What Is the Digital Services Act? – Definition and How It Works

Learn what the DSA requires of platforms, and how its researcher data access provisions change web data work.

The Digital Services Act is the EU regulation governing online intermediaries and platforms. It concerns illegal content, transparency, and accountability — and it contains provisions that are directly relevant to anyone who studies platform data. This page is a general explanation, not legal advice.

1. Tiered Obligations

Obligations scale with size and role:

VLOPs carry the heaviest duties, including systemic risk assessment, independent audits, and researcher data access.

2. Core Requirements

• Notice and action mechanisms for illegal content.
• Statements of reasons for content moderation decisions.
• Transparency reporting on moderation activity.
• Advertising transparency, including a public ad repository for VLOPs.
• Recommender system transparency, and for VLOPs a non-profiling option.
• Dark pattern prohibitions in interface design.
• Trader traceability on marketplaces.

3. The Provision That Matters Most Here

Article 40 requires VLOPs to provide vetted researchers with access to data for studying systemic risks in the EU.

This is a genuine shift. Historically, studying platform behaviour meant scraping, because platforms restricted or withdrew API access — and researchers were left in an uncomfortable position, technically breaching terms of service to do public-interest work.

The DSA creates a lawful route: vetted researchers, affiliated with a research organisation, applying through a defined process for data relevant to systemic risk.

4. What That Changes in Practice

For academic researchers: apply for access rather than scraping. It is lawful, the data is richer than anything scrapeable, and it does not depend on evading defences.

For commercial data collection: it does not apply. The mechanism is for vetted researchers studying systemic risks, not for competitive intelligence or product data. Commercial actors remain where they were.

For everyone: the existence of a sanctioned route slightly weakens the argument that scraping was the only option — at least for the research use case, and at least for VLOPs.

5. Advertising Repositories Are Publicly Useful

VLOPs must maintain public repositories of the advertisements they carry, including the advertiser, the content, the period, and targeting parameters.

These are public, structured, and sanctioned — which makes them a legitimate data source for studying advertising without any collection ambiguity. For anyone researching ad practices, this is a better starting point than scraping ad surfaces.

6. Transparency Reports as Data

Platforms publish moderation transparency reports on a defined cadence, and the DSA pushes toward a common format. That produces comparable cross-platform data on moderation volumes and outcomes — again, published rather than collected.

The general pattern is worth noticing: the DSA converts several things that previously required scraping into published datasets. Checking whether a required disclosure already exists is now a sensible first step.

7. What It Does Not Do

• It does not create a general right to platform data. Access is for vetted researchers, for a defined purpose.
• It does not legalise scraping. Terms of service, copyright, database rights, and privacy law all still apply.
• It does not cover non-EU platforms outside its scope, though reach is broad where EU users are served.
• It does not address commercial data collection at all.

8. Summary

The Digital Services Act imposes tiered obligations on online intermediaries around illegal content, moderation transparency, advertising disclosure, and recommender systems, with the heaviest duties on very large platforms.

Its most consequential provision for data work is vetted researcher access to platform data for studying systemic risks — a lawful alternative to scraping for public-interest research, though not for commercial collection — and its mandated ad repositories and transparency reports turn several previously scraped datasets into published ones worth checking first.

Potential Related Glossary

• Web Crawling
• Data Fusion
• Unstructured Data
• Data Wrangling