← Back to blog

California’s DROP Changes the Data Broker Deletion Pipeline

California’s DROP platform turns one consumer request into a recurring deletion obligation for data brokers. Here is what developers should build—and first determine.

California’s Delete Act has changed deletion from a one-off support workflow into a recurring system obligation for covered data brokers. Since January 1, 2026, residents can submit a single request through the state’s Delete Request and Opt-out Platform, or DROP, asking active data brokers to delete their personal information.

The scope is narrower than “any company that processes analytics.” California defines a data broker around a business that collects consumers’ personal information from various sources and sells it to other companies, even when the consumer did not interact with that business directly. Whether a particular analytics provider or site operator meets the definition depends on its actual data flows and business model, not the word “analytics” in its product description.

What changed in 2026

The state platform is the consumer-facing intake point. Beginning August 1, 2026, a data broker must access the platform at least once every 45 days. After receiving a request, it generally has 45 days to process the deletion and must direct associated service providers or contractors to delete the relevant personal information as well.

The Act also makes deletion ongoing: after an initial request is processed, a broker must delete the consumer’s personal information at least every 45 days, subject to statutory exceptions. It also restricts selling or sharing new personal information about that consumer after the request, unless an exception applies or the consumer asks otherwise. A single deletion job is therefore not a sufficient implementation.

Determine whether your system is in scope

Start with a data-flow inventory, not a vendor label. Identify where personal information originates, whether you collect it directly from the person or receive it from other sources, who receives it, and whether your business sells it. Review each entity’s role separately: a customer site, an analytics provider, and a downstream data broker may have different obligations.

Do not treat every third-party analytics script as a data broker, and do not assume that pseudonymous event records are automatically outside California’s definition of personal information. The right legal classification depends on the information and context. If your business appears to meet the statutory definition, use the California Privacy Protection Agency’s registration and DROP materials to confirm the applicable process.

There are two state lists to understand. The Data Broker Registry records businesses that operated as brokers during the previous calendar year and is published on an annual schedule. DROP reflects businesses with active platform accounts and is updated as those accounts change. They answer different questions; a registry export is not a substitute for checking the operational deletion channel.

Build deletion as a durable workflow

Model each verified request as a durable instruction with a received time, processing status, and limited audit evidence. Keep identity matching in a restricted workflow, then propagate a deletion task to the systems that hold the covered records: primary databases, derived tables, caches, exports, and service providers. Record completion or a documented exception per destination rather than marking the entire request complete when one database responds.

The recurring check matters just as much as the first pass. Schedule platform access within the statutory interval, make retries safe, and surface failed destinations before the processing window closes. A request may need to reach a processor that has its own queue or service-level boundary; contracts and operational interfaces should make that propagation possible.

Design the workflow to minimize new data. Do not copy a full consumer profile into every processing system just to route a deletion. Use a short-lived, access-controlled request reference where possible, and retain only the evidence needed to demonstrate that the request was handled. Deletion logs themselves should have a defined purpose and retention period.

For analytics teams, this is a reason to document what is collected and where it goes before building a request handler. A small, first-party dataset with clear ownership is easier to classify and reconcile than a chain of tags, enrichment vendors, and long-lived exports. DROP does not make every analytics operator a data broker; it makes role clarity and deletion propagation operational requirements for the businesses that are.

Sources

Comments

Loading comments…