buyer4 min read

Data marketplaces: how they work, and which one fits your need

Snowflake, Databricks, AWS, or an independent marketplace? What each one actually delivers, what you can verify before you sign, and what the EU Data Act changed for anyone brokering data.

4 min read ⏱

Data marketplaces, explained

Cloud-native, independent, vertical — and what each trades away

11 slides · swipe or use the arrows

d-nvest.com1/11

The shift

Buying data stopped meaning buying a file

A marketplace does not hand you a CSV. It brokers a LICENCE and a delivery mechanism — which is why the same dataset can cost very different amounts depending on where you buy it.

d-nvest.com2/11

Family 1

Cloud-native marketplaces

Snowflake, Databricks, AWS Data Exchange, Google Analytics Hub. The data lands INSIDE your warehouse, with no copy to move.

Fastest path: often live in minutes, no pipeline to build
Strong governance — access is revocable, usage is auditable
The trade-off: you are limited to what YOUR cloud hosts

d-nvest.com3/11

Family 2

Independent marketplaces

Cloud-agnostic catalogues. Broader supply, and room to negotiate terms a self-service listing cannot offer.

Wider catalogue, including holders with no cloud presence
Scope, duration and exclusivity are negotiable
The trade-off: you integrate the delivery yourself

d-nvest.com4/11

Family 3

Vertical marketplaces

Built around one industry — health, mobility, finance, geospatial. Smaller catalogues, but the schema and the compliance work are already done for that field.

d-nvest.com5/11

What you actually get

Three delivery models

The model decides freshness, cost, and what happens when the licence ends.

Live share — the seller grants access; no copy leaves their estate
API — you pull on demand, you pay per call or per seat
File drop — a snapshot, cheapest and stalest of the three

d-nvest.com6/11

The question nobody asks

Can you verify it BEFORE you pay?

A listing is a claim. A sample, a schema, a row count and a provenance statement are evidence. Ask for all four — a seller who cannot produce them is telling you something.

d-nvest.com7/11

Provenance

Where did this data come from?

Resale is a NEW purpose. It does not inherit the consent given for the original collection.

Who collected it, from whom, and under what notice
What legal basis allows it to be transferred to you
Whether the seller may sub-licence it at all

d-nvest.com8/11

The EU angle

Two laws changed who may broker data

The Data Governance Act imposes NEUTRALITY on data intermediaries: an intermediary may not use the data it brokers for its own ends. The Data Act, applicable since September 2025, opens access to connected-product data and frames B2B sharing.

┌ European Commission — Data Act & Data Governance Act

d-nvest.com9/11

Pricing

Why the same dataset has three prices

Most disputes come from scope, not from price.

Scope — one team, one entity, or the whole group
Duration — a snapshot, or a refreshed feed
Exclusivity — non-exclusive is the norm and the cheapest

d-nvest.com10/11

Takeaway

Which one for which need

Start by seeing what is actually available.

Already on Snowflake/Databricks and need speed - cloud-native
Need supply that no cloud hosts, or custom terms - independent
One industry, compliance-heavy - vertical
In every case: sample, schema, provenance, licence - before paying

d-nvest.com11/11

Browse available datasets

Questions about monetising or buying data?

Talk to an expert — no strings attached.

Book a free 30-min call

The full guide

Data marketplaces have moved from a niche idea to a normal way of buying data, and the vocabulary has not caught up. Buying data on a marketplace does not mean receiving a file: it means acquiring a licence and a delivery mechanism. That distinction explains why the same dataset carries very different prices depending on where you buy it. There are three families. Cloud-native marketplaces - Snowflake Marketplace, Databricks Marketplace, AWS Data Exchange, Google Analytics Hub - deliver data inside the warehouse you already run. Nothing is copied, access is revocable, and you can be querying in minutes. The limit is supply: you only see what your cloud hosts. Independent marketplaces are cloud-agnostic, with a broader catalogue and terms you can negotiate, at the cost of integrating delivery yourself. Vertical marketplaces cover one industry and arrive with the schema and compliance groundwork already done. Whatever the family, three delivery models exist: a live share, where the seller grants access and no copy leaves their estate; an API, priced per call or per seat; and a file drop, the cheapest and the stalest. The model decides freshness, cost, and what happens the day the licence ends. The question buyers skip is verification. A listing is a claim; a sample, a schema, a row count and a provenance statement are evidence. Provenance is the one that bites: who collected the data, under what notice, and whether the seller may sub-licence it at all. Resale is a new purpose - it does not inherit the consent given for the original collection. Two European laws now shape this. The Data Governance Act imposes neutrality on data intermediaries, which may not exploit for themselves the data they broker. The Data Act, applicable since September 2025, opens access to data generated by connected products and frames business-to-business sharing. Finally, price. The same dataset carries different prices because scope, duration and exclusivity differ - and most disputes come from scope, not from the number. If you already run Snowflake or Databricks and need speed, start cloud-native. If you need supply no cloud hosts, or terms a self-service listing cannot offer, go independent. In every case, ask for the sample, the schema, the provenance and the licence before you pay.