Smart Bidding Learning Phase: 50 Conversions

TL;DR: Smart Bidding calibration depends on how many conversions a strategy has seen, not on calendar weeks. Trade press reports that Google's help article now says up to roughly 50 conversion events or 3 conversion cycles, but the page I checked still showed older "1-2 conversion cycles" wording, so treat 50 as reported, not confirmed. An account converting 10 times a month would take about 5 months to reach 50, so low-volume accounts should move the primary conversion earlier in the funnel.

When a client asks "why hasn't my new bid strategy kicked in yet," the honest answer was never really about weeks on a calendar, it's about how many conversions that bid strategy has actually seen. Google's own help documentation is starting to say the same thing more directly, and it's worth walking through what changed, what I could actually confirm myself, and why the number matters more than the calendar does.

What's being reported

Trade press (first flagged by marketer Hana Kobzová and picked up by Search Engine Land and other PPC outlets in late September 2026) reports that Google rewrote its help article on how long a Smart Bidding strategy takes to calibrate. The reported change: instead of framing calibration mainly in elapsed time, "up to 3 weeks," or "1-2 cycles," the updated wording describes it as taking up to roughly 50 conversion events or 3 conversion cycles, with the explicit caveat that it can be faster if there's enough historical conversion data.

What I could verify myself, and what I couldn't

I checked Google's current help article directly, "Duration of the learning period for campaigns and what affects it," rather than take the trade press summary at face value. As of writing this, that specific page still reads: calibration "can take a few conversion cycles (1-2 typically)... although it can be faster depending on the amount of conversion data present." I didn't find the "50 conversions" figure sitting in that sentence.

A few things worth being straight about here:

Google's related article, "Tips on measuring Smart Bidding performance," already tells you to assess results over "a month, or at least 50 conversions," and to allow "1-2 conversion cycles" before adjusting anything further. So the number 50 wasn't foreign to Google's documentation before this, it was already being used as a performance-assessment benchmark, just not explicitly as the calibration threshold itself.
Google has also, for a while, published a table specifically for Display campaigns that ties conversion volume directly to expected learning time: 50 conversions maps to "up to 3 weeks," 100 to 500 conversions maps to "up to 2 weeks," and fewer than 30 conversions maps to "up to 4 weeks." That's on a different, Display-specific page, but it shows Google has already been connecting conversion counts to week-based estimates elsewhere in its own docs.
The documentation applies to automated bid strategies, Target CPA, Target ROAS, Maximize Conversions, Maximize Conversion Value, and the campaign types built on them. It explicitly does not apply to Manual CPC, which has no learning period at all.

My honest read: either the rewrite hasn't reached every locale or page yet, it's being tested on a subset of accounts, or it's already been partially rolled back since being spotted, Google does this with help-center wording more often than advertisers realize. I'm not going to present "50 conversions or 3 conversion cycles" as settled, word-for-word Google policy when my own check of the live page didn't match it. What I will say is that the direction of the change, tying calibration to a conversion count rather than a week count, lines up with language Google already uses elsewhere in its own documentation, and that's useful regardless of which exact sentence is live on which page this week. If you're reading this more than a few weeks out, it's worth clicking through to the live page yourself and checking the current wording before quoting a specific number to a client.

I couldn't independently confirm a separate claim, that Google also added Performance Max to its conversion-goal best-practices documentation around the same time. Performance Max already has its own, fairly extensive conversion-goal guidance (including a dedicated page on best practices for lead generation), and I found no clear evidence that's a recent addition rather than existing content. I'm flagging it rather than including it as a confirmed fact.

Why this is mainly ammunition for a client conversation, not a settings change

Here's the part that actually matters for how you run an account, and it's true regardless of whether the exact number Google settles on is 50 or something close to it.

When a bid strategy is new or just changed, the question that matters isn't "how many weeks has it been running." It's "how many conversions has it actually collected in that time." Those two questions can point to wildly different answers depending on account volume.

An account converting 10 times a month hits 50 conversions in roughly 5 months. "Up to 3 weeks" and "5 months" are not the same conversation to have with a client.
An account converting 50 times a month hits that mark in about a month, much closer to what people assume "learning phase" means.
An account converting 200 times a month can clear it in under two weeks.

Same bid strategy, same Google algorithm, three completely different real-world timelines, and the only variable is conversion volume. That's the gap between "weeks" and "conversions" that trips up a lot of client expectations, and it's why I'd rather set expectations around the conversion count from day one than let a client count calendar weeks and get anxious when week four rolls around with nothing to show for it.

Three things worth actually doing with this

Don't touch the bid strategy before it's actually seen real conversion volume. Resist the urge to judge, or worse, change, a Target CPA or Target ROAS strategy at the two-week mark just because that's what used to sound like the finish line. Check conversion count, not the calendar, and give it a real chance to calibrate before adjusting targets or swapping strategies.

For low-volume campaigns, move the conversion action up the funnel. If a campaign is only generating a handful of purchases or qualified leads a month, the primary conversion action is working against itself, it can't accumulate enough volume to ever finish learning in a reasonable window. Pointing the bid strategy at an earlier-funnel action, add-to-cart instead of purchase, a qualified lead form instead of a closed deal, gives Smart Bidding enough signal to actually calibrate, while the harder, lower-volume action can still be tracked as a secondary conversion for reporting. I go through the primary-versus-secondary distinction in more detail in what is a Google Ads conversion.

Don't launch a new bid strategy right before Q4. A strategy switched on in early November doesn't have time to calibrate, under either the old "weeks" framing or the new "conversions" one, before Black Friday and holiday traffic hits. Walking into your highest-volume, highest-stakes weeks of the year with a bidding algorithm that's still learning is a bad trade. Make strategy changes in September if you're planning for Q4, not after the rush has already started.

Frequently asked questions

How long does Smart Bidding actually take to learn? Google's documentation ties this to conversion volume and conversion-cycle length rather than a fixed number of days. Trade press reports Google's updated guidance points to roughly 50 conversions or 3 conversion cycles, though the specific page I checked directly still used older "1-2 conversion cycles" wording as of this writing, so treat the exact figure as reported, not confirmed word-for-word.

Does the learning period apply to all bid strategies? It applies to automated Smart Bidding strategies, Target CPA, Target ROAS, Maximize Conversions, and Maximize Conversion Value, and the campaigns built on them. Manual CPC has no learning period at all, since there's no algorithm calibrating anything.

What should I do if my account only gets a handful of conversions a month? Move your primary conversion action earlier in the funnel, something like add-to-cart or a qualified lead instead of a completed purchase, so Smart Bidding has enough volume to calibrate in a reasonable amount of time. Keep the harder, more valuable action tracked as a secondary conversion so you're still reporting on it.

Why shouldn't I launch a new bid strategy right before Q4? Because it won't have time to calibrate before your highest-traffic weeks of the year arrive, whether you measure that in weeks or in conversions. Make bidding changes in September if Q4 matters to your business, not in November.

Is the "50 conversions" figure official Google policy? It's reported by multiple PPC trade outlets as a rewrite of Google's help documentation, not something Google announced directly. When I checked the specific Google help page myself, it still showed older wording, so I'd treat the number as a useful directional signal rather than a guaranteed, fixed threshold.

If you're not sure how many conversions your account is actually generating a month, or whether your primary conversion action is set up to ever clear a learning phase, that's exactly what conversion tracking is for. And if you want someone watching bid strategy timing against your actual campaign calendar, not just toggling settings and hoping, that's part of what Google Ads management covers.