Google call-only ads are going away: you can't create new ones since February 2026, and every existing call-only ad stops serving in February 2027. Google's replacement is a responsive search ad with a call asset. The migration looks like a format swap. It isn't. A call-only ad had one job and one outcome, the phone ringing. The replacement gives searchers a second option, clicking through to your website, and you pay for that click too.
If your business lives on inbound calls (plumbing, HVAC, legal intake, vet clinics, towing, roofing), the risk isn't that calls stop. It's that some of them quietly turn into website visits that never call, while your cost per call creeps up and nobody notices until the monthly report.
Here's how we're handling the move for the service-business accounts we manage, and the parts that break if you just let Google's suggested migration run.
What exactly is changing with call-only ads?
Google announced the deprecation in its call ads to call assets transition notice, and there are two dates:
- February 2026: creation of new call-only ads was disabled. That already happened.
- February 2027: all existing call-only ads stop receiving impressions.
Between those dates your existing call-only ads still run. But they're frozen. Editing an ad in Google Ads retires the old ad and creates a new one, and new call-only ads can't be created. So the ad copy, number and settings you have today are what you're stuck with until they switch off.
The window to migrate on your own schedule is right now, not in January.
Why isn't a call asset just a call-only ad with extra steps?
Because the economics of the click are different.
With a call-only ad, the searcher sees your business name, a couple of lines of copy and your number. Tap anywhere and the phone dials. There's no website. Every paid click is a call attempt.
With a responsive search ad plus a call asset, the searcher sees a normal search ad with a call button attached. They can:
- Tap the call button. Same outcome as before: a call.
- Click the headline. They land on your website. You pay for that click, the same as any search click.
Some of those headline clicks will call from the page. Some will fill out a form. A lot of them, especially on mobile, will look around for eight seconds, hit back and call the competitor whose number was easier to find.
Here's a simple way to see the problem. Say a call-only ad got 100 clicks and every one of them was a call attempt. Put the same budget behind an RSA with a call asset and maybe 55 of those 100 clicks tap the call button and 45 go to the site. If the landing page turns a third of those 45 into calls, you get about 70 call attempts for the same spend. That's a 30% jump in cost per call, and nothing in the account looks broken.
Those percentages are illustrative, not a benchmark. Your split depends on the query, the device and how badly someone needs a plumber at 11pm. But the direction is the point. After this migration, your landing page is part of your call funnel whether you planned for it or not.
How do you migrate call-only ads without losing calls?
This is the order we run it in.
1. Export the call-only baseline before it disappears
Pull 12 months of call-only performance by campaign, ad group, device and hour of day: calls, cost per call, average call length and, if your call tracking or CRM has it, booked jobs. Once these ads stop serving in February, that comparison gets a lot harder to rebuild.
Pay special attention to hour-of-day. Call-only campaigns are often scheduled tightly around office hours, and that schedule needs to carry over to the call asset.
2. Build a page whose only job is to make the phone ring
The RSA needs a final URL, so give it one that behaves like a call-only ad:
- Tap-to-call number in the first screen on mobile, large, not in the footer
- A sticky call bar that follows the scroll
- The same number as the call asset, in server-rendered text (this also keeps you clear of the unverified phone number disapproval)
- Service area, response time and one trust signal (reviews, license number, years in business)
- No main navigation, no blog links, nothing that invites browsing
A homepage is the wrong final URL for a former call-only campaign. Treat the page as a delivery mechanism for a phone number.
3. Write RSAs that sell the call
Call-only ad copy was written to get a tap. RSA copy tends to drift toward generic brand messaging because Google suggests it. Push back. Headlines like "Licensed Plumber, On Site in 60 Min" or "Call Now, Live Answer 24/7" keep intent pointed at the phone. Pin a call-focused headline to position 1 if your RSAs keep rotating into weaker combinations.
Attach the call asset at the ad group or campaign level rather than relying only on an account-level number, so emergency services, estimates and different locations can each show the right number.
4. Schedule the call asset to hours someone actually answers
Call-only ads that ran 24/7 into a voicemail box were wasting money already. Now's a good time to fix that. Schedule the call asset to staffed hours. Outside them, the RSA keeps serving and sends people to the landing page, where a form or an after-hours voice agent can pick them up.
If you've been meaning to answer after-hours calls with AI, this migration is a natural point to do it. Just make sure the AI-answered calls still get attributed back to the ad. We wrote up the handoff that keeps that intact in AI voice agent call tracking. If a live answer isn't possible yet, missed call text back is the minimum.
5. Fix conversion tracking before shifting budget
This is where most migrations quietly fail. Set it up in this order:
- Turn on call reporting so the call asset uses a Google forwarding number and calls show up in reporting at all.
- Create a "Calls from ads" conversion action with a minimum call length that reflects a real lead. Google's default is 60 seconds. For a business where booking takes three minutes, 60 seconds may be too generous. For a quick-quote business it might be too strict.
- Add website call conversions for the landing page, so a visitor who clicks the headline and then calls from the page still counts.
- Decide what's primary. If Smart Bidding optimizes toward every 15-second hang-up equally, it'll find you more hang-ups.
If you already import qualified leads or booked jobs from a CRM, keep those as the primary goal and let call conversions inform rather than drive. This is the same principle we apply to filtering invalid leads before Smart Bidding sees them: feed the algorithm outcomes, not activity.
6. Run both side by side, then switch off the old ads yourself
Launch the RSA version in the same ad groups while the call-only ads are still serving. Give it two to four weeks. Compare cost per qualified call (not CTR, not cost per click). When the new version holds up, pause the call-only ads on your schedule rather than having Google switch them off in February 2027 while bidding scrambles to relearn.
What should you check in the first 30 days after switching?
Four numbers tell you whether the migration worked:
- Call asset taps vs headline clicks. If headline clicks dominate on mobile, your copy or your asset isn't pushing the call hard enough.
- Landing page call rate. What share of site visitors from these campaigns call from the page? If it's low, the page is leaking.
- Cost per qualified call vs your call-only baseline. This is the one that pays the bills.
- Average call length. A drop usually means lower-intent callers, often from broader queries the RSA is now matching.
Watch the search terms report too. Google is pushing broad match and AI Max hard on Search campaigns this year, and an RSA can be matched to more queries than your old call-only setup. "Emergency plumber near me" and "how to fix a leaking pipe" are not the same caller.
Who gets hit hardest by the call-only ads sunset?
From what we see in service-business accounts, three groups:
- Businesses with no real landing page. Plenty of local advertisers ran call-only ads precisely because their website was weak. That shortcut is gone.
- Accounts with call-only ads on autopilot. Campaigns built years ago, never touched, still producing calls. They'll stop in February with no warning beyond the notice nobody read.
- Multi-location businesses. Each location's number needs to land on a location-specific page and asset, and that mapping often only existed inside the call-only ads.
If you also ran Local Services Ads, there's overlap here. Google is pulling those into the main Google Ads account as well, which we covered in LSAs moving into Google Ads. Two call-heavy channels are changing in the same twelve months, so audit them together.
Our take
Call-only ads were one of the last formats where a small business could buy a phone call almost directly. Google is folding that into the same automated RSA format as everything else, and you can see why from Google's side: one format, more surface area, more clicks to sell.
For advertisers, the call still exists. It just has a website standing in front of it now. The businesses that keep their cost per call flat through this will be the ones that treat the landing page, the call asset schedule and the conversion setup as one system, and build it before February rather than in the week after their calls drop.
Get a free automation audit
If you're running call-only ads and want the migration done without a dip in calls, we'll review the account, the landing page and the call tracking setup, and show you where calls will leak once the old ads switch off. We'll also tell you whether an after-hours voice agent makes sense for your volume.
Get a Free Automation Audit and we'll map out the migration before Google does it for you.