ATTEXDocsattexdata.ai ↗
Contents

Moving from the legacy mobile export

For customers who integrated the mobile export before October 2026. The current mobile export is requested with "version": "v2" on Create an export. The legacy export (v1) ended on 11 November 2026: its days up to then stay available, and a request for a later day is a 400 V1_NOT_AVAILABLE_AFTER. A job that sends no version keeps getting the legacy export until you switch. Desktop is unchanged. Versions →

What changes

Column Change What to do
demographics.date_of_birth Removed Use age in demographics. The current export carries no date of birth.
evidence_metadata Removed Use brand, brand_evidence and ad_text.
brand New The advertiser by its public name.
brand_evidence New The exact text the brand was read from.
other_brands New Other brands named in the ad, [] when there are none.
industry New Top level of the IAB Ad Product Taxonomy 2.0.
industry_path New The full IAB category.
ad_text New The advertiser's own words in the ad.
creative_url New Link to the ad's image or video file.
app_name New The app by the name it shows on the phone.
ad_format New Where in the app the ad ran: feed, reel, story, in-stream and so on.
media New video or display.
size_group New The ad's shape by height ÷ width, from banner to full.
display_size New The standard display size a browser ad matches, for example 300x250.

Every other column keeps its name and type.

What the rows now mean

last_seen is when the ad left the screen: the device's next screen change after the ad was last captured, in the same app or another, at most 60 seconds later and 10 minutes after first_seen, and earlier when the camera stopped. In the legacy export it was the last capture of the ad. duration_s, eyetracker_enabled, fixation_duration_ms and is_fixated cover first_seen to last_seen, so the same impression shows longer time on screen and more fixation than in the legacy export. Compare within one export, not across the two.

impression_id values do not pair between the two exports, and a rebuilt day can give its rows new ids. Merge by day, not by id.

How to switch

  1. Add "version": "v2" to the create call. Same key, same endpoints, same files.
  2. Update your reader. Drop date_of_birth and evidence_metadata; add the twelve columns.
  3. Move to the change feed. A first pull, then a daily pull of what changed, upserted on impression_id: the change feed. Staying on the day export, replace whole days instead, as in Day export.
  4. Fetch from 13 March 2026. v2 data starts on 13 March 2026 in Sweden and Denmark, 26 March in Norway and 19 May in Finland; a from_date before 13 March is a 400 V2_FROM_DATE_NOT_AVAILABLE with the first day in available_from.
  5. Tell us at data@attexdata.ai when you have switched.