If you ask your Head of Sales and your Head of Product: "How many active accounts do we have?" do they give you the same number?
In 90% of B2B SaaS companies, the answer is no.
The Product team is looking at track events in Amplitude. The Sales team is looking at "Last Login Date" in Salesforce. The Data team is stuck in the middle, writing one-off SQL queries to try and reconcile the two, creating a "Truth Gap" that slows down every revenue decision.
I believe that metrics should be infrastructure, not opinions.
That is why I built and open-sourced GrowthCues Core. It is a warehouse-native semantic layer designed to give B2B SaaS teams a resilient, standardized foundation for growth.
What is GrowthCues Core?
Technically, it is a robust dbt project that connects to your Snowflake or BigQuery warehouse. It takes the raw chaos of your event stream (from Segment or Rudderstack) and transforms it into clean, business-ready signals.
But strategically, it is the common language for your GTM team.
Instead of arguing over definitions, GrowthCues Core provides pre-built, industry-standard logic for:
- Account Health: Stickiness, Utilization, and Churn Risk.
- Growth Velocity: 7, 14, and 30-day trends for every metric.
- User Lifecycle: Automatically tagging users as New, Resurrected, Dormant, or Churned.
Here is how this standardized layer unlocks value for every part of your Revenue organization.
For Sales: Spotting Expansion Signals (Before the Renewal)
The Problem: Product-Led Sales (PLS) relies on timing. Calling a customer 30 days after they added 50 new users is too late. You need to know the moment momentum shifts.
The GrowthCues Solution:
We don’t just count users; we calculate Velocity. The Core model includes a metric called net_new_users_7d.
- The Play: When an account shows a positive spike in netnew_users_7d, it means they are actively expanding their team _right now.
- The Action: Reverse ETL this signal to Salesforce to trigger a "High Momentum" alert for the Account Executive.
For Customer Success: Detecting "Silent Churn"
The Problem: Most CS teams rely on "Green/Red" health scores that are often subjective or lagging. By the time an account cancels, they’ve usually been inactive for weeks.
The GrowthCues Solution:
I introduced a metric called Volumetric Churn (volumechange_ratio_7d). It measures the _intensity of usage compared to the previous week.
- The Play: If an account is technically "Active" (someone logged in), but their volume ratio drops below 0.5 (usage halved), they are in immediate danger.
- The Action: This signal allows CS to intervene on declining usage, not just zero usage.
For Marketing: identifying the Real Champions
The Problem: Case studies and beta programs need "Power Users." But finding them usually involves asking Engineering to run a CSV export, which takes days.
The GrowthCues Solution:
Our user model automatically calculates an L14 Frequency score (active_days_last_14) and ranks users within their company (usage_rank_in_account).
- The Play: You can instantly query for "Users ranked #1 in Enterprise Accounts who are active 10+ days every fortnight."
- The Action: Build hyper-targeted campaigns for your most engaged advocates without bothering the data team.
Ready for the AI Era
There is one more massive benefit to standardizing your metrics: AI Readiness.
Everyone wants to enable "Self-Serve Analytics" where executives can ask questions to an AI agent. But AI models hallucinate when they don't understand the data schema.
GrowthCues Core solves this by treating Context as Code. The project includes a specialized schema.yml file, heavily tagged with prompt-engineered definitions ([Definition], [Context]).
You can feed this schema directly to Gemini, Claude, or ChatGPT. Because the AI understands your strict business logic (e.g., "Churn Risk means Active last month but 0 events last week"), it can write accurate SQL for you.
Why Open Source?
I'm building a commercial platform, GrowthCues, that goes a step further, adding No-Code Milestones and Customer Journey tracking, and Predictive AI to forecast future churn and conversion.
But I realized that the descriptive layer, the math that defines "What happened yesterday", shouldn't be a trade secret. It should be a standard. Every GTM Engineer shouldn't have to rewrite the same "Monthly Active User" SQL from scratch.
By open-sourcing the Core, I hope to give every B2B SaaS startup a solid foundation to build upon.
Star the Repository on GitHub and start building your growth stack on a rock-solid semantic layer today.