Hightouch is the leading reverse ETL platform and, increasingly, the leading composable customer data platform — a warehouse-native answer to monolithic CDPs like Segment. If you're reading about reverse ETL in 2026, Hightouch is the company that has done the most to define what the category looks like at scale and to push it beyond simple sync into the broader activation, audience, and personalization layers of the marketing stack.

What It Actually Is

At its core, Hightouch does the same thing as Census: take a SQL query against your data warehouse, push the results to a SaaS destination on a schedule, and handle every operational detail in between. Connect to Snowflake, BigQuery, Databricks, or Redshift on one side; connect to Salesforce, HubSpot, Marketo, Iterable, Braze, Google Ads, Facebook Ads, TikTok Ads, Klaviyo, Zendesk, or one of ~250 other destinations on the other side; define a sync; let it run.

What pushes Hightouch beyond pure reverse ETL is the upward expansion the team has done since about 2022. The product now includes:

• Audiences: a no-code audience builder that lets marketers (not analysts) define cohorts ("US users who completed onboarding but haven't subscribed in 14 days") on top of warehouse tables, with parent/child audience hierarchies and overlap analysis.
• Customer Studio: a marketer-facing UI for building, previewing, and activating audiences across destinations, designed to feel like Segment Personas or Adobe Audience Manager.
• Identity resolution: stitching anonymous and identified user records together at the warehouse level, which historically required a separate identity tool.
• Personalization API: a real-time HTTP layer that lets product teams query "what audiences is this user in?" from within an application, turning the warehouse into a runtime personalization engine.
• AI Decisioning: ML-driven decisioning on top of warehouse data for next-best-action use cases.

The collective bet is that the modern CDP is a thin marketing UI on top of the warehouse, not a separate database with its own ingestion. Hightouch is the most aggressive proponent of that worldview.

The Origin Story

Hightouch was founded in 2018 by Tejas Manohar (formerly an engineer at Segment), Kashish Gupta, and Josh Curl. The founding insight came directly from Tejas's time at Segment: customers were paying Segment to ingest behavioral data into a warehouse, then asking Segment to push that data back out to other tools, and the second half of that loop was poorly served. Segment, the dominant CDP, was structurally not great at warehouse-first activation because Segment had its own database and the warehouse was an afterthought.

Hightouch went through Y Combinator (W19) and launched in 2019, more or less simultaneously with Census. The two companies have been the duopoly of the category ever since. Hightouch raised aggressively — Sequoia, Bain, ICONIQ, Sapphire — and reached unicorn status in 2022.

The strategic move that defined Hightouch's trajectory was the explicit pivot to "composable CDP" in 2022-2023. Rather than positioning Hightouch as one tool in the data engineer's belt, the founders re-cast the company as the answer to "should we buy Segment, or should we build the CDP layer on top of our warehouse?" That repositioning — aimed at the marketing buyer rather than the data team buyer — materially expanded Hightouch's TAM and put it in direct competition with Segment, mParticle, Tealium, and the broader CDP market.

The Opinionated Take

Hightouch is winning the battle for the CDP layer because the warehouse is winning the battle for the customer database. The strategic logic is clean. Five years ago, the customer database was a separate thing — a CDP that ingested events and stored them in its own opinionated schema. Today, the warehouse already contains the canonical customer record (after all, Fivetran pulled in Stripe, Salesforce, and product events; dbt modeled them; the result is a beautiful users table). Maintaining a second customer database in a CDP is increasingly indefensible. Hightouch made the bet that the warehouse would win this contest, and that bet looks correct.

The risk for Hightouch is the same risk Census faces: the warehouses themselves are absorbing the activation surface. Snowflake has built outbound connectors and acquired related companies. Salesforce launched Data Cloud, which is essentially a reverse-ETL-flavored CDP from the destination side. Databricks has shown interest in the activation layer. If a warehouse vendor shipped a fully credible activation product as part of the platform, Hightouch's wedge would shrink.

Hightouch's response has been to expand fast: the audience builder, identity resolution, personalization API, and AI decisioning are all attempts to be too valuable to absorb. The bet is that being the complete marketer-facing CDP UI on top of the warehouse is a deeper moat than being a sync pipeline.

The other thing to say plainly: Hightouch out-marketed Census. Both products do the same core job. Hightouch picked a bigger story (composable CDP, displace Segment) and a bigger buyer (the CMO, not the head of data) and grew faster as a result. It's a textbook case of category creation through narrative.

Where Hightouch Fits in the Stack

Hightouch sits between the warehouse and the operational SaaS / advertising / marketing layer. It reads from any major warehouse via SQL and writes to ~250 destinations across CRMs, marketing automation, ad platforms, customer support tools, and data destinations. It is not an ingestion tool, not a transformation tool, and (in its base form) not a BI tool.