Hightouch

Overview

Hightouch started as a reverse ETL tool and no longer describes itself as one. Its own framing today is an agentic marketing platform, with reverse ETL listed as one product among eleven and offered as the entry-level free tier.

The architectural promise is unchanged and still the reason data teams pick it. By default no customer data is copied into Hightouch; queries run directly against your warehouse, and it stores only the minimal metadata needed to work out what changed.

Event collection is the documented exception, and a sensible one: its own SDKs collect events but write them into your warehouse as tables, so your warehouse remains the source of truth.

Identity resolution runs in the warehouse too, building a deterministic identity graph and a golden record table, with probabilistic matching processing some data outside it.

Destination coverage is the widest of the three, more than 300 tools across 24 published categories, with webhook and HTTP destinations for anything not listed.

Modelling is flexible. Audiences can be built in SQL, in a visual builder for non-SQL users, or driven from dbt Cloud, Fivetran, Airflow, Dagster or Prefect.

Read the plan gating before designing around frequency: free and self-serve plans cap syncs at hourly, and minute-level cron scheduling needs the business tier, as does identity resolution.

Syncs are incremental by change capture, comparing this run's results against the last and sending only what was added, changed or removed, with real-time personalisation on separate infrastructure.

Its privacy controls are the strongest of the three. Columns carry privacy levels of blocked, redacted or sync-only, redacted values never appear in previews, and potential personal data is flagged automatically for review.

The trade-off is price transparency: Hightouch publishes no monetary figures at all, in any currency, and bills on active syncs rather than tracked users.

Key features and capabilities

The same headings are used for every reverse etl & activation entry, so two tools can be read side by side.

Where data comes from

• Warehouses including Snowflake, BigQuery, Databricks, Redshift and Azure Synapse
• Databases including PostgreSQL, MySQL, Oracle, SQL Server and SingleStore
• Files and object storage on S3, Google Cloud Storage, Azure Blob and SFTP, plus Sheets and Excel
• Query engines and tools including Looker, Sigma, ClickHouse and Trino
• It also collects events itself, through SDKs for web, mobile and server, and streaming sources

Business tools it syncs to

• More than 300 tools, the widest coverage of the three
• 24 published categories including advertising, analytics, CRM, email, payments, sales and lakehouses
• Webhook and HTTP request destinations cover anything not pre-built
• Free and self-serve plans put no limit on destinations, destination types, seats or sync runs
• Destination support tiers are not published

Defining audiences and models

• Models defined in SQL against the source, or driven from dbt Cloud, Fivetran, Airflow, Dagster or Prefect
• A visual audience builder for people who do not write SQL, plus natural-language exploration
• Deterministic identity resolution producing queryable identity graph tables in your warehouse
• A golden record gives a unified one-row-per-identity view
• Identity resolution is business tier only, and graphs are per-source rather than global

Sync behaviour and scheduling

• Scheduling by interval, custom recurrence or full cron expressions, at minute granularity
• Free and self-serve plans are capped at hourly; minute-level cron needs the business tier
• Incremental by change capture, sending only rows added, changed or removed since the last run
• Real-time personalisation runs on separate infrastructure rather than a batch schedule
• Configurable alerts, a completed-with-errors status for rejected rows, and documented retry handling

Security, consent and observability

• SOC 2 Type 2 and ISO 27001, with support for HIPAA, GDPR and CCPA compliance
• SAML single sign-on, SCIM provisioning, granular role-based access control and searchable audit logs
• Per-column privacy levels of blocked, redacted or sync-only, with redacted values hidden from previews
• Potential personal data columns are flagged automatically for review
• Least privilege by design, needing only limited read access, with bring-your-own-bucket for retention control

Beyond syncing

• A full composable customer data platform, with audience building, match boosting and suppression
• AI decisioning uses reinforcement learning to choose message, channel and timing per individual
• Ad Studio generates ads from approved assets, with content assembly and lifecycle marketing
• Real-time personalisation, including same-session, with the personalisation API as a paid add-on
• Event collection and an analytics product of its own

Where it runs and what it costs

• Cloud only, with no self-hosted option; execution pushes down against your warehouse
• Nine regions across AWS, Google Cloud and Azure, including Ireland, Belgium, Mumbai and Sydney
• Region is chosen when the workspace is created and cannot be changed afterwards
• Private connectivity on AWS and Google Cloud is region-limited
• Billed on active syncs, meaning any sync that ran at least once in the previous calendar month

Pricing

Price on requestFree for two active syncs

Hightouch publishes no monetary figures anywhere, in any currency. The structure is usage-based on active syncs, with no limits on sources, destinations or seats and no tracked-user charges. The free tier covers two active syncs with unlimited destinations and seats, capped at hourly frequency. Self-serve allows ten active syncs and 100 million operations a month, at an unpublished price. The business tier is a platform fee plus consumption, with add-ons for customer studio and the personalisation API, all quote-only. Trial length is not published.

Vendor pricing page →

Demos and videos

• AI Decisioning four-minute demo Product demo · hightouch.com
• AI Decisioning deep dive webinar Webinar or talk · hightouch.com

About Hightouch

Hightouch was founded by Tejas Manohar, Josh Curl and Kashish Gupta, though no founding year and no single headquarters are published; it lists offices in San Francisco, New York, Denver and London. It is private and heavily funded, raising a $12.1m Series A in 2021, a $40m Series B led by ICONIQ, $38m in 2023, an $80m Series C in February 2025 at $1.2bn, and a $150m Series D in April 2026 at $2.75bn led by Goldman Sachs and Bain Capital Ventures. Its about page publishes an unusually precise 803 paying customers, including Spotify, Warner Music, DoorDash and Grammarly.

hightouch.com