Upstream CRM records · built on Snowflake

The records that never belonged in the CRM

shunt /ʃʌnt/ · verb To move a rail car onto the track it needs. Shuntline moves leads, consent and campaigns out of the CRM and next to the warehouse, where they already had to go.

Some CRM records are referentials, not workplaces: leads, consent, the campaign catalogue, and the ones your org added over the years. Shuntline does two things with them. CRM Insights detects which entities in Salesforce should not be there, why, and what moving them saves. Solutions give them a sustainable home: adaptable Snowflake Native Apps, installed from the Snowflake Marketplace into your own account. The CRM keeps engagement and receives only the outcome.

Shunting diagram · detected in Salesforce, hosted on Snowflake
Leads Consent Campaigns & UTM
Snowflake Native AppsSnowflake MarketplaceHybrid tablesStreamlitOpenflow · SnowpipeRow access policies · SSOCoWork · Cortex Agents · MCPReverse ETL
Designed for Snowflake

Shuntline is built for Snowflake. Every solution is a Snowflake Native App, installed from the Snowflake Marketplace: records in hybrid tables, screens in Streamlit, and the warehouse as the referential they read from and write to.

The warehouse is the beginning

The data warehouse used to end the cycle: operations ran in the CRM, data was copied out, analytics came after the fact. Shuntline turns it around. CRM referentials and data land in Snowflake first, business users start there, and operations are triggered from there.

Yesterday · the warehouse ends the cycle

Operations in the CRMand in every tool around it, each with its own users and rules
Copied to the warehouseat night, by integration jobs
Analytics, after the factfor data teams; nothing goes back

With Shuntline · the warehouse starts it

Land in SnowflakeCRM referentials and data: leads, consent, campaigns, the context from CRM, ERP and usage
Focus in the Hubbusiness users see their own data, through screens, CoWork artifacts and an assistant, under one set of rules
Trigger operationsin specialized solutions on Snowflake, or down to the tools outside: CDP, CRM, marketing execution

The landing zone, its access rules and how AI agents obey them: Shuntline Hub →

The problem

They used to sit in Salesforce. They should not.

Leads, consent and campaigns all ended up in the CRM because nowhere else was ready. Yet they share a pattern that a CRM is the wrong tool for: they are referentials shared across the whole company, worked mostly by automation, and read far more than written. Next to the warehouse, each becomes one table, one app and one hand-off, instead of a custom object, a tangle of triggers and a seat per reader.

Cross-company

One referential, every business unit

Every brand, country and partner feeds the same list and reads the same answer. In a CRM: record types, sharing rules and a copy per team. In the warehouse: one governed dataset.

High automation

System of record, not of engagement

Few people open these records; their life is receive, check, file, hand off. A seat, a page layout and an assignment rule are the wrong price for a job nobody sits in front of.

Near the warehouse

Upstream dependencies, simplified

The context they need, ERP status, usage, history, attribution, is already in Snowflake. Reading it in place removes the inbound sync and the integration per sender.

1 · Detect, in Salesforce

CRM Insights

An automated audit of your org. Every object and record type gets a scorecard: how much automation it carries, who creates it and who only reads it, which integrations feed it, what it costs in seats. The objects explorer ranks the candidates against the pattern; the visits analysis shows the read-only populations that need a referential, not a CRM seat.

  • which entities should move, and why
  • costs and users per entity, from your contract lines
  • findings as native tables in your Snowflake, re-run at will
CRM Insights →
Scorecards: processes, records, complexity, licences and users at a glance
Scorecards: objects, records, automation, licences, users.
Objects explorer: every object with its records, operations, users and cost
Objects explorer: records, operations, users and cost per object.
scorecardsobjects explorervisits & visited objectscohortslicence inventorycold code & flows
2 · Host, on Snowflake

Solutions

Each entity that should move gets a Snowflake Native App, installed from the Snowflake Marketplace into your account: hybrid tables for the records, Streamlit for the screens, the warehouse for the joins. A handful of screens, rules an administrator edits, an intake for senders, tools for agents. Standalone from the catalogue, or custom on the same skeleton.

  • installed from the Snowflake Marketplace, in your account
  • configured, not coded; every version reviewed by Snowflake
  • the CRM receives the outcome only
Solutions →
Lead management: the lead register
Lead management: every lead as sent, flagged and filed.
Campaign referential: the group dashboard
Campaign referential: one catalogue across business units.
Lead managementOpt-inCampaign referential & UTMcustom, same skeleton
3 · Subscribe, fixed and monthly

Pricing & Services

Fixed monthly prices. CRM Insights is priced per org, with Snowflake pattern detection; each solution has a fixed monthly price, independent of volumes. Custom add-ons are quoted when you need them, and integration partners take over when your project needs hands on top of the software.

  • nothing per seat, no licence renewal
  • updates and support included
  • a written quote within a day
Pricing & Services →
CRM Insightsper org · per monthaudit toolkit · unlimited runs · Snowflake pattern detection
Solutionper solution · per month · fixedindependent of volumes · updates · support
Custom add-onson quoteextensions · custom solutions · integration through partners

No seat. No volume metering. Snowflake compute on your own contract.

Questions about Shuntline

Short answers to what architects, CRM owners and marketing ops ask first. Everything here is also in the pages above, in more depth.

What is a reverse-CRM experience?

It moves users and records upstream. People start from their enterprise data in Snowflake, land in one governed space, Shuntline Hub, where they view, analyze or ask Snowflake CoWork, then go down to the solution that does the job: a Shuntline solution on Snowflake, or a specialized app such as Salesforce or a CDP, which receives only what it needs to engage. Reverse ETL already made the warehouse the starting point for marketing, building audiences there and activating them in ad, email and CDP tools; a reverse CRM applies the same logic to the CRM's users and records. Shuntline Hub → · Read the essay: The Reverse CRM →

What is Shuntline?

Upstream CRM records on Snowflake, in two moves. CRM Insights detects which entities in Salesforce are referentials rather than workplaces, why they should leave and what moving them saves. Solutions give them a home: adaptable applications (lead management, opt-in and consent, campaign referential and UTM, or custom ones on the same skeleton) installed as Snowflake Native Apps in your own Snowflake account, sold as a fixed-fee monthly subscription.

What does "upstream CRM records" mean?

Records that exist before and around the CRM rather than inside it: the lead as it arrived, the consent a person gave, the campaign that was defined once for the whole group. They are cross-company referentials, worked mostly by automation and read far more often than written, which is the opposite of what a system of engagement like Salesforce is built for.

Does Shuntline replace Salesforce or the CRM?

No. The CRM keeps the work it is good at: accounts, opportunities, the conversations of a sales team. Shuntline takes the referentials out of it, hands the CRM what it needs (for example the converted opportunity, never the raw lead) and lets the warehouse join the rest.

Where does Shuntline run and who hosts it?

In your Snowflake account: the applications are Snowflake Native Apps installed from the Snowflake Marketplace, their records live in hybrid tables, their screens are Streamlit, and the warehouse is the referential they read from and write to. Nothing is hosted by Shuntline. A hybrid design with AWS, Google Cloud or Azure alongside Snowflake can be studied on demand.

How do the solutions connect to marketing platforms?

By reverse ETL. Audiences are built in the warehouse from leads, consent and campaign codes, then pushed to Marketing Cloud or other downstream solutions, which execute. Attribution comes back as a join on the campaign code rather than a reconciliation.

What is CRM Insights?

An automated audit of your Salesforce org, delivered by Shuntline. It detects the entities that should not belong in a system of engagement, explains why, and prices the seats, automations and integrations they carry, attributed to the users who touch them. The findings stay on that machine by default; turn on sharing and they are published to your own Snowflake account or AWS S3 bucket, so every team reads the same evidence.

Is Shuntline limited to the solutions listed?

No. The catalogue grows wherever the pattern applies, and a custom solution can be co-designed on the same skeleton (a Snowflake Native App with hybrid-table records, a Streamlit console, a rules engine and agent tools) for any entity the audit surfaced. Some of those custom solutions surface here, in the catalogue, once the record proves common enough; others stay private to the customer who asked for them, delivered with their source code.

See your own data take the shortcut

Bring a few real payloads: a lead, a consent event, a campaign brief. We run them through the apps live, in a Snowflake account, and show where each one lands and what reaches Salesforce.

  • A working session with the team that builds it
  • Your payloads, routed and joined to warehouse context
  • Snowflake architecture, security and hosting questions answered
Book a demo Write to us with your company, your Snowflake region, and the systems that send you leads, consent and campaigns today.