The export is where ad research either becomes an asset or disappears into a slide deck. A strong export keeps the human-readable CSV, the machine-readable JSONL, the full JSON context, and a manifest that records how the dataset was produced.

Why this is not generic advice

This page is written from Pullmesh source-package behavior, buyer handoff boundaries, and the actual operational controls described on the product page: capability numbers, export surfaces, approval gates, provenance, and exclusions.

Key takeaways

  • CSV is for fast review; JSONL and JSON preserve richer structure.
  • Every export needs query metadata and run boundaries.
  • Asset rows and transparency fields should stay connected to the source ad.
  • A manifest prevents future readers from guessing how the dataset was made.

Use each format for its real job

CSV is the lowest-friction format for strategists and analysts. It is easy to open, filter, and annotate. It is also easy to flatten too much information and lose nested fields.

JSONL is better for pipelines because each record can be processed independently. JSON is better for complete context, manifests, grouped summaries, and nested fields such as assets, variants, transparency rows, or advertiser rollups.

Preserve the run context

A useful export records the query, country, filters, date range, platform selection, active status, target count, collection time, job identifier, and any partial-run notes.

Without that context, the dataset becomes hard to defend. A buyer cannot tell whether an observed pattern comes from the market or from the filter choices used during collection.

Keep creative and transparency fields connected

Ad intelligence depends on the relationship between the ad, the advertiser, the creative, the destination, and the transparency fields. Splitting those into files is fine, but the keys need to join cleanly.

The Pullmesh Meta package is positioned around exports that include ads, advertisers, assets, transparency rows, summaries, manifests, and optional raw records so analysts can move between review and verification.

Treat the manifest as part of the product

A manifest should say what is inside the export, when it was generated, which filters were applied, how many records were included, and which files represent raw versus derived output.

That small file is what makes the export usable by someone who did not run the original job. It also gives procurement, legal, or technical reviewers a clean starting point for questions.

Export structure checklist

  • Ads table with stable identifiers and primary review fields.
  • Advertiser rollup with page or advertiser-level summary fields.
  • Asset rows joined back to source ad identifiers.
  • Transparency and regional fields preserved when available.
  • JSONL for pipeline ingestion and raw-record inspection.
  • Manifest with query, filters, counts, timestamps, and exclusions.

pullmesh package

Meta Ad Library collection console

This article maps to the working source package rather than a generic content campaign. Review the product scope, proof points, exclusions, and handoff path.

Open product

FAQ

Which export format should I use first?

Use CSV for fast human review, JSONL for data pipelines, and JSON for complete context or nested output. Serious handoffs often need all three.

Why keep raw records if the CSV exists?

Raw records give technical reviewers a way to verify derived fields, inspect nested data, and recover details that were flattened out of the CSV.

What makes an export client-ready?

Clear filters, stable identifiers, readable summaries, source-linked assets, and a manifest that explains exactly what the client is receiving.