Dynamic Search Ads are the cheapest keyword research you can buy in B2B. They are also one of the worst places to spend real budget.
Dynamic Search Ads, usually shortened to DSA, are Google Search campaigns that run without keywords. Google crawls the pages you point it at, matches searches to those pages, and writes the headline itself from your page content. For B2B advertisers with long sales cycles and small qualified audiences, the right use is narrow: run DSA as a small, fenced research campaign to find the queries your keyword list is missing, move the good ones into real keyword coverage, and never let it become the campaign your pipeline depends on.
That verdict matters more this year than last, because DSA is on a clock. Google's own DSA help pages now carry a notice that starting in February 2027, campaigns using Dynamic Search Ads will be automatically upgraded to AI Max, Google's keywordless search product (Google Ads Help, About Dynamic Search Ads). So the research window is finite, and the accounts that get value out of DSA between now and then are the ones treating it as a query-mining exercise rather than a growth channel.
What Dynamic Search Ads actually do
DSA replaces the keyword with the page. Instead of telling Google which searches you want, you tell it which URLs describe your business, and Google decides which searches are close enough.
You control the pages several ways, per Google's targeting documentation:
→ URL_Equals points at one specific page, which is the tightest option and the right one for a first test.
→ URL_Contains targets every page whose URL includes a string, so `/solutions/` captures a whole section.
→ Categories are page groupings Google generates from your site, including a "Landing pages from your standard ad groups" set that simply reuses the pages your existing search campaigns already send traffic to.
→ Page feeds are an uploaded spreadsheet of URLs, which is the most controlled version of all.
Two things stay in your hands even without keywords. You can add negative keywords to a DSA campaign, and you can exclude pages. Google's own guidance for targeting recommends excluding pages containing terms like "sold out," which tells you a lot about who the product was designed for.
Ready for paid ads that pay off?
Book your free auditWhy DSA is a bad converting campaign in B2B
The design assumption behind DSA is a large catalog of pages describing things people search for by name. That is an ecommerce shape, or a hotel-chain shape, which is the exact example Google uses in its documentation. Very few B2B software or services companies have that shape.
What a B2B site has instead is a marketing site. Blog posts, a careers page, a pricing page, integration pages, a support center, a glossary someone built for SEO three years ago. Point Google at that and it will find matches, because there is always a match. The matches are researchers, students, job seekers, and people looking for a free version of something you sell for five figures a year.
The second problem is the headline. Google generates the DSA headline from the page, and the strongest signal it uses is the HTML title tag. Title tags were written for search engines and browser tabs, not to qualify a buyer. A good B2B search ad does work on the way in, repelling the wrong clicks before they cost you anything. A dynamically generated headline pulled from a page title cannot do that, because nobody wrote it with a wrong buyer in mind.
Third, DSA gets graded on the wrong scoreboard. It usually posts a low cost per click and a respectable cost per conversion, because it is winning cheap, loose queries that convert at a low bar. Then the sales team quietly stops working those leads. I have seen accounts where DSA looked like the best performing campaign in the account on platform metrics and produced nothing anyone in revenue could point to.
Run it as a research campaign instead
The version that works is deliberately small and boring.
1.) Create a separate DSA campaign, never a DSA ad group bolted onto a campaign that is working. You want its spend, its search terms, and its results isolated so you can read them and kill them independently.
2.) Cap the budget at research money. A few percent of monthly search spend is enough. The output you want is a list of queries, not volume.
3.) Target the pages that describe what you sell, using URL_Equals or a page feed. Product pages, solution pages, use-case pages. Exclude the blog, careers, support docs, the glossary, and anything free. Google's targeting guidance pushes you toward broader targets for more incremental traffic, which is good advice for a retailer and bad advice for you. You are not buying traffic here.
4.) Import your master negative keyword list on day one. Job titles, "free," "template," "salary," "vs open source," student and course terms, and every consumer-intent pattern you already know about. If you do not maintain that list as a weekly habit, build it before you turn DSA on, not after.
5.) Bid manually or on a conservative target, and keep it low. DSA on an aggressive automated target will spend fast on the loosest queries available.
6.) Read the search terms report weekly and do something with it every single week. This is the entire point of the campaign.
The graduation rule
A query you find in DSA is not a result. It is a candidate. Here is the rule I use to decide what happens to it.
Graduate it if the query names a problem you solve, a product category you compete in, or a competitor, and it came from someone who could plausibly buy. Add it as an exact or phrase match keyword in the standard search campaign that owns that theme, with the landing page you actually want and ad copy you actually wrote. Now it has a real headline, a real page, and a real bid.
Block it if it is informational, free-seeking, job-related, or from a segment you do not sell to. It goes in the negative list, which makes the whole account better, not just DSA.
Leave it if it is a rare, oddly specific query that converted once. Long tail with no repeat volume is not worth a keyword and not worth a negative. Check it again next month.
The healthy end state is that DSA gets quieter over time. Every good query you graduate leaves the DSA campaign, because your own keyword now outranks the dynamic target for that search. Every bad query gets negated. A DSA research campaign that is still finding a lot of new spend after four months is not doing research, it is just running.
The February 2027 deadline is the reason to do this now
The automatic upgrade to AI Max changes what DSA is. AI Max keeps the keywordless matching and adds more automation on top, including expanded query matching and asset generation, with controls you have to go set yourself. My take on AI Max has not changed: turn it on where it fits, then fence it in with URL inclusions, brand controls, and negatives.
But the sequencing matters. Query discovery is worth more to you before the upgrade than after, because right now DSA is the least automated way to see what Google thinks your pages are about. Once those campaigns move to AI Max, you are reading query data through a system that is also rewriting creative and choosing landing pages, which makes it harder to tell whether a query performed because it was a good query or because the automation found a decent page for it.
So the honest plan for the next couple of quarters is to mine DSA hard, get the recurring buyer queries into explicit keywords with copy you control, and shrink your dependence on keywordless matching before keywordless matching becomes the default. This is the same argument as choosing controlled search over a black box campaign when your conversion volume is low. Control is worth paying for when every lead matters.
When to skip DSA completely
Not every account should run it.
Skip it if your site is mostly thought leadership and a handful of product pages, because there is not enough indexed product content for Google to match on and the matches you get will come from blog posts.
Skip it if you sell to a very small number of companies. When the total addressable audience is a few thousand accounts, discovery by query volume is the wrong instrument, and the money belongs in account-level targeting.
Skip it if your conversion tracking cannot tell a qualified lead from a form fill. DSA's whole failure mode is looking good on cheap conversions. Without a qualified signal, you have no way to catch it.
And skip it if nobody on the team has a standing weekly slot to read search terms. An unread DSA campaign is a budget leak with a nice cost per lead.
What to do this week
Pull your search terms report from the last 90 days and count how many distinct converting queries you found that were not covered by a keyword. If the answer is more than a handful, you have a discovery gap and a small DSA campaign will pay for itself in the keyword list it produces. If the answer is close to zero, your coverage is fine and DSA will just buy you traffic you already decided not to want.
Then put a 30 minute recurring block on the calendar for the search terms review, because that block is the actual product here. The campaign is just what generates the list.
If you want a second read on whether your account has a coverage gap or a waste problem, that is most of what we look at in a B2B Google Ads audit, and we will tell you which one it is before you spend anything on finding out.
