Alation is the company that, more than anyone else, turned "data catalog" from a dusty metadata tool into a real software category. Founded in 2012 in Redwood City by Satyen Sangani (ex-Oracle), Venky Ganti (ex-Microsoft Research), Aaron Kalb (ex-Apple, where he helped build Siri), and Feng Niu (a Stanford PhD), Alation was the first of what we now call the "second wave" catalog vendors — arriving four years after Collibra but several years before the modern-stack era of Atlan, DataHub, and Select Star.

For most of the 2010s, Alation was the answer. If you were a Fortune 1000 data team looking for something more usable than Informatica or IBM InfoSphere, you bought Alation. The company essentially defined what an enterprise catalog should look like in the cloud era and was rewarded with a unicorn valuation and hundreds of large customers.

In 2026 Alation sits in an uncomfortable but valuable position: too modern to compete with Collibra for the most regulated RFPs, too enterprise-flavored to compete with Atlan and DataHub for cloud-native startups. It owns the middle — and the middle is both the largest and the hardest place to defend.

The Big Idea: Behavioral Metadata

Alation's founding insight, which remains its most distinctive feature, is behavioral metadata. Rather than relying only on what humans tell the catalog about a table (descriptions, tags, glossaries), Alation watches the query logs of the underlying warehouse and infers context from actual usage.

If 40 analysts run a SELECT against finance.revenue_by_region every Monday morning, Alation concludes that the table is important, rank it high in search, and surfaces the top queries written against it. If a column has never been referenced in a SELECT, Alation knows it is probably dead. If a particular analyst keeps being the source of complex joins against dim_customer, that person becomes an implicit subject-matter expert the system suggests when someone else asks a question.

This was a genuinely novel idea in 2012 and it still holds up. Alation called the resulting pattern the Behavioral Analysis Engine, and it remains a distinctive piece of the product. Modern catalogs have largely caught up on query-log parsing, but Alation's implementation is still regarded as among the deepest and is the spiritual ancestor of most usage-based ranking in the category.

What Alation Actually Does

Alation is the standard set of catalog features executed at enterprise depth:

Search and discovery. A Google-like search box over every table, column, dashboard, query, and report, ranked using the behavioral signals described above. Alation was one of the first catalogs to make search the front door, an idea Amundsen later pushed even harder.

Business glossary and stewardship. A glossary of business terms (ARR, active user, region) linked to the underlying technical assets, with assignable stewards, approval workflows, and certifications. This is the piece that most modern startups have underinvested in and Alation has spent a decade polishing. For regulated customers, it is often the main reason they pick Alation over newer vendors.

Lineage. Alation offers table-level lineage across warehouses, ETL tools, and BI platforms, with column-level lineage added in more recent releases. The column-level implementation is serviceable but is generally considered less deep than Atlan's or DataHub's, particularly across complex dbt projects.

Compose and query authoring. Alation includes a SQL authoring tool called Compose that lets analysts write queries directly against connected warehouses with in-editor suggestions pulled from the catalog's knowledge. This is an unusual feature — most catalogs stop at discovery and leave the querying to a BI tool — and it reflects Alation's original conviction that the catalog should be where analysts actually work.

Alation Cloud Service (ACS). Originally deployed on-prem or single-tenant, Alation now offers a managed cloud service. The rollout has been slower than some competitors would like, partly because many of Alation's biggest customers explicitly want single-tenant or self-hosted deployments for regulatory reasons.

Alation AI Agents. Like every catalog vendor in 2025–2026, Alation has added a suite of AI assistants that generate descriptions, answer questions about metadata in natural language, and suggest stewards — a competitive necessity that every vendor in the category now ships in some form.

The Opinionated Take

Alation is the established middle, and the established middle is a hard place to be. The company built a genuinely excellent product, earned its unicorn status honestly, and remains one of the two or three catalogs every enterprise shortlists. The problem is not the product; it is the market structure.

When a regulated Fortune 500 buys a catalog for the first time, they tend to shortlist Collibra first because Collibra has spent 15 years being the default answer at banks and pharma. When a cloud-native company buys a catalog for the first time, they tend to shortlist Atlan and DataHub because those are the tools their peers talk about on podcasts and at dbt Coalesce. Alation's ideal customer — a large-enterprise-but-not-a-bank company modernizing onto Snowflake — exists in real numbers, but it is a narrower niche than either of the adjacent buyer segments.

Where Alation wins. It wins where behavioral metadata, stewardship workflows, and deep warehouse integration matter more than slick UI. It wins at large companies with long procurement cycles that want a catalog vendor with a real services arm and a track record of surviving CDO transitions. It wins on feature depth when the buyer actually reads the spec sheet.

Where Alation loses. It loses to Atlan on design and on new cloud-native logos. It loses to Collibra on pure governance weight and regulated-industry brand. It loses to DataHub on technical depth and open-source credibility. And it loses to warehouse-native catalogs (Unity Catalog, Snowflake Horizon) when the customer is comfortable being locked into a single platform.

The honest prediction is that Alation continues to be a significant private company with deeply entrenched customers, occasionally loses the sexiest new deals to Atlan, and eventually is either acquired by a larger data platform vendor or finds a durable niche as the catalog for traditional-enterprise-Snowflake customers. It is a good company, not a dying one, but the narrative momentum in the category has clearly moved elsewhere.