A bid strategy change is a fair A/B test in Google Ads: same ads, same landing pages, same budget, one setting different, and a result you can read in a couple of weeks. A switch from optimizing on MQLs to optimizing on SQLs usually isn't. The deeper goal takes weeks to produce each conversion, so for most of the test both versions chase the same people and neither learns much. In long-cycle B2B, we commit to the goal switch and watch it.
The bid strategy test that worked
On a B2B SaaS account we worked on in late 2024, a campaign on Maximize conversions was getting few conversions, and few impressions and clicks with them. The founder wanted to know whether Maximize clicks would bring in more converted leads. So we tested it: a second campaign with only the bid strategy changed. The ads, the URLs and the budget stayed identical, and both ran for two weeks.
We stopped Maximize clicks. It brought people happy to click and much less likely to fill in a form, which fits Google's definition: Maximize clicks isn't Smart Bidding, a separate trap we cover in your campaign says it's optimizing for leads.
Why a goal switch is a different kind of change
On a fulfillment account we run across three markets, we moved optimization from MQLs to SQLs with no side-by-side test. Someone becomes an MQL in a day. An SQL can take three weeks: follow-up calls, a booked meeting that people often skip, and finally a completed call. The switch took around two months on average to show results.
A two-month MQL campaign against a two-month SQL campaign wouldn't have shown us much. With the same bid strategy, both reach the same kind of people at the start, because the SQL signal hasn't built up enough to steer delivery anywhere different. Our read is that different bid strategies overlap less, since Maximize clicks and Maximize conversions go after different people.
To be precise about the overlap: Google says keywords in one account don't compete with each other in the auction, and only one can trigger an ad for a search. So duplicates don't bid up your CPCs. The cost is that Google, not a random split, decides which campaign serves each search.
The exception is a short cycle. The marketer who asked us had a two to three week sales cycle, and there a two-week goal test, comparing which campaign brings more qualified leads, can be worth running.
Run the test in Experiments, not a duplicate
Google's custom experiments create a trial from your original campaign that shares its traffic and budget, and Google says they're typically used to test Smart Bidding, keyword match types, landing pages, audiences and ad groups. From the docs:
- The split. On Search you choose a cookie-based or search-based split. Google says the split controls auction eligibility but doesn't guarantee equal impressions or spend.
- Eligibility. The campaign must be active, on its own daily budget rather than a shared one, and on a standard bid strategy rather than a portfolio. Only one experiment can run at a time.
- Timing. Allow 7 to 14 days for the trial to stabilize. If there's no result yet, Google recommends at least 4 to 6 weeks.
- Volume. Google's page says reliable results need minimum conversion volume, citing "more than 100 daily conversions" as an example. Most B2B accounts are nowhere near that. Google also lets you choose whether the winner is decided on directional results or on statistical significance.
Google's pages don't spell out conversion goal changes as a test type, and our objection isn't about the tool anyway. A new goal needs weeks of conversions before either arm tells you anything.
The takeaways
- A bid strategy change is a clean A/B test: one variable, readable in weeks.
- An MQL-to-SQL switch in a long sales cycle usually isn't worth testing side by side. Commit to it and monitor it.
- With a two to three week cycle, a two-week goal test can work.
- Use Google's custom experiments rather than a duplicate campaign, and check eligibility and conversion volume first.
Common questions
Can you A/B test a bid strategy in Google Ads?
Yes. Google's custom experiments create a trial from your original campaign that shares its traffic and budget, and Google says they are typically used to test Smart Bidding, keyword match types, landing pages, audiences and ad groups. For Search campaigns you can split by cookie or by search, and Google says the split controls auction eligibility, though it doesn't guarantee equal impressions or spend.
Should I A/B test switching my conversion goal from MQL to SQL?
With a long sales cycle, we usually don't. An SQL can take weeks to happen, so the new goal takes a long time to build enough conversions to steer delivery, and in the meantime both versions chase the same people. We make the switch and monitor it. With a short cycle of two to three weeks, a two-week side-by-side test can be worth running.
Why not just duplicate the campaign and change one setting?
Google says that when several keywords in one account can match the same search, they don't compete in the auction and only one of them triggers an ad. So duplicates don't bid each other up, but Google, not a random split, decides which campaign serves each search. A custom experiment splits traffic between the original and the trial instead.
How long should a Google Ads experiment run?
Google's custom experiments page says to allow 7 to 14 days for the treatment arm to stabilize and notes the bid strategy learning phase typically lasts 7 days. If results still show as in progress, undecided or unavailable, Google recommends letting the experiment run at least 4 to 6 weeks.
Which campaigns can use a custom experiment?
Google says the campaign must be active, use its own daily budget rather than a shared budget, and use a standard bid strategy rather than a portfolio strategy. Standard custom experiments aren't supported for App or Shopping campaigns, and you can only run one experiment at a time.