Dynamic Search Ads: how they still work and the migration path into AI Max

Dynamic Search Ads still run on Google Ads, generate headlines and landing pages from your site content, and remain fully targetable and excludable, but they now sit on a fixed migration path into AI Max that begins in February 2027. This guide focuses on how DSA works today, how to keep it under control, and how to structure your upgrade and experiments before you are moved into AI Max, which is covered in more depth in this AI Max for Search resource.

How Dynamic Search Ads currently build ads and what you control

Dynamic Search Ads use your website content to target and build ads. Per Google Ads Help, headlines and landing pages are generated from the site's content, with the HTML page title named as a leading signal among those used to generate the headline. The advertiser writes the description lines, and remains responsible for the final ad text that users see.

When a user's query matches your DSA targeting, Google selects a relevant landing page from your site and generates a headline from that page's content, especially its title tag. The final URL is that landing page, and the display URL domain is derived from that final URL. You do not write the headline directly, and Google notes that it cannot guarantee which headline variation will serve, since it may modify generated text for policy compliance.

Your direct controls are the descriptions, the targeting configuration, and the exclusions. You choose which parts of the site are eligible, which pages are excluded, and which queries are blocked with negative keywords. You also control bids and budgets at the campaign or ad group level, but those are standard search controls rather than DSA-specific behavior.

A common misconception is that you write DSA headlines. In reality, you influence them through your HTML title tags and on-page content, while writing only the descriptions. This is why title-tag hygiene is a core DSA optimization lever, since the title is a leading signal for the generated headline.

Targeting options and keeping DSA on-topic

Google's documentation states that DSA offers several targeting options that all use your site content. You can target by URL rules, by Google-generated categories, or by page feeds that you upload. Each option controls which URLs are eligible to serve as landing pages and headline sources.

URL-based targeting includes URL_Equals and URL_Contains rules. URL_Equals lets you specify exact URLs that can trigger dynamic ads. URL_Contains lets you match URL substrings, for example a directory path, so that any URL containing that string is eligible. These rules are how you include or exclude site sections at a structural level.

Categories are themed sets of landing pages that Google generates from your site. You can select categories that align with your offerings and bid differently by category. Since categories are inferred, they depend on how well your site is structured and how clearly content themes are expressed in URLs and titles.

Page feeds are uploaded lists of URLs with labels. Per Google Ads Help, you can have up to 100 page feeds per account. In DSA, you can target specific labels or all URLs in a feed, which gives you precise control over which pages are eligible, independent of site navigation. This is often the cleanest way to keep DSA tightly scoped.

To keep DSA on-topic, Google recommends regular search terms review plus negative keywords and DSA exclusions to prevent unwanted pages from serving. You can create DSA exclusions to block specific URLs or URL patterns from being used as landing pages. Negative keywords prevent your dynamic ads from showing on irrelevant queries, even if a page technically matches.

Another misconception is that DSA is uncontrollable or that exclusions are weak. In practice, you can keep DSA tightly focused by combining URL rules or page feeds with a disciplined search terms review and negative keyword process, which is covered more broadly in this negative keyword reference.

Where DSA still fits and where it never did

Google Ads Help describes DSA as best suited to well-developed websites or large inventories. The system performs well when there is enough crawlable, text-based content for Google to understand what each page is about and to generate relevant headlines and landing pages.

DSA is not recommended for rapidly changing sites, such as daily deals, because the content that Google indexes can lag behind what is live. It is also a poor fit for sites that are mostly image-based or that gate content behind logins, since the crawler and ad systems rely on accessible text content.

Title-tag hygiene is a fit criterion as well. Google suggests keeping titles in the 60 to 90 character range. Clean, descriptive titles within that range help DSA generate clearer headlines and match queries more accurately. Sites with missing, duplicated, or vague titles tend to produce weaker DSA performance because the leading headline signal is low quality.

Some advertisers who struggled with DSA were never a good fit, for example those whose entire offer changed daily or whose key content was not indexable. Others treated DSA as a fully automatic solution and skipped the search terms and exclusions work that Google recommends, which led to off-topic queries and landing pages.

Migration timeline from DSA to AI Max

On April 15, 2026, Google announced that AI Max was out of beta and positioned it as the next generation of Dynamic Search Ads. In that announcement, Google stated that DSA, automatically created assets, and campaign-level broad match campaigns would begin auto-upgrading starting in September 2026.

On June 11, 2026, Google revised that timeline. Per the updated blog post and the DSA help page header, the DSA sunset and auto-upgrade were deferred to begin in February 2027, while the September 2026 wave proceeded for automatically created assets and campaign-level broad match. Any content that still cites September 2026 as the DSA auto-upgrade date is out of date relative to this revision.

When the auto-upgrades begin in February 2027, new DSA creation will end across Google Ads, Google Ads Editor, and the Google Ads API. Google states that in-account notifications will appear beforehand, so advertisers will see prompts before their campaigns are upgraded.

A common misconception is that DSA is already dead or that it auto-upgrades in September 2026. In fact, DSA still runs and serves today, and the sunset and auto-upgrade begin in February 2027, after being deferred from the original September 2026 plan.

Google's documentation also notes that there is no permanent opt-out from the February 2027 auto-upgrade. The choice for advertisers is when and how to migrate, not whether to migrate.

What upgrading does to DSA campaigns and components

When you upgrade a DSA campaign to AI Max, dynamic ad groups transition to standard ad groups. Google states that ad group settings and historical data are ported into the new structure, so your configuration and performance history are preserved at the ad group level.

All three AI Max features are enabled on the upgraded campaign, with legacy URL controls preserved. Existing Dynamic Search Ads are converted into responsive search ads. Google uses text customization to generate the minimum static assets required for RSAs, based on your existing DSA descriptions and site content.

DSA targets that are unsupported in AI Max become read-only. The same applies to legacy URL rules such as page title contains and page content contains. These read-only elements can be removed but not edited after the upgrade. This means your historical targeting logic is visible but cannot be modified in its old form.

Google describes a transition safety behavior. Legacy DSAs keep running temporarily while the new RSAs are generated and pass policy review. Once the new RSAs are active, the DSAs are automatically paused into a view-only state. This counters the misconception that the upgrade deletes your DSA setup, since settings and data are ported, URL controls are preserved in some form, and the old DSAs remain visible in a paused, view-only status.

For the RSA mechanics that your DSAs become, see this responsive search ads reference for details on how those ads work once generated.

Experiment paths to test AI Max against DSA

Google Ads Help describes two experiment paths to test AI Max against your current DSA setup. The first is to create a custom experiment by using the Upgrade campaign option from a draft that copies your DSA campaign settings. This lets you run a controlled split between the original DSA campaign and an AI Max variant.

The second path is to use the experiment template. In the Experiments section, you can choose Campaign features and settings, then AI Max, which applies a standard upgrade configuration in the treatment arm. This is designed to simplify testing of AI Max features against your existing campaign.

There are structural caveats. Campaigns that are 100% dynamic ad groups cannot use the templated experiments. For mixed campaigns that contain both DSA and non-DSA ad groups, the template only upgrades the non-DSA ad groups in the treatment arm. This means that for pure DSA campaigns, you rely on the custom experiment path rather than the template.

Google cites internal data from 2026, for non-retail advertisers, that AI Max campaigns using the full feature suite see an average 7% more conversions or conversion value at similar CPA or ROAS compared with using search term matching alone. That figure is directional context for experiments, not a guarantee for any individual account.

What DSA advertisers should do before February 2027

TF's migration posture, the same one applied to the Salesforce legacy sunset, is to move before deadlines force it, on your own test design rather than Google's schedule. For a DSA advertiser that means running the experiment path now against AI Max, reading it on ROAS over a proper window, and migrating the winners deliberately instead of letting February 2027 do it wholesale.

In practical terms, that means selecting your material DSA campaigns, cloning them into experiments, and using the Upgrade campaign or AI Max experiment template where eligible. You then compare performance on conversions or conversion value and CPA or ROAS, using your own thresholds for success, before applying the upgrade to the base campaigns.

Before or alongside experiments, clean up your site inputs. DSA's quiet lesson carries over, that generated targeting is only as good as the site it reads, so title-tag and landing-page hygiene, and the search-terms-plus-negatives discipline, matter as much under AI Max's keywordless matching as they did under DSA. This includes aligning titles with the 60 to 90 character range Google suggests and ensuring that key pages have clear, descriptive titles and crawlable content.

Finally, plan for the structural changes. Expect your dynamic ad groups to become standard ad groups, your DSAs to become RSAs, and some legacy URL rules to become read-only. Use the time before February 2027 to document which URL rules and page feeds are critical, to refine exclusions, and to decide where AI Max's broader feature set is acceptable versus where you may want tighter controls through other campaign types. For a broader account-level review cadence that supports this, see the Google Ads audit workflow.