The Signals Your CRM Can't See

What Is a Signal?

A signal is any behavioral, usage, or business indicator that predicts a customer outcome, separate from company characteristics. Signals differ from firmographics (company size, industry, location) because they measure what a customer is actively doing and how they're changing. A signal shows that a customer has crossed a threshold, opened a new need, or is at elevated risk of an outcome. Signals are typically multi-dimensional. It's never just one data point; it's a combination of usage pattern, buying behavior, timing, and company characteristics that together predict whether a customer will expand, churn, or become your best account.

What Is a Fit Score?

A fit score is a numeric ranking that quantifies how closely a customer matches the profile of your most successful accounts. Fit scores combine multiple signals and company characteristics into a single number, typically between 0 and 100, that tells your team which customers have the highest potential. A fit score isn't just company size or industry; it's a weighted model that factors in product adoption, buying behavior, market timing, and any other variables that correlate with revenue success in your customer base.

Why Can't Most CRM Standard Fields Predict Customer Value?

Two customers with similar contract values and the same sub-industry label can have completely different trajectories. One is actively adopting new features, expanding usage, engaging with support on advanced use cases, and showing the buying patterns of your highest-value accounts. The other has been flat for two years, uses a fraction of the product, and matches the profile of customers who churn.

Your CRM treats them identically. Same sub-industry, same company size bracket, same product tier. Without deeper intelligence, your team has no way to distinguish them.

A 2025 study from Validity found that 44% of CRM users believe their database contains inaccurate or duplicate records, highlighting the gap between what standard fields capture and what actually matters. The issue isn't that your CRM lacks fields; it's that standard firmographic fields don't predict outcomes. Filtering takes the fields you have and sorts by them. Intelligence discovers the signals that actually predict value, signals that often don't exist as fields in your CRM at all.

What Signals Actually Predict Customer Value in B2B?

GoodWork has found that three categories of signals consistently provide the most predictive power when modeling B2B customer bases against real outcome data. Most of them surprise the teams who've been relying on firmographics.

Product usage patterns. Usage patterns are often the strongest predictor of customer value, and the most overlooked. Which features they've adopted, how deeply they use them, how usage is trending over time, whether they're bumping up against plan limits. In real customer modeling, a specific pattern of feature adoption predicted cross-sell readiness with far more accuracy than any demographic or firmographic field. Your product data is telling you who's ready to expand. Most teams aren't listening.

Buying velocity and engagement behavior. Buying velocity, deal cycle length, which products they evaluated, how they engaged during the sales process, and where they came from all carry signal. Customers who came through referral behave differently than those who came through paid channels. Customers who bought quickly behave differently than those with long evaluation cycles. These patterns cluster, and the clusters predict outcomes.

Business evolution and timing indicators. This is where timing signals matter. A customer who just added a second location, hired an operations lead, expanded to a new market, or got acquired by a PE-backed platform is a different customer than they were six months ago. Macro signals (funding, leadership changes, M&A activity, market expansion) and micro signals (usage spikes, support escalations, feature requests, engagement shifts) tell you when a customer is ready for something they weren't ready for before. Over 70% of B2B companies are expected to rely on predictive analytics for lead enrichment and targeting by 2025, signaling a fundamental shift toward signal-based decision-making.

These signals are almost always multi-dimensional. It's never just one field. It's a combination of usage pattern, buying behavior, timing, and company characteristics that together explain why some customers become your best and others quietly churn.

How Does AI Find Patterns Humans Can't See at Scale?

When you have hundreds of customers and the signals that predict value come from product data, buying behavior, engagement patterns, timing indicators, and company characteristics, the combinations are too numerous to analyze manually.

Consider the scope. Product usage has dozens of dimensions. Buying behavior has its own patterns. Timing signals are constantly changing. Company characteristics multiply across sub-industries, sizes, and stages. Many of the most valuable combinations will be non-obvious.

Humans default to analyzing the dimensions they think matter, usually company size and sub-industry, and miss the interactions that actually predict outcomes. AI models all of the signals against actual revenue, churn, and expansion data and surfaces the ones that matter.

This is not theoretical. In practice, GoodWork sees findings like: "Customers who adopted Feature X within 90 days, came through a specific channel, and show accelerating usage represent 12% of the base and contribute 30% of revenue." That level of specificity is what makes intelligence actionable. And it's a finding no team would surface by sorting spreadsheet columns.

How Do You Get Customer Signals into Your CRM Workflows?

The analysis is only useful if the intelligence reaches your team where they work. That means the scores, the segments, and the signals need to live in your CRM as data your team can filter, sort, and act on.

This is where most approaches fall apart. A consulting firm delivers a deck with segment definitions. An internal team builds an analysis in a spreadsheet. The insights are real, but they never make it into the daily workflows.

The approach that works delivers the intelligence directly into the CRM as native fields. Every contact scored for fit, tagged with signals (cross-sell candidate, expansion potential, churn risk, product-fit indicators), and assigned to a segment. Marketing builds campaigns by segment instead of blasting the entire base. Sales prioritizes outreach based on fit scores and expansion signals. Customer success knows which accounts match the profile of your best multi-product buyers.

GoodWork delivers all of this as native Salesforce or HubSpot fields, with Smart Lists that update dynamically as the intelligence evolves. Contacts flow in and out of lists based on continuous monitoring. The intelligence isn't a one-time analysis that decays. It updates as your data changes. New customers get scored automatically. Existing customers get re-evaluated as their usage evolves, as their businesses change, as new signals surface.

When you can show your board exactly which 12% of your base drives 30% of revenue, what makes them different, and what you're doing about it, that's a fundamentally different conversation than reporting top-line numbers by sub-industry.

Key Takeaways

• Standard CRM fields like company size and industry provide very little signal about which customers will drive growth, expand, or churn.
• The signals that predict customer value come from product usage patterns, buying behavior, and business timing indicators that standard CRM fields weren't designed to capture.
• 44% of CRM users report their databases contain inaccurate or duplicate records, highlighting the inadequacy of firmographic-only segmentation.
• AI-driven signal modeling reveals non-obvious patterns that no human team would discover through manual analysis, delivering actionable specificity like "12% of customers drive 30% of revenue."
• GoodWork embeds signals directly into your CRM as native fields and dynamic Smart Lists, making intelligence operational rather than analytical.
• Once signals are live in your CRM, your entire go-to-market team can prioritize and act based on actual predictors of customer value instead of standard filters.