How Snowflake built a data-driven account scoring foundation
Established in 2012, Snowflake has rapidly emerged as a leader in the cloud data platform market with Snowflake Data Cloud, and is a major player in business intelligence. At the core of Snowflake's strategy is a focus on leveraging data to refine targeting in sales and marketing, setting new standards in data-driven decision-making.
The challenge at Snowflake's scale is real: managing a global account universe with inconsistent firmographic and technographic data, then scoring and prioritizing those accounts without manual intervention. Building scoring models on incomplete CRM data produces unreliable territory assignments, a problem that compounds as the account universe grows. ZoomInfo, an all-in-one AI GTM Platform, is a key part of that journey, helping Snowflake transform its account scoring, capitalize on real-time market signals, and design a new wave of personalization tools grounded in ZoomInfo's intelligence layer.
The account propensity scoring model Snowflake built sits at the center of this story: a system that ingests enriched data at scale, produces scores that feed directly into CRM, and drives territory allocation decisions for field operations leads worldwide.
Central to Snowflake's own data strategy was the development of tools that used in-depth firmographic and technographic data. Firmographic data covers the structural attributes of an account: industry classification, employee count, company size, and revenue range. Technographic data captures the technology stack a company runs, a critical signal for identifying accounts that fit Snowflake's product profile. These tools generate data-informed scores that feed into Snowflake's CRM system, enabling district and regional leaders to sharpen their territory planning efforts.
Snowflake's commitment to treating data as an asset (not just for understanding the global account landscape, but also for optimizing internal processes and decision-making) has been pivotal in turning a complex challenge experienced by many businesses into a strategic opportunity. Building scoring models on incomplete CRM data produces unreliable territory assignments, and that structural problem is exactly what Snowflake set out to solve.
Account propensity scoring: architecture, data inputs, and results
A key element in realizing this vision is Snowflake's Account Propensity Scoring (APS) model, a tool designed to measure a client's potential fit with Snowflake's offerings. Its APS model isn't just predictive, but transformative: it redefines how the company manages territory planning and account distribution.
As you might expect, comprehensive data is the foundation of the APS system. Snowflake Sales Data Science Manager David Gojo, one of the driving forces behind Snowflake's own data strategies, notes that at least one-third of the most critical data features in the APS model come from expansive firmographic data and technographic data from the ZoomInfo Data Cube, feeding over 70 different fields.
This is a B2B account scoring model built on a DaaS pipeline, not a batch import. ZoomInfo's Data-as-a-Service pipeline delivers firmographic and technographic fields as a continuous feed, which means the APS model operates on current signals rather than stale snapshots. For a RevOps team managing a global account universe, that distinction is the difference between territory assignments that reflect reality and territory assignments that reflect what the data looked like six months ago.
"We use enriched data to understand the universe of accounts worldwide. Once our APS system produces a score, we put it in front of field operations leads so they can allocate those accounts as efficiently as possible," Gojo says.
Accounts monitored using ZoomInfo-powered scores showed 90% higher opportunity open rates and 2x higher customer conversion rates compared to unscored accounts. The APS score flows directly into the CRM, where field operations leads use it for territory allocation. High-value account activity triggers Scoops notifications to account owners in real time, surfacing the signal at the moment it matters.
"We notify the account owner of any high-value customer activity detected as part of our Scoops usage," Gojo says.
The data enrichment for account scoring that makes this work is ZoomInfo's 70+ field feed from the Data Cube, augmented with technographic signals that identify which accounts are running technology stacks aligned with Snowflake's product. Accounts with the highest propensity scores performed better in every category and helped make sellers more productive.
Snowflake's data-driven approach also extends to analyzing territory productivity, helping the company reveal key performance drivers. To that end, Snowflake has built visibility into the team's sales activities and how much time is spent inside each sales tool. These insights not only help improve seller productivity and sales forecasting, but also quickly surface challenge areas to coach and support sales reps.
By monitoring productivity, Snowflake analyzes how much top-of-funnel activity plays a role in moving prospects down the funnel and closing the deal. "Does engagement actually translate into any meaningful performance at the end of the year? That's the hypothesis we work from that helps us understand if engagement actually makes sense," Gojo says.
How ZoomInfo's intelligence layer powers Snowflake's future data strategy
Looking to the future, Snowflake's commitment to data is set for continued evolution, with a keen focus on advancing AI-driven account scoring, signal-based personalization, and the long-term development of a sales assistant: all grounded in the verified firmographic and technographic data ZoomInfo provides. Snowflake's future strategy includes a continued emphasis on furthering what's already working well, including the data quality it uses to power its initiatives.
"For me, it's continuing to do what we do. We know that it's working, but we're also quite happy where we are. We've landed on three use cases for where we want to develop generative AI capabilities with the long-term goal of developing a sales assistant that we want to explore," Gojo says.
Gojo and his team have also identified areas for continued innovation, including:
Integrating and maximizing data usage
Determining additional data-driven use cases to drive growth
Enhancing personalized communications, including prospecting
Maximizing every customer engagement so it's as impactful as possible
ZoomInfo's value to Snowflake extends beyond the data layer itself. The GTM Context Graph fuses Snowflake's CRM records, behavioral signals, and ZoomInfo's third-party intelligence into a unified reasoning layer, processing 1.5B+ data points daily. It captures not just what happened in an account, but why, connecting CRM activity, conversation intelligence, and intent signals into a single reasoning surface that makes the B2B account scoring model more reliable over time. This is the difference between enrichment (adding fields) and intelligence (understanding account behavior across every signal source).
Because ZoomInfo surfaces that intelligence through APIs and DaaS feeds as well as native products, Snowflake's data science team can consume it programmatically inside their own scoring infrastructure without being locked into a single front-end. That's the Universal Access principle in practice: the same verified data and the same intelligence layer, available through whichever consumption lane fits the team's architecture. For a data science team building custom models, that means ZoomInfo functions as infrastructure, not a tool they have to log into.
ZoomInfo's data pipeline meets ISO 27001, SOC 2 Type II, and GDPR/CCPA compliance requirements, table stakes for enterprise data pipeline decisions where data governance and auditability are non-negotiable.
"I give a lot of credit to the data services team at ZoomInfo. They've done really a great job working with us to get us what we need," Gojo says.
Measurable outcomes from Snowflake's ZoomInfo partnership
Snowflake's journey vividly illustrates the transformative power of data in sales and marketing. Its partnership with ZoomInfo has not only fueled its growth but also revolutionized its approach to business intelligence.
Snowflake's use of APS and additional data-driven strategies have not only enhanced its territory planning and account management, but also provided a clear ROI by improving rep productivity and efficiency.
ZoomInfo's data quality is independently validated: recognized as a Forrester Wave Leader for Intent Data Providers in Q1 2025 with the highest scores across 8 criteria, and holding 133 No. 1 G2 rankings including Data Quality and Account Data Management. For a RevOps team evaluating whether to build a scoring model on top of a third-party data pipeline, that independent validation matters, it confirms the data foundation is auditable and defensible, not just internally benchmarked.
Building a scoring model on third-party data feeds requires reliable API contracts, consistent field mapping, and continuous enrichment. Snowflake's team validated ZoomInfo's DaaS pipeline against these requirements before committing to the architecture. That due diligence is honest about the technical tradeoffs involved in any enterprise data pipeline decision, and it reflects the kind of evaluation any RevOps team should run before building scoring infrastructure on an external data source.
As Snowflake continues to harness the potential of data-driven strategies, its story is a compelling example of pushing innovation forward with data. With its eyes set on the future, Snowflake's partnership with ZoomInfo and its commitment to evolving its data capabilities promises to set new benchmarks in its industry.
To see how ZoomInfo's data and intelligence platform can power your account scoring model, request a demo.
Frequently asked questions
How does Snowflake use ZoomInfo data in its Account Propensity Scoring model?
Snowflake's APS model draws on ZoomInfo's DaaS pipeline to feed over 70 firmographic and technographic data fields, with at least one-third of the model's most critical features coming from ZoomInfo data. The enriched scores are surfaced to field operations leads for territory allocation, and high-value account activity triggers real-time Scoops notifications to account owners. Accounts scored using ZoomInfo-enriched data showed 90% higher opportunity open rates and 2x higher customer conversion rates, see Snowflake's full case study for the complete breakdown.
What firmographic and technographic data does ZoomInfo provide for account scoring?
ZoomInfo's data platform covers 500M contacts, 100M companies, and 30,000+ technologies tracked across 200+ categories for technographic enrichment. Firmographic data fields include industry classification, employee count, revenue range, and company location, the structural signals that define whether an account fits a given ICP. These fields are delivered via ZoomInfo's DaaS pipeline and can be consumed programmatically via APIs, making data enrichment for account scoring available as infrastructure rather than a manual lookup process.
How does ZoomInfo's DaaS pipeline integrate with a CRM for territory planning?
ZoomInfo's Data-as-a-Service pipeline delivers enriched firmographic and technographic data as a continuous feed into a CRM system, not a batch import. In Snowflake's case, the enriched data populates 70+ fields that feed the APS model, and the resulting scores are pushed back into the CRM for territory planning by field operations leads. This continuous enrichment model means territory assignments reflect current account signals rather than a six-month-old snapshot, and ZoomInfo's pipeline meets ISO 27001, SOC 2 Type II, and GDPR/CCPA compliance requirements for enterprise data pipelines.
What results did Snowflake see from ZoomInfo-powered account scoring?
Accounts monitored using ZoomInfo-powered scores showed 90% higher opportunity open rates and 2x higher customer conversion rates compared to unscored accounts. The APS model also improved seller productivity by surfacing high-value account activity in real time, enabling district and regional leaders to allocate accounts more efficiently. Snowflake's Sales Data Science Manager David Gojo credits ZoomInfo's data services team with enabling the data infrastructure that powers these outcomes.
How does ZoomInfo support data-driven territory planning for enterprise sales teams?
ZoomInfo provides the firmographic and technographic data foundation that enterprise sales teams use to build territory models, TAM analyses, and account scoring systems. By delivering data as a continuous feed rather than a batch import, ZoomInfo ensures territory assignments reflect current account signals rather than a stale snapshot. The GTM Context Graph adds a reasoning layer on top of the data, connecting CRM activity, behavioral signals, and third-party intelligence to surface which accounts are most likely to convert, making data enrichment for account scoring a foundation for ongoing territory planning decisions, not a one-time exercise.