Dynamic Search Ads Best Practice and AI Max

Last updated: 19 August 2026

Executive Summary

Dynamic Search Ads (DSAs) remain a genuinely useful Search advertising tactic in 2026 — but they now come with a firm expiry date. Google announced on 15 April 2026 that DSAs would be retired and migrated to AI Max for Search, and while a June 2026 reprieve extended the timeline, the forced automatic migration of remaining DSA campaigns is currently scheduled to begin around February 2027.^[8] New DSA campaign creation is expected to be blocked around January 2027.^[8]

The verdict for a senior practitioner is nuanced: DSAs are still worth running in 2026 for the right sites, used selectively, with a clear plan for the AI Max transition. They remain the most direct tool for large-catalogue or frequently changing sites where keyword-level coverage is impractical, and for capturing converting long-tail queries that standard keyword campaigns miss.^[5][12] That case has not weakened simply because Google has announced a migration path. What has changed is the calculus around investment: every DSA build or expansion decision made today should be weighed against the February 2027 migration window and the operational cost of re-doing that work inside AI Max.

The following principles govern sound DSA use through to migration:

• Keep tight control from day one. Use a custom page feed with custom labels as your primary targeting mechanism, not whole-site auto-targeting. Exclude non-commercial pages and seed an aggressive negative keyword list before launch.^[2][5]
• Use Smart Bidding. Maximise Conversions (with an optional target CPA) or Maximise Conversion Value (with an optional target ROAS) are the correct bid strategies when conversion tracking is in place. Manual CPC is not recommended for mature DSA programs.^[8]
• Start small, scale deliberately. Begin DSAs at roughly 10–20% of Search spend during the discovery phase, then reallocate toward labels and targets that meet CPA or ROAS thresholds, scaling in 10–20% increments.^[8]
• Run AI Max side by side now. The February 2027 migration deadline is close enough that running an AI Max for Search experiment now is prudent, not premature. Voluntary migration on your terms beats an automatic upgrade on Google’s default settings.^[8]
• Separate the DSA timeline from the September 2026 wave. The September 2026 auto-upgrade applies to campaigns using automatically created assets (ACAs) and campaign-level broad match — not DSAs. Conflating the two timelines is the most common planning error right now.^[8]

For sites that are small, static, or already fully covered by tightly managed keyword campaigns, the cost-benefit of building new DSA infrastructure in late 2026 is questionable. For large e-commerce, travel, real estate, recruitment, and other high-SKU or high-churn environments, DSAs still earn their place — provided the setup is disciplined and the team is actively preparing for migration.

How Dynamic Search Ads Work

DSAs are a distinct Search ad format in which Google’s systems, rather than the advertiser, determine two of the three core ad components: the headline and the landing page. The advertiser retains full control over the description lines — up to four descriptions of 90 characters each, consistent with Google’s current Search ad specifications.^[5][11]

The targeting mechanism is fundamentally different from keyword-based Search campaigns. Instead of a keyword list triggering auction entry, DSAs use Google’s index of your website content — or a structured page feed you supply — to determine when a user’s search query is relevant to a specific page on your site.^[5][12] When a match is found, Google dynamically constructs a headline derived from the most relevant content on that page and routes the user directly to the matched URL.^[5]

The Matching Process Step by Step

• Google crawls your site (or reads your page feed) and builds an understanding of the topic, intent, and content of each indexable page.^[5][12]
• When a user enters a search query, Google evaluates whether the query’s intent is likely served by one of your crawled or listed pages.^[5]
• If a match meets the relevance threshold, Google enters the auction, dynamically generates a headline from the page content, and serves the ad with your written descriptions.^[5][12]
• The user lands on the matched page, not a single designated landing page — meaning landing-page relevance is a structural output of the DSA system, not something you set manually per ad.^[5]

What DSAs Need to Work Well

Because headline generation and page selection are automated, the quality of the output is entirely dependent on the quality of the site content Google can read. DSAs perform reliably only when the underlying site meets these conditions:^[5][12]

• Well-structured HTML page titles that clearly and specifically describe the topic of each page. The page title is the primary source for dynamically generated headlines.
• Clear, crawlable on-page content that is not hidden behind JavaScript rendering, login walls, or session-dependent parameters.
• A coherent site architecture with consistent URL patterns, relevant category structures, and minimal orphan or thin pages.
• Commercial intent on each targeted page — product pages, category pages, service pages, and location pages perform best; blog posts, help articles, and careers pages should be excluded.

Ad Format and Advertiser Control

Within the DSA format, advertiser control is concentrated in the description lines. These should be written to reinforce the commercial intent of the page category being targeted, since the headline will vary by query and page. Generic descriptions that could apply to any page on the site waste the format’s capacity to pre-qualify clicks. Strong descriptions clarify the value proposition, the conversion path, and any relevant offer — with the understanding that they must remain relevant across the range of pages a given ad group might serve.^[5][11]

Core Use Cases

• Large catalogues or inventory-heavy sites where building individual keywords and ads for every product or service page is operationally impractical or impossible.^[12]
• Frequently changing inventory or promotions, where a page feed keeps DSA targeting current without manual keyword updates each cycle.^[2][5]
• Filling coverage gaps in keyword campaigns by surfacing relevant queries that no existing keyword would trigger.^[12]
• Query discovery, where the search terms report reveals high-converting queries and landing pages that can be promoted into standard keyword campaigns.^[12]

Targeting Options and Page Feeds

DSA targeting operates along a spectrum from fully automated (Google indexes and targets your entire site) to tightly controlled (you specify every eligible URL through a page feed). In practice, the tighter end of the spectrum is almost always preferable for efficiency and brand safety. The table below summarises the four main targeting modes.^[2][3][5]

Targeting Mode	How It Works	Control Level	Best For
All webpages	Google indexes and targets every crawlable page on the domain	Low	Small, clean sites where every page is commercially relevant
Google-recommended categories	Google groups pages into topic categories derived from site content; you select which categories to activate	Medium-low	Mid-sized sites with reasonably consistent page structure
Specific webpages by URL rules	You define rules based on URL pattern, page title text, page content, or category name	Medium-high	Sites with predictable URL structures (e.g. /products/, /services/)
Custom page feed	You supply a spreadsheet of exact URLs with optional custom labels; DSAs serve only from that list	High	Any account with meaningful spend, mixed-content sites, or commercial-priority requirements

Page Feeds: The Recommended Control Mechanism

A custom page feed is a structured spreadsheet uploaded to Google Ads (via the Business Data section) that lists the exact URLs eligible to serve as DSA landing pages.^[1][2] You can attach custom labels to each URL, which allows you to segment URLs into logical commercial groups — for example, by product category, margin tier, geographic focus, or promotional status — and then build separate DSA ad groups targeting each label.^[1][4]

Google also supports a mode that restricts DSA serving exclusively to URLs in the page feed, preventing the system from crawling and serving other pages on the domain even if they are indexable.^[5] This is the most conservative and recommended configuration for accounts where landing-page governance matters — regulated industries, large e-commerce sites with mixed-quality pages, or any site with a significant volume of non-commercial content.

Custom Labels: Structuring for Performance

Custom labels transform a page feed from a simple URL list into a bidding and optimisation framework. Practical label structures include:^[4][1]

• By commercial tier — top sellers, high-margin products, clearance items
• By category — mapped to site navigation or product taxonomy
• By geography — location-specific service pages or store pages
• By conversion value — pages associated with high average order value versus low
• By promotion status — current sale pages, used to push spend during campaigns and then deprioritise when promotions end

Each label should map to a single DSA ad group so that descriptions, bids, and exclusions can be managed at the appropriate level of granularity. If a label starts returning a mix of high- and low-intent traffic, split it into smaller labels before scaling spend on it.^[3][4]

URL Exclusions

URL exclusions at the campaign or ad group level block specific pages or URL patterns from serving as DSA landing pages, regardless of what the page feed or auto-targeting would otherwise include.^[6] Standard pages to exclude include:^[6][2]

• Contact, about, and team pages
• Privacy policy, terms and conditions, and legal pages
• Careers and job listing pages
• Help centre and support articles
• Login, account, cart, and checkout pages
• Internal search results pages
• Thin or duplicate pages
• Out-of-stock or unavailable product pages (Google explicitly recommends excluding pages containing terms like “sold out” or “unavailable”)^[2]

Ad Group Structure Best Practice

The recommended structure is one DSA ad group per page feed label, with each ad group containing descriptions written specifically for that commercial segment.^[4] This means bidding, creative, and exclusion decisions can be applied at the right level without creating conflicts between segments. A flat, single-ad-group DSA campaign covering the whole site is a common and costly structural error that makes optimisation extremely difficult.

Setup and Bidding Best Practice

Pre-Launch Checklist

Before a DSA campaign goes live, confirm the following are in place:^[8][2][5]

• Conversion tracking is active and verified — Smart Bidding requires accurate conversion signals to optimise effectively. Launching DSAs without conversion tracking means bidding blind.
• A custom page feed is built with custom labels assigned to commercial page groups.
• URL exclusions are configured for all non-commercial page types listed in the section above.
• An initial negative keyword list is seeded — at minimum, branded terms already covered by dedicated campaigns, competitor terms, informational intent qualifiers (e.g. “how to”, “what is”, “guide”), career-related terms, and support-related terms.
• Site crawlability is confirmed — check that pages in the feed are not blocked by robots.txt or noindex tags, and that page titles are descriptive and specific.
• Campaign structure matches keyword campaigns — DSAs should run in a separate campaign to prevent cannibalisation and to allow clean performance comparison.

Bid Strategy Selection

Smart Bidding is strongly preferred over manual CPC for any DSA program with sufficient conversion data.^[8] The correct strategy depends on business objective:

Business Objective	Recommended Bid Strategy	When to Add a Target
Maximise conversion volume within budget	Maximise Conversions	Add target CPA once the campaign has at least 30–50 conversions and CPA is stable
Maximise conversion value within budget	Maximise Conversion Value	Add target ROAS once value data is reliable and ROAS is consistent over 30+ days
New campaign, no conversion data yet	Maximise Conversions (no target) initially, then transition	Transition to target CPA after the learning period

Manual CPC may be appropriate in very early testing with no conversion history, but it should be treated as a temporary state, not a long-term strategy. Google’s current guidance is explicit that bid strategy choice should reflect the business objective and that Smart Bidding requires accurate conversion signals to function correctly.^[8]

Budget Allocation and Scaling

DSAs should begin as a discovery and coverage layer, not as a primary spend vehicle. The recommended phasing approach — which reflects operational best practice rather than a published Google rule — is:^[8]

• Launch phase: Allocate approximately 10–20% of total Search spend to DSAs. This gives the Smart Bidding algorithm sufficient signal without risking meaningful budget on unvalidated targeting.
• Optimisation phase: Review the search terms report daily for the first 30 days. Add negatives aggressively. Identify which page feed labels are meeting CPA or ROAS targets.
• Scaling phase: Increase budget allocation to proven labels in increments of 10–20%, only after performance has stabilised over at least two to four weeks. Do not scale all labels simultaneously.
• Reallocation: Reduce or pause labels that consistently fail to meet efficiency targets. The DSA structure should self-select toward commercially productive page groups over time.

Ad Copy: Writing Descriptions That Work

Because Google generates the headline automatically from page content, your four description lines (up to 90 characters each) carry the full burden of communicating value proposition, offer, and conversion intent.^[5][11] Effective DSA descriptions:

• Are written for the category being targeted, not for a single page — they must remain relevant across the range of pages in the ad group’s page feed label.
• Include a clear call to action and a reason to click rather than bounce.
• Avoid generic phrases (“Learn more”, “Find out today”) that add no qualifying information.
• Reference competitive differentiators — price, range, shipping, service — that are true across the page group.
• Are reviewed whenever the page feed label is updated, to ensure continued relevance.

Preventing Cannibalisation of Keyword Campaigns

In Google Ads auction logic, an exact-match keyword in a standard Search campaign will take precedence over a DSA opportunity for the same query, provided the keyword is eligible and the ads are in different campaigns or ad groups.^[9] However, this protection is not absolute, and DSAs can still consume budget on queries adequately covered by keyword campaigns if the campaign structure is not carefully managed.^[9]

To prevent cannibalisation:^[9][2]

• Add your core keyword list as negative keywords in the DSA campaign so those queries route to keyword campaigns rather than DSAs.
• Segment DSA campaigns by commercial intent and restrict them to page groups not yet covered by keyword campaigns.
• Monitor the DSA search terms report for queries that overlap with keyword campaign traffic and resolve conflicts by adding negatives at the query or URL level.
• Do not rely on campaign priority settings alone to resolve conflicts — use explicit negative keywords.

Controlling DSAs: Negatives, Exclusions and Search Terms

Control is the defining discipline of a well-run DSA program. Without structured negative management, DSAs will drift toward low-commercial-intent traffic and inflate cost per acquisition quickly. The DSA system provides two distinct exclusion mechanisms — negative keywords (blocking queries) and negative URL targets (blocking pages) — and both need to be maintained actively.^[6][19]

Negative Keywords

Negative keywords in DSA campaigns function the same way as in keyword campaigns: they prevent the ad from serving when the user’s query contains the excluded term.^[6] DSA campaigns typically require a broader and more aggressive negative keyword list than keyword campaigns, because the matching system is inherently less precise. The foundational categories to negate include:

• Informational intent: “how to”, “what is”, “guide”, “tutorial”, “explained”, “definition”, “review”
• Support and service: “help”, “support”, “contact”, “complaint”, “refund”, “return”, “cancel”
• Career and internal navigation: “jobs”, “careers”, “login”, “sign in”, “my account”, “sitemap”
• Research and comparison: “vs”, “versus”, “compare”, “best”, “top 10” (where these are not converting intent for your category)
• Brand terms already in dedicated campaigns — exclude your own brand terms to prevent budget split between DSA and brand campaigns
• Competitor terms — unless you have a deliberate competitor-targeting strategy, exclude these to avoid irrelevant traffic

Note that negative keywords in DSA campaigns should not conflict with keywords in the same campaign structure. Google’s guidance cautions that if a negative keyword matches a regular keyword in the same ad group, that keyword may be prevented from serving.^[2]

Negative Dynamic Ad Targets (URL Exclusions)

Negative dynamic ad targets block specific pages or URL patterns from serving as DSA landing pages, operating at either the campaign or ad group level.^[6][19] These are distinct from negative keywords and address a different failure mode: a page that ranks for a query but should not serve as a DSA destination.

URL exclusions should be applied using the same logic as page feed URL selection — if a page is not in your commercial page feed, it should also be explicitly excluded as a negative dynamic ad target, particularly if whole-site or category targeting is active in any ad group.^[6]

The Search Terms Report: Cadence and Actions

Google provides a dedicated DSA search terms report that shows both the query that triggered the ad and the landing page that was served.^[16] This dual visibility — query plus URL — is one of DSA’s most valuable diagnostic tools and enables two types of exclusion action directly from the report:^[16]

• Add as negative keyword — blocks the query from triggering DSAs in future
• Exclude landing page — adds the served URL as a negative dynamic ad target

Recommended review cadence:^[16][6]

• Daily for the first 30 days — DSAs can accumulate spend on irrelevant queries rapidly in the early phase. Daily review allows fast iteration and prevents budget waste from compounding.
• Weekly thereafter — once the negative list is mature and performance has stabilised, weekly review is sufficient provided volume is meaningful.
• After any page feed update — adding new URLs to the feed can open new query patterns; review the search terms report within 48–72 hours of a feed update.

The search terms report should also be used proactively for query mining: high-converting queries surfaced by DSAs that are not covered by existing keyword campaigns are strong candidates for addition to exact or phrase-match keyword campaigns, where they can be managed with tighter bid control.^[12]

Structural Controls Summary

The table below maps the control mechanism to the problem it solves.

Control Mechanism	Problem Addressed	Applied At
Negative keywords	Irrelevant or low-intent search queries triggering DSAs	Campaign or ad group level
Negative dynamic ad targets (URL exclusions)	Non-commercial or unsuitable pages serving as landing pages	Campaign or ad group level
Page feed (URL-restricted mode)	DSAs crawling and serving pages outside approved commercial inventory	Campaign level
Custom labels + separate ad groups	Inability to bid, optimise, or exclude at page-category level	Ad group level
Search terms report review	Detecting new irrelevant queries and underperforming landing pages	Ongoing operational process

The DSA Retirement and AI Max Migration Timeline

The migration timeline has changed twice since April 2026. Understanding the current state — and distinguishing between the two separate auto-upgrade tracks — is essential for planning.

Timeline of Events

Date	Event
15 April 2026	Google announces DSA retirement; all eligible DSA campaigns, automatically created assets, and campaign-level broad match to be auto-upgraded to AI Max beginning September 2026[8]
11 June 2026	Google delays the DSA-specific migration and restores the ability to create new DSA campaigns[8]
15 June 2026	New DSA campaign creation re-enabled in Google Ads UI, Ads Editor, and API[8]
~1 September 2026	Auto-upgrade begins for campaigns using automatically created assets and campaign-level broad match — DSAs are excluded from this wave[8]
~January 2027	New DSA campaign creation expected to be blocked in Google Ads UI, Ads Editor, and API[8]
~February 2027	Automatic migration of remaining DSA campaigns to AI Max for Search begins[8]

Two Separate Upgrade Tracks — Do Not Conflate Them

The most operationally significant distinction in the current timeline is that there are two separate auto-upgrade tracks running on different schedules:^[8]

• September 2026 track: Campaigns using automatically created assets (text customisation) and the campaign-level broad match setting. These are being auto-upgraded to AI Max from approximately 1 September 2026, regardless of whether those campaigns also contain DSAs. If you have Search campaigns using ACAs or campaign-level broad match, that upgrade is imminent and should be your immediate planning priority.
• February 2027 track: Dynamic Search Ads specifically. DSA campaigns that have not been voluntarily migrated will be automatically converted to AI Max for Search beginning around February 2027. New DSA creation is expected to be blocked approximately one month earlier, around January 2027.

Advertisers who conflate these two tracks risk either acting too urgently on DSAs (which have more runway) or, more dangerously, being caught off-guard by the September 2026 ACA and broad match upgrades.

What Happens During Automatic Migration

Based on Google’s published migration materials and practitioner reporting, the automatic DSA migration to AI Max is expected to preserve existing URL controls and convert dynamic ad groups into standard ad groups.^[8] Page feeds are explicitly positioned by Google as the primary bridge between legacy DSA URL targeting and AI Max’s automated destination selection — existing page feeds are expected to carry forward.^[8]

However, the scope of what is preserved versus reset to AI Max defaults during automatic migration has not been fully specified by Google at time of writing. This uncertainty is the primary argument for voluntary migration on your own timeline rather than waiting for the forced upgrade. Key items to verify before relying on automatic migration include:^[8]

• URL inclusion rules and page feed label assignments
• Final URL expansion behaviour and whether it respects existing exclusions
• Asset behaviour, including which descriptions carry forward and how text customisation interacts with existing copy
• Landing-page compliance for regulated categories
• Negative keyword lists — confirm these transfer intact

Recommended Pre-Migration Actions

• Audit all campaigns now for ACA use, campaign-level broad match, and DSA status — treat these as three separate inventories with different urgency levels.^[8]
• Prioritise the September 2026 ACA and broad match audit — these are the nearer-term risk.
• Run an AI Max for Search experiment alongside current DSAs to gather performance data before the forced migration window.^[8]
• Document your current DSA page feeds, labels, URL exclusions, and negative keyword lists — this documentation is essential for validating what carries forward in migration.
• Monitor account notifications from Google — Google has indicated that pre-migration banners will appear in the UI ahead of the DSA transition window.^[8]

DSA vs AI Max for Search vs Performance Max

With Google actively migrating DSAs to AI Max and Performance Max already established as the fully automated campaign type, practitioners need a clear comparison of what each format offers and where each sits in a 2026 Search strategy.

Dimension	Dynamic Search Ads	AI Max for Search	Performance Max
Ad format	Search only	Search only	All Google channels (Search, Display, YouTube, Shopping, Maps, Discover)
Headline generation	Dynamically generated from page content	AI-generated from assets and page content; text customisation enabled	AI-generated from asset groups; broader creative variation
Landing page selection	Matched to site page by crawl or page feed	Final URL expansion with page feed support; broader automation of destination	Final URL expansion across the site; less granular control
Keyword requirement	None — site content or page feed drives matching	Optional — search term matching is the primary mechanism, keywords can inform	None — audience signals and asset content drive targeting
Advertiser control over query matching	Medium — negatives and URL exclusions; no direct keyword targeting	Medium — search term matching with negative keywords and URL controls	Low — negative keywords at account level; limited query-level visibility
Page feed support	Yes — primary recommended control mechanism	Yes — carries forward from DSA and supports URL governance in AI Max	Yes — supported but landing page control is less precise than DSA
Search terms report visibility	High — full query and landing page visibility	Medium — search term matching provides visibility; some detail may be aggregated	Low — partial; “Search categories” rather than individual queries in many cases
Retirement status	Being retired; new creation ends ~January 2027, auto-migration ~February 2027	Active and expanding; the destination for DSA migrations	Active and evolving; Google’s primary fully-automated campaign type
Best suited to	Large catalogues, frequently changing inventory, query discovery on Search	Search-focused advertisers wanting automation with more transparency than PMax	Advertisers wanting full-funnel coverage across all Google inventory with minimal management

The Practical Difference Between DSA and AI Max for Search

AI Max for Search is best understood as DSA expanded and rebuilt on a broader automation foundation.^[8] Where DSA’s automation was primarily limited to headline generation and landing-page matching based on site content, AI Max adds search term matching (broader query matching informed by AI), text customisation across the full ad, and deeper integration with Google’s campaign-level AI signals. In theory, AI Max should match or exceed DSA performance for most use cases — but that comparison needs to be validated with real data from your account before the forced migration, not assumed.^[8]

Where Performance Max Sits

Performance Max operates at a fundamentally different level of abstraction. It trades granular control over queries and landing pages for breadth of coverage across all Google inventory. For advertisers whose primary objective is Search coverage and landing-page relevance, Performance Max is not a direct substitute for DSA — it is a different product for a different objective. AI Max for Search is the closer functional replacement.^[8]

Who Should Still Use DSAs (and Who Should Not)

Strong Candidates for DSA in 2026

• Large e-commerce catalogues — retailers with thousands of product pages, where keyword-level coverage of every SKU is impractical. DSAs provide automatic coverage and surface high-converting long-tail queries that would otherwise be missed.^[12]
• Frequently changing inventory — travel sites, real estate platforms, recruitment sites, event ticketing, and similar verticals where page content changes regularly. Page feeds allow targeting to track inventory without manual keyword updates.^[2][5]
• Advertisers with well-structured, crawlable sites — DSA quality is proportional to site quality. Sites with clear HTML titles, descriptive page content, and logical URL structures will get better headline generation and more relevant query matching.^[5][12]
• Accounts that want query discovery — DSAs remain one of the most effective ways to identify converting queries for promotion into keyword campaigns. This use case is valid even for accounts where DSA is not the primary spend vehicle.^[12]
• Accounts already running DSA campaigns that are performing — if a DSA campaign is meeting CPA or ROAS targets today, there is no reason to pause it ahead of the February 2027 migration. Run it alongside an AI Max experiment and migrate voluntarily when the data supports it.

Poor Candidates for DSA in 2026

• Small sites with fully mapped keyword coverage — if every commercially relevant page is already covered by tight keyword campaigns, DSAs add noise rather than discovery value.
• Sites with poor crawlability or thin page content — badly structured pages produce low-quality dynamic headlines and irrelevant query matching. The DSA system will not compensate for poor site architecture.
• Sites with significant non-commercial content mixed in — if extensive exclusion work is required to make the site safe for DSAs, the operational overhead may outweigh the benefit, particularly this close to the retirement date.
• Advertisers building new DSA campaigns in late 2026 with no existing DSA infrastructure — with new DSA creation ending around January 2027 and the AI Max migration beginning around February 2027, the runway for a new DSA build to reach performance maturity before forced migration is very short. Starting with AI Max for Search directly may be more efficient.
• Regulated industries with strict landing-page compliance requirements — DSA’s automated landing-page selection requires careful page-feed and exclusion management. If automated landing-page selection creates compliance risk and the operational controls cannot be resourced adequately, DSA is not appropriate regardless of potential performance.

Common Mistakes to Avoid

• Using whole-site auto-targeting without exclusions. Targeting all webpages without URL exclusions is the most common and costly DSA error. Non-commercial pages will accumulate spend rapidly, inflating CPA and reducing overall campaign efficiency. Always configure URL exclusions before launch.^[6][2]
• Neglecting the search terms report in the first 30 days. DSA campaigns can expand into unexpected query patterns very quickly. Failing to review the search terms report daily during the first month allows irrelevant spend to compound. Set a calendar reminder and treat the report review as a non-negotiable daily task during launch.^[16]
• Writing generic description lines. Because the headline is automated, the descriptions are doing more work than in a standard Search ad. Generic descriptions that could apply to any page on the site waste the format’s ability to pre-qualify clicks and reduce ad relevance.^[5][11]
• Running DSA and keyword campaigns without negative keyword separation. Without explicit negatives in the DSA campaign blocking queries covered by keyword campaigns, DSA may compete for — and win — auctions that should route to keyword ads. Always seed DSA campaigns with your core keyword list as negatives.^[9][2]
• Scaling budget before performance has stabilised. DSAs on Smart Bidding require a learning period, during which CPA and ROAS may appear erratic. Scaling budget aggressively before at least 30 days of stable data (and at least 30–50 conversions) risks permanently disrupting the Smart Bidding learning state.^[8]
• Using a single ad group for all page feed labels. Collapsing all commercial pages into one ad group prevents label-level bidding, makes description relevance impossible to manage, and makes it difficult to diagnose which page groups are driving performance. Build separate ad groups per label from the start.^[4]
• Treating the September 2026 and February 2027 migration dates as the same event. The September 2026 auto-upgrade applies to ACAs and campaign-level broad match — not DSAs. Confusing the two timelines leads either to unnecessary urgency around DSA migration or, more dangerously, to missing the ACA and broad match changes that are imminent.^[8]
• Assuming automatic migration will preserve all settings correctly. Google’s migration materials indicate page feeds and URL controls should carry forward, but the full scope of what is and is not preserved in automatic migration has not been fully specified. Relying on automatic migration without prior testing of AI Max is a risk that can be avoided by running a side-by-side experiment now.^[8]
• Abandoning a performing DSA setup because of the retirement announcement. If a DSA campaign is meeting targets, the retirement announcement is not a reason to pause it immediately. It is a reason to run an AI Max experiment alongside it, document the setup thoroughly, and plan a voluntary migration ahead of the February 2027 deadline.^[8]

What Changed Recently (Last 30 Days)

The most significant development in the last 30 days is the confirmation and formalisation of the two-track migration timeline and the practitioner guidance around the September 2026 auto-upgrade for campaigns using automatically created assets and campaign-level broad match.^[8]

Key Changes and Developments

• September 2026 ACA and broad match auto-upgrade confirmed as imminent. Google’s developer blog has formalised that the September 1, 2026 auto-upgrade applies to campaigns with automatically created assets (text customisation) and those using the campaign-level broad match setting. The rollout is described as progressive rather than instantaneous, meaning not all eligible campaigns will migrate simultaneously. Advertisers with these campaign types should have reviewed account settings and banners by now; if not, this is the immediate priority.^[8]
• DSA auto-migration remains confirmed for ~February 2027. The delay first announced in June 2026 has not been further extended as of the research available. The February 2027 date should be treated as the current authoritative deadline, with the caveat that Google has revised this timeline once already and may do so again — defer to account-level notifications and Google’s official channels for any future changes.^[8]
• Page feeds confirmed as the primary URL governance bridge into AI Max. Google’s migration materials and practitioner coverage in the last 30 days consistently position page feeds as the continuity mechanism between DSA URL targeting and AI Max’s automated destination selection. Advertisers who have not yet built page feeds for existing DSA campaigns — whether running on whole-site or category targeting — should prioritise this immediately, as it is the control mechanism that carries the most weight in the AI Max environment.^[8]
• Practitioner guidance is converging on voluntary migration as the safer path. Reporting from the practitioner community over the last 30 days consistently advises that voluntary migration — where the advertiser controls the timing, validates settings, and compares performance before and after — produces better outcomes than waiting for Google’s automatic migration defaults to apply. The practical recommendation is to run an AI Max for Search experiment now alongside existing DSAs, review final URL expansion behaviour, text customisation outputs, and asset carryover before committing to migration.^[8]
• New DSA creation remains open but the window is narrowing. As of mid-June 2026, new DSA campaigns can be created in the Google Ads UI, Ads Editor, and API. That window is expected to close around January 2027. Advertisers considering new DSA builds should factor the short operational runway into their decision — new campaigns will have fewer than six months before creation is blocked and fewer than seven months before automatic migration begins.^[8]

The Practical Implication for Right Now

If your account has campaigns using automatically created assets or campaign-level broad match, the September 2026 auto-upgrade is your immediate concern. Audit those campaigns, review what AI Max will change, and decide whether to opt in voluntarily or let the automatic upgrade proceed — but make that an informed, active decision, not a default by inaction.

If your account has DSA campaigns, the February 2027 migration gives you time to run a proper AI Max experiment. Use that time. A side-by-side test started in September or October 2026 gives you three to four months of data before the forced migration window — enough to validate performance, tune the AI Max setup, and migrate on your own terms rather than Google’s defaults.

DSAs are a still-working tactic with a defined shelf life. The discipline is to keep running them tightly while actively building what comes next.

---

References

1. [1] https://developers.google.com/google-ads/api/docs/dynamic-search-ads/dsa-page-feeds developers.google.com
2. [2] https://support.google.com/google-ads/answer/7166527?hl=en support.google.com
3. [3] https://support.google.com/google-ads/editor/answer/9100019 support.google.com
4. [4] https://searchengineland.com/adwords-dynamic-search-ads-page-feeds-expanded-text-ads-27208… searchengineland.com
5. [5] https://support.google.com/google-ads/answer/2471185?hl=en support.google.com
6. [6] https://clickpatrol.com/google-ads-character-limits-2026-guide-headlines/ clickpatrol.com
7. [7] https://yeezypay.io/blog/how-to-set-up-dynamic-search-ads-in-2026 yeezypay.io
8. [8] https://blog.google/products/ads-commerce/dsa-upgrade-to-ai-max-2026/ blog.google
9. [9] https://blog.google/products/ads/dynamic-search-ads-are-now-more/ blog.google
10. [10] https://experienceleague.adobe.com/en/docs/advertising/search-social-commerce/campaign-man… experienceleague.adobe.com
11. [11] https://support.google.com/google-ads/answer/17092074?hl=en support.google.com
12. [12] https://developers.google.com/google-ads/api/docs/dynamic-search-ads/overview developers.google.com
13. [13] https://zenweb.my/blog/dynamic-search-ads/ zenweb.my
14. [14] https://adspreview.us/guides/google-ads-character-limits adspreview.us
15. [15] https://www.webtonic.io/blog/dynamic-search-ads www.webtonic.io
16. [16] https://support.google.com/google-ads/answer/9054156?hl=en support.google.com
17. [17] https://support.google.com/google-ads/answer/2497706?hl=en support.google.com
18. [18] https://support.google.com/google-ads/answer/7334472?hl=en support.google.com
19. [19] https://support.google.com/google-ads/answer/7185083?hl=en support.google.com
20. [20] https://developers.google.com/google-ads/api/docs/dynamic-search-ads/dsa-page-feeds?authus… developers.google.com
21. [21] https://support.google.com/sa360/answer/14667146?hl=en support.google.com
22. [22] https://halacreative.agency/dynamic-ad-google/ halacreative.agency
23. [23] https://developers.google.com/google-ads/api/docs/dynamic-search-ads/dsa-page-feeds?authus… developers.google.com
24. [24] https://support.google.com/google-ads/answer/2472708?hl=en support.google.com
25. [25] https://www.bluewinston.com/introduction-product-campaign-google-search/ www.bluewinston.com
26. [26] https://support.google.com/google-ads/editor/answer/65470?hl=en support.google.com
27. [27] https://experienceleague.adobe.com/en/docs/advertising/search-social-commerce/campaign-man… experienceleague.adobe.com
28. [28] https://storage.googleapis.com/think/intl/ALL_uk/docs/Google_DynamicSearch_Playbook_v8.2_D… storage.googleapis.com
29. [29] https://searchengineland.com/google-retire-dynamic-search-ads-ai-max-474262 searchengineland.com
30. [30] https://commonthreadco.com/blogs/coachs-corner/google-ads-june-2026-dsa-migration-smart-bi… commonthreadco.com
31. [31] https://www.sixthcitymarketing.com/2026/04/29/google-ads-is-retiring-dynamic-search-ads-an… www.sixthcitymarketing.com
32. [32] https://searchengineland.com/google-sets-ai-max-migration-timeline-for-search-campaigns-48… searchengineland.com
33. [33] https://www.digitalapplied.com/blog/google-delays-dynamic-search-ads-ai-max-migration-feb-… www.digitalapplied.com
34. [34] https://ppc.land/google-ads-broad-match-campaigns-face-ai-max-auto-upgrade-on-september-1/ ppc.land
35. [35] https://ads-developers.googleblog.com/2026/ ads-developers.googleblog.com
36. [36] https://farvero.com/blog/google-ai-max-migration/ farvero.com
37. [37] https://ads-developers.googleblog.com/2026/08/migrate-campaign-level-broad-match-and.html ads-developers.googleblog.com
38. [38] https://ppcnewsfeed.com/ppc-news/2026-04/dynamic-search-ads-upgrading-to-ai-max/ ppcnewsfeed.com
39. [39] https://www.linkedin.com/pulse/google-retiring-dynamic-search-ads-september-2026-heres-wha… www.linkedin.com
40. [40] https://www.trackbee.io/blog/google-ai-max-search-better-data www.trackbee.io
41. [41] https://www.linkedin.com/posts/bushletter_google-will-move-eligible-search-campaigns-activ… www.linkedin.com
42. [42] https://www.digitalapplied.com/blog/google-ads-ai-shopping-campaigns-stay-in-control-2026 www.digitalapplied.com
43. [43] https://support.google.com/google-ads/answer/7185083?hl=en-GB support.google.com
44. [44] https://business.google.com/us/resources/articles/reaching-the-right-customers-on-search/ business.google.com
45. [45] https://services.google.com/fh/files/misc/smart_biddiing_one_sheeter.pdf services.google.com
46. [46] https://support.google.com/google-ads/answer/6167140?hl=en support.google.com
47. [47] https://support.google.com/sa360/answer/9838179?hl=en support.google.com
48. [48] https://developers.google.com/google-ads/api/docs/dynamic-search-ads/create-dynamic-search… developers.google.com
49. [49] https://www.bounteous.com/insights/2013/03/11/demystifying-dynamic-search-ads-set-targetin… www.bounteous.com
50. [50] https://www.negator.io/post/dynamic-search-ads-negative-keyword-mastery-taming-googles-aut… www.negator.io
51. [51] https://www.thinkwithgoogle.com/intl/en-apac/marketing-strategies/search/dynamic-search-ad… www.thinkwithgoogle.com
52. [52] https://ppc.land/google-ads-ai-max-auto-upgrade-lands-september-1-will-ads-all-look-alike/ ppc.land
53. [53] https://www.linkedin.com/posts/hoick-media_google-ads-updates-activity-7457577573980540928… www.linkedin.com
54. [54] https://deeperinsights.com/ai-blog/google-ads-replaces-dsa-with-ai-max/ deeperinsights.com
55. [55] https://ppc.land/advertisers-push-back-as-google-sunsets-dsa-for-ai-max/ ppc.land
56. [56] https://commonthreadco.com/blogs/coachs-corner/google-ai-max-search-auto-upgrade-september… commonthreadco.com
57. [57] https://strikesocial.com/blog/dsa-to-ai-max-replacing-dynamic-search-ads/ strikesocial.com
58. [58] https://adverge.com/google-is-retiring-dynamic-search-ads-for-ai-max-in-2027-is-your-accou… adverge.com
59. [59] https://www.digitalapplied.com/blog/google-ai-max-ga-dsa-sunset-agency-playbook-2026 www.digitalapplied.com
60. [60] https://www.aubado.com/blog/dsa-to-ai-max-migration www.aubado.com