Sales organizations. And this requires an additional level of complexity in how it’s built.
Some predictive vendors offer professional services to build the models over two to three months on behalf of their customers — and they promise a lot. Although this is an alluring proposition to some, it’s a black box: You don’t have direct access to design those models. You also have to spend a lot of time and resources in back and forth dialogue with the vendor to make sure they’re as correct as possible, and your organization ultimately needs to rely on the hope that it’s correct, adopted, trusted, and passes the test of time.
At Demandbase, we not only assist you with the setup, we put more power and control over the creation and iteration of predictive models in the hands of your marketing and data science teams. You might think this would be a laborious process, but creating sophisticated predictive models in under 24 hours is part of our methodology — and it couldn’t be simpler. Let’s review the predictive model build process for both Pipeline Predict and Qualification Score.
The Pipeline Predict score represents the propensity for an account to become an opportunity for a specific product within the next 30 days. The science behind the scoring warrants its own blog, but in summary, historical activities are evaluated against the pipeline creation date and the signals observed are then applied to all other accounts. Out-of-the-box options exist; however, for our more mature ABM customers, you have the ability to define the model yourself.
When models ingest behavioral signals across different products and activity types, the predictive model ends up too generalized. It doesn’t apply to any specific product, but a combination of products. This makes the actionability of the predictive models confusing. Defining the engagement activities that should apply to its respective model eliminates any confusion regarding which accounts are in the market for a specific solution. In turn, this specificity helps Marketing and Sales teams create more actionable playbooks for each account. By defining product-specific engagement activities within Pipeline Predict, we’ve minimized the guesswork required whenever an account is exhibiting in-market signals. The below illustrates an example of which engagement filters should be used with a Pipeline Predict model related to our Advertising product.
The Qualification Score represents the account fit for a specific product or segment. In short, it will create and rank your total addressable market. Out-of-the-box options exist; however, for our more mature ABM customers, you have the ability to completely define the model.
Multiple predictive models are game changing, not only for our experienced customers, but also for anyone looking to start their account-based experience journey. The innovative use of AI and machine learning to process specific signals to predict in-market accounts and to rank account lists has never been more impactful. For our customers that need assistance, we have out-of-the-box configurations in addition to our onboarding and professional service teams. For our customers that want to take control, we are uniquely positioned to accommodate this through an intuitive user experience.
With this latest release, we are continuing to encourage the maximum ABM adoption and value across your organization’s go-to-market teams. If you’re ready to get started, ask your Customer Success Manager or contact us for more information.
Director of Product Marketing, Demandbase
This article was published in:
Why Data Will Eat the World
3 mins read
How to Use Sales Intelligence Technology to Increase Sales Pipeline
Sales Intelligence Technology: Helping Your Sales Teams Be a Little Smarter Than Everyone Else
Welcome to Demandbase, HubSpot CRM and Microsoft Dynamics!
How Financial Services Teams Identify Their Best Prospects
Better Together: New Demandbase Unified Interface in Salesforce
We have updated our Privacy Notice. Please click here for details.