Skip to content

How to process DROP requests

Download the batch ZIP from DROP, standardize and hash your own records with DROP's rules, match them to each list, delete or opt out every match, then upload one Id,Status row per work item before the reporting window closes. Check DROP's validation results and re-upload any rejected rows.

Last reviewed 22 November 2026

This guide follows one batch from download to completion. It applies whether you work in the DROP portal or through the API; the files are the same. If you want the wider picture first, read the DROP compliance guide.

Before your first batch

  • Select your lists. In the DROP account, select every consumer deletion list whose identifier type you hold: NDZ (name, date of birth and ZIP), Email, Phone, MAID, NameVIN and CTVID. You may leave one out only if it would match a fully duplicative set of consumers. If you hold none, select at least one and report every row as not found.
  • Add a shared mailbox as a contact. Upload results arrive by email to the primary and secondary contacts. A personal inbox on holiday is how rejected rows go unnoticed.
  • Map your systems. List every database, warehouse, CRM, file share and vendor that holds each identifier type. The match has to cover all of them.
  • Test your hashing. Run CalPrivacy's published test vectors through your routine before you touch a real batch. The hash checker shows the expected hash for any value.

1. Download the batch

In the portal, generate the batch and download the ZIP. Through the API, call GET /data/download; the first call asks DROP to build the package and returns 202 with a Retry-After header until it is ready. See the DROP API for the full set of responses.

CalPrivacy says your processing clock begins once you completely download the batch. Two practical rules follow:

  • Store the ZIP before doing anything else. While a batch is open, DROP will not serve it again on request; the API answers 409 until the batch is complete, and a re-download needs a request to DROP support.
  • Write down the download date. Every deadline in the batch counts from it. The deadline calculator turns it into a report-by date.

2. Read the files

The ZIP holds one CSV per selected list, named <YYYYMMDD>_<DataBrokerId>_<DataType>.csv, for example 20261120_4821_Email.csv. A list with nothing new still has its file, with only the header row. Each list file looks like this:

Id,Hash
A7kP2xQ9Lm4R,KA18MT/ph6IHYjzT9zwETySDQyvSh87YuoSBpOQtkhE=
bK8rT3nV6pZa,mKDnDvwF2inxrKcK1hJN2TRkxPfL6kzNNTtU12eH8Bw=
  • Id is the work item ID: 12 letters and digits, case-sensitive. A work item is one consumer and one identifier. Two consumers who share a phone number arrive as two work items with the same hash.
  • Hash is the SHA-256 of the standardized value, in standard Base64: 44 characters ending in =. For NDZ and NameVIN it is a composite hash of several fields.
  • Your first batch for a list is the whole list. Later batches contain only what is new or changed since your last completed batch.

If consumers withdrew requests, the ZIP also has a ..._Removed.csv with the columns Id,Hash,List. You report nothing for those rows. Use them to take the identifier off your suppression list for that request, unless the same hash is still on a list for another consumer.

3. Standardize and match

DROP standardizes each identifier before hashing it, so you must apply the same rules to your own records and compare hashes exactly. A few rules catch people out:

  • Email: remove all whitespace and lowercase. Keep dots and plus signs; do not fold Gmail addresses.
  • Phone: remove any extension first, keep digits only, then take the last 10.
  • ZIP: drop a ZIP+4 suffix, keep letters and digits, strip leading zeros, then keep the first five characters. 00712345 becomes 71234.
  • Names: Unicode NFKC, lowercase, DROP's transliteration table, then letters and digits only. María José becomes mariajose.

Every rule, with worked examples, is in DROP hashing and standardization. Where you hold several first names, last names or ZIP codes for a person, hash each combination; the regulations ask you to apply any standardization you know will make a match more likely.

For each work item, count the distinct consumers in your records whose hash equals the one on the list. That count decides what happens next.

4. Act on matches

Consumers matchedWhat you doStatus
NoneKeep the hash on your suppression list.5 Not found
One, with non-exempt dataDelete all non-exempt personal information, including inferences, in every system; direct service providers and contractors to delete.3 Deleted
One, all data exemptKeep the data, use it only for the exempt purpose, record the basis.2 Exempted
More than oneOpt every matched consumer out of sale and sharing; tell service providers.4 Opted out

Do not contact the consumer to verify the request. Deletion can mean erasing, deidentifying or aggregating; backups may wait until they are restored or accessed. See DROP status codes for the borderline cases and DROP exemptions for what may be kept.

Keep a record of who confirmed each deletion and when. An auditor will ask, and the status you upload is a statement that the deletion happened.

5. Build the response files

Each response file reuses the downloaded file name and has exactly this header:

Id,Status
A7kP2xQ9Lm4R,3
bK8rT3nV6pZa,5
  • One row for every work item in every list file of the batch. Nothing for the Removed file, and nothing for header-only files.
  • Each ID at most once per upload, and only IDs that DROP delivered to you.
  • Status is 2, 3, 4 or 5. Codes 0 and 1 are unused.
  • UTF-8 CSV, uploaded as individual files, not zipped.
  • A second file for the same list needs a suffix: 20261120_4821_Email_part02.csv. The suffix is up to 10 letters or digits after an underscore. DROP refuses a file name it has already seen for the current download.

The response file checker runs these checks in your browser before you upload.

6. Upload and check

Upload in the portal or with POST /data/upload. One request takes up to 10 files and 100 MB; a batch takes up to 100 files across all uploads. DROP checks files first and rejects a whole file for a wrong type, a bad header, an unreadable encoding or a duplicate name. Then it checks every row and sends the result by email (and by webhook, if you set one up).

Rejected rows are listed in the DROP portal through a link in that email. Fix them and send them again through the normal upload, in a new file with a new suffix. Do not use the amend route for rejected rows; amend is for changing a status DROP already accepted. Wait for one upload's validation to finish before sending the next.

The batch is complete when nothing is outstanding. Only then can you download the next one, so a forgotten rejected row also blocks your next cycle.

7. After the batch

  • Suppress. Keep unmatched requests' hashes and screen new data against them before you sell or share it. See the DROP suppression list.
  • Amend when things change. If a request you reported as 5 later matches data you acquire, report the new status with /data/amend (or the portal's Amend button) at your next access. Amending does not restart any clock.
  • Keep the evidence. The download date and counts, your match results, deletion confirmations, the files you uploaded and DROP's answers. Keep them at least six years for the Delete Act audit.
  • Book the next download inside the access window. See DROP deadlines.

CalDrop Compliance runs these seven steps in one place, matching in your browser so your customer file is not uploaded. The live demo walks a practice batch through them.

Questions

How long do I have to process a DROP batch?

CalPrivacy asks brokers to report within 45 days of the download, and every response for the open batch must be in before the next download. From 1 January 2027 brokers must download at least every 30 days under AB 883, which shortens the practical window.

What goes in a DROP response file?

A UTF-8 CSV with the header Id,Status and one row per work item: the 12-character work item ID from the downloaded file and a status of 2, 3, 4 or 5. The file reuses the downloaded file's name, with a suffix if you send more than one file for a list.

Do I report a status for removed work items?

No. Rows in the Removed file are withdrawn requests. You report nothing for them and take the identifier off your suppression list for that request.

What do I do with rejected rows?

Correct them and upload them again through the normal upload, in a new file with a new suffix. The amend route is only for changing a status DROP has already accepted.

Can I download the same batch twice?

Not on demand. While a batch is open, DROP refuses another download until the batch is complete. A re-download has to be requested from DROP support, so store the ZIP safely as soon as you have it.

Do I need to contact the consumer to verify a DROP request?

No. The DROP regulations say a broker must not contact the consumer to verify the request. Where an identifier matches more than one consumer, you opt them all out of sale and sharing instead.

Sources

  1. CalPrivacy: Processing DROP requests
  2. CalPrivacy: DROP operations and API documentation
  3. CalPrivacy: DROP technical reference
  4. CalPrivacy: DROP help for data brokers
  5. DROP regulations, 11 CCR §§ 7600–7622 (PDF)
  6. California Civil Code § 1798.99.86

Written by the CalDrop Compliance team at Preptics LLC and last reviewed 22 November 2026. This guide explains the rules; it is not legal advice. CalDrop Compliance is an independent product of Preptics LLC. It is not affiliated with, endorsed by or operated by the State of California or the California Privacy Protection Agency (CalPrivacy).