Smart Bidding learning period: what triggers it and how long it lasts

What is the Smart Bidding learning period?

The Smart Bidding learning period is the calibration window after you change an automated bid strategy. Google Ads may show the bid strategy status "Learning" while the campaign or portfolio adjusts to its new goal, and performance can fluctuate during that time.

Three details from Google's help pages change how you should read it:

• It is measured in conversion cycles, not days. Google says it "can take a few conversion cycles (1-2 typically)" for the strategy to calibrate, and that it can be faster when there is more conversion data.
• Learning does not stop when the label disappears. Google states that its algorithms "continue to learn even when the bidding status no longer shows 'Learning'".
• Manual CPC has no learning period. It only applies to automated bid strategies such as Target CPA, Target ROAS and Maximize conversions.

To see the status, go to Tools > Bid strategies and check the "Status" column, then hover over "Learning" to see the reason.

Which changes trigger a "Learning" status?

Google lists four reasons for a "Learning" status. All of them are changes to the bid strategy itself or to what it controls, so the hover reason tells you which kind of edit started the clock.

Notice what is not on the list: budget edits, new ads, a new landing page or a broken conversion tag. These do not appear as a learning reason, yet they can move results just as much. That is why a campaign can perform worse with no "Learning" label at all, and why the label alone never proves what caused a change.

How long does Smart Bidding learning take?

Google does not publish a fixed duration. Its help page says learning typically takes 1–2 conversion cycles and names three factors: conversion volume, conversion cycle length (the time between a click and its conversion) and the bid strategy type.

Google's "Learning and review periods" page adds a practical rule: allow around 50 conversion events before making bidding changes, and wait 50 conversions before making changes to the campaign.

Put together, the length depends on your account. The examples below are illustrative, not Google figures:

Estimate your own conversion cycle from your data (the usual time between a click and the conversion), and plan the evaluation window from that number rather than from a generic "wait 7 days" rule.

How do I read performance during and after learning?

Judge a bid strategy change only after the learning period and the conversion lag have both passed, and compare like with like. Reading results in the first days mostly measures noise and missing conversions that have not been reported yet.

1. Write down the change date and the reason shown on hover. "Setting change on 6 October" is a fact you can test later; "the campaign was learning" is not.
2. Exclude the learning window from the comparison. Compare the period after learning with a period of the same length before the change, not the days in between.
3. Wait for the conversion lag. If most conversions arrive within 7 days of the click, the last 7 days of any window will look worse than they are.
4. Look at the metric the strategy optimizes. A Target ROAS change should be judged on conversion value and ROAS; a Target CPA change on CPA and conversion volume. Higher CPA with more conversions can be the intended trade-off.
5. Check volume, not just efficiency. A lower target often reduces traffic. If impressions and clicks fall, the strategy may be bidding less, not "learning badly".
6. Do not stack changes. If you change the target again before the first change settles, you restart the clock and lose the ability to attribute the result to either edit.

The same discipline applies on Meta, where significant edits reset the learning phase. See which Meta Ads edits reset the learning phase for the comparison.

When should I use seasonality adjustments or data exclusions?

Use a seasonality adjustment for a short, predictable jump in conversion rate, and a data exclusion when conversion data was broken. Both tell Smart Bidding to treat a period differently, so both belong in your change log.

Seasonality adjustments

Google says to use them "only if you expect major changes to conversion rates", that they are ideal for short events of 1–7 days, and that they may not work as well for periods longer than 14 days. Smart Bidding already accounts for regular seasonality on its own. A one-day flash sale is a good fit; a whole holiday season is not.

Data exclusions

Data exclusions tell Smart Bidding to ignore clicks and conversions from a period when tracking failed, for example a broken tag, a site outage or failed offline uploads. Google notes that excluded conversions still appear in your reports, that exclusions set in advance usually take effect within a day, and that you should not use them often or remove one once it is applied. They are supported for Search, Display, Shopping and Performance Max, and not for manual bidding. You create them under Tools > Adjustments > Exclusions.

If you are planning a peak period, our post on the Black Friday change freeze covers which edits to lock and when.

How do I prove whether the learning period or another change moved results?

Put every change from the days before the shift on one timeline, rule out tracking first, then test one hypothesis at a time. "It's still learning" is a hypothesis, not an explanation, and it is often wrong.

1. List everything that changed. Bid strategy edits, budgets, ads, keywords, auto-applied recommendations, landing page releases, tag or consent changes, CRM imports and outside events such as a competitor promotion. Google Ads change history covers the account side for the past 2 years; site, tracking and CRM changes need to be logged by the team.
2. Rule out tracking first. If conversions fell on the same day across all campaigns, including those not in learning, suspect a tag or conversion action change before the bid strategy.
3. Match the timing. A learning effect should start on the change date and only in the campaigns attached to that strategy. A drop that started two days earlier, or that also hit Manual CPC campaigns, has another cause.
4. Test one hypothesis. For example: "The Target CPA cut from €40 to €30 on 6 October reduced volume in the three campaigns of the portfolio." Check impressions, average CPC and conversions for those campaigns only, against the same weekdays before the change.

A shared marketing change log makes this a five-minute check instead of an afternoon of digging. In Changeline, Google Ads edits can be imported automatically with the before and after value, who made them and the source (interface, API, scripts, automated rules or recommendations), and the team logs site, tracking and CRM changes on the same timeline. If you have to explain the result to a client, the guide on explaining a conversion drop to a client shows how to structure that conversation.

What should I check before changing a Smart Bidding strategy?

Most learning-period problems are planning problems: the change was made at the wrong time, together with other changes, or without a way to measure it. Run through this list first.

• Conversion tracking is stable. No tag, consent or conversion action change in the last cycle, and the primary conversion actions are the ones you want to optimize.
• Enough volume. Estimate how long it will take to reach around 50 conversions at the current rate.
• No peak in the window. Avoid starting a new strategy right before a sale or a seasonal spike, when you will not be able to tell learning from seasonality.
• One change at a time. Change the target or the strategy, not both, and keep budgets and creative stable while it settles.
• Small steps. Moving a target by a modest amount is easier to evaluate than a large jump, because the expected effect on volume is easier to predict.
• Logged with a reason. Record the date, the old and new value, the owner and what you expect to happen, so the evaluation later has a hypothesis to test.

If CPA rose after a change and you are unsure why, how to find the change behind a CPA spike walks through the full diagnosis.