Salesforce to Snowflake integration
Sesame Software builds your Snowflake tables from Salesforce, then captures every change on the schedule you set. No subscriber to maintain, no field mapping, and running in under an hour.
Book a 30-minute demo built around your org and your Snowflake account.
No commitment. Response within 1 business day.
How it works
No developers and no implementation project.
Authorize your org. Objects, custom fields and custom objects are discovered automatically.
Tables and data types are created for you. Choose how often it refreshes.
The first full load runs, then only changes move. Every run is logged.

Full, incremental and Pardot jobs, each running on its own schedule.
Change data capture
The first run loads everything. After that, each run captures just the records that changed, which is the difference between a warehouse refreshed overnight and one that's current all day.
Detection
Each run reads only records changed since the last successful sync. A large org is never re-read end to end for a day's edits.
Cadence
Set the interval per job: seconds for operational reporting, hourly or nightly for batch. No external orchestration tool needed.
Loading
Changes are written with Snowflake bulk loaders, not row-by-row inserts. New fields and objects are added to the schema as they appear.
Recovery
Every job is logged with timestamps and record counts. A failed run resumes from the last successful sync, with no full reload.
Native Change Data Capture is a capable event stream. On its own, though, it isn't a Snowflake pipeline. Here's what each leaves to your team.
Which you need: for sub-second, event-driven triggers in another app, use native Change Data Capture. For a current, queryable copy of Salesforce in Snowflake, use Sesame Software. Note that Snowflake holds the current state of each record; prior versions aren't kept, because history tracking isn't available for Snowflake.
| Native Salesforce CDC | Sesame Software | |
|---|---|---|
| Getting to Snowflake | You build and run a subscriber (Pub/Sub API or CometD) | Point-and-click, no subscriber code |
| Catching up | Events retained for 72 hours | Resumes from the last successful sync |
| Objects covered | Five selected objects by default; more needs an add-on | Your full schema, custom objects included |
| Limits | Counts against a daily event delivery allocation | Incremental reads keep API use low |
| Timing | Near real-time events | Your schedule, as often as every few seconds |
Why it doesn't strain your org
"API-efficient" is a claim every vendor makes. Salesforce meters API calls per org per day, and a tool that ignores that budget either runs out or gets throttled mid-job. Four mechanisms keep Sesame Software inside it.
01
After the first load, runs read only what changed. On a large org, this is the biggest factor in daily API use.
02
Throughput comes from parallel work, not more calls, so volume grows without matching API pressure.
03
Extract and load are tuned separately, so neither side becomes the bottleneck that forces retries.
04
A failed run picks up where it stopped instead of re-reading records and spending calls twice.
Consumption is managed for you, not left to engineering to figure out.
From thousands to hundreds of millions of records, without re-platforming.
Sync as often as every few seconds, with no external orchestration.
Runs in your environment. Data never passes through Sesame Software's servers.
Why teams do this
Tableau, Power BI or Looker on a Snowflake copy, with full history and no load on Salesforce.
Salesforce next to finance and operations data in one warehouse and one query layer.
Your complete, structured Salesforce dataset, ready for your data science team.
Reporting and models query Snowflake, so they never compete for Salesforce API calls.
What's included
Custom fields, custom objects, Pardot and complex org structures, without bespoke development or upkeep.
Add a field in Salesforce and it appears in Snowflake. No mapping, no maintenance.
Fast ingestion into Snowflake that sidesteps common API bottlenecks.
Every few seconds to once a year, with built-in Cron support.
A full one-time load, or incremental refreshes for a live copy.
CRM and marketing data in Snowflake together, one platform.
Straight into your Snowflake account. You control access and retention.

Salesforce objects and custom fields mapped to Snowflake columns automatically.
From teams who run it
"Job creation is very modular, so we can add and remove steps easily to jobs on the fly if needed. Sync is mostly hands off and we will go months without having to log into the server or application for any kind of issues."
Verified user
Investment Banking, G2 review
"Custom objects added in Salesforce are automatically added as new tables in the replicated database. Custom fields added in Salesforce are automatically added as new columns in the replicated database."
Sam F., Senior Web Application Developer
Enterprise, G2 review
"I found it to be the most reliable tool in syncing data to/from Salesforce compared to other named ETL platforms."
Mayank S., Senior Manager CRM Data
Enterprise, G2 review
30 minutes. We'll show a full load and a change being captured, start to finish.
Common questions
You authorize your Salesforce org, point Sesame Software at Snowflake, and it creates the Snowflake tables, converts data types, and runs the pipeline on a built-in scheduler. No scripts, no manual field mapping, and no engineering time to maintain it. New Salesforce fields and objects are added to Snowflake automatically.
Two ways. Incremental sync moves only records changed since the last successful run instead of re-reading whole tables, and patented multithreading gets throughput from concurrency rather than from more API calls. Together they keep high-volume pipelines inside your org's daily API budget without manual tuning.
Yes. After the first full load, each run captures only the records that changed since the last successful sync and writes them to Snowflake with bulk loaders. You set the interval, as often as every few seconds. If a run fails, the next one picks up from the last successful sync, so no changes are skipped and no full reload is needed.
Native Change Data Capture publishes change events, but you build and run the subscriber that writes them to Snowflake. Events are retained for 72 hours, and the default allocation covers five selected objects. Sesame Software replaces the subscriber with a no-code pipeline that catches up from the last successful sync and covers custom objects automatically. For event-driven triggers in another app, use native Change Data Capture. For a current, queryable copy of Salesforce in Snowflake, use Sesame Software.
Automatically. New custom fields and objects are detected and added to Snowflake as new columns or tables, with no manual mapping.
Yes. Run a full one-time load into Snowflake, or keep incremental refreshes running for a live copy. Both use the same setup and are typically running in under an hour.
You're alerted, and every job is logged with timestamps and record counts. The next run resumes from the last successful sync, without data loss or duplication.
No. History tracking, which keeps prior versions of changed records, is available for row-oriented databases but not for Snowflake, which is columnar and loaded in bulk. Change data capture still works: Snowflake always holds the current state of each record. If you need change history, we'll walk through the right approach in your demo.
No. Data moves directly from Salesforce into your Snowflake account. Sesame Software never stores it on its servers, and you control access, retention, and security.
See a full load and a change capture running against an org and Snowflake account like yours.
Book a demo 30 minutes. No commitment. Most teams are running in under an hour.