Introduction
RevSure’s Pipeline Propensity AI engine predictics pipeline conversion probabilities for Leads and Account to help prioritisation at different stages of the GTM Motion
Pipeline Propensity AI Model Overview
RevSure’s AI models use historical data of stage-by-stage conversions along with different attributes of the leads, accounts, and opportunities to make the predictions.
RevSure analyses the stage by stage and lead to opportunity (pipeline) conversion and journey behavior at the aggregate and granular levels across volume, value, velocities, and conversions.
RevSure’s ML (machine learning) models are trained contextually for each customer’s GTM motion, funnel stages, lead, and opportunity life cycles. Our recommendations and predictions are specific to the customer and not a generic capability. Further, ReSure' rigorously test the robustness of the models.
The factors we look at include contact, lead level, and account level firmographic, demographic and technographic attributes along with GTM funnel activities, GTM sources, and momentum attributes, including region, company size, channel, lead sources, activities done, campaigns touches, campaign spends, lead source categories, industry vertical, aging, velocities, sales cycle lengths, opportunity sizes, ACVs, Pricing and Product, etc.
We adapt to different data availability and data quality scenarios by using a combination of AutoML and Heuristic Models.
The AutoML model learns at the most granular level the quality of the different attributes as well as the statistical importance of the different attributes in driving conversion behavior so it automatically selects the best attributes to predict conversion behavior.
The Heuristic model looks at the GTM Motion level conversion patterns at aggregate levels (thus minimizing the impact of potentially bad attribute data) to make sure that the granular insights from the AutoML models are guided by the macro patterns of the GTM motion.
By reconciling the AutoML and Heuristic models RevSure’s AI methodology maximizes the predictive power of its AI methodology while insuring against different data quality and availability scenarios.
What is RevSure’s unique take on Predictive Pipeline Propensities
Our unique take is that predictive pipeline propensities are unique to every business, their product, their macro and micro GTM motion and their category
So we Auto-train our models to every customer individually using our Auto-ML pipelines
Our predictive pipeline propensity is based on Machine Learning models that take into account first party data across the CRM, MAP, Website interactions, Ad systems (Google, Meta, LinkedIn etc.) any interactions across sales engagement (Outreach, Salesloft), Gifting campaigns, etc.
RevSure looks a 100s of "features" across
Firmographic attributes
Technographic
Interactions (channel, activities, etc.)
Journey (unique sequence of interaction)
Time based (age, duration, momentum)
The Macro GTM motion of the customer (lead/account life cycles, opportunity stage life cycles, etc.)
etc.
Our pipeline propensity is time dependent so we give pipeline conversion probabilities across different quarters
We use an ensemble of models that captures macro trends as well as lead/account level journeys and state transitions
It is a combination of multi-variate, time series classification, regression and probabilistic models
How are RevSure’s pipeline propensity models different from those of 6sense and other tools
RevSure’s AI Pipeline Propensity models are not ideological to the ABM motion. We can work with customers with different and hybrid GTM motions
We generate predictive pipeline propensities at multiple levels - lead, account and booking conversion propensities for each opportunity
We don't rely on third party buyer intent. We can bring that in but it is not necessary. Third party buyer intent (is nowadays getting less interest) and is only useful for maybe attracting accounts but not really useful for prioritisation