ultra realistic man looking through a computer screen displaying a complex algorithm with worried expression

The Google Ads Algorithm and the (Dreaded) Learning Period

Google Ads' advanced algorithms learn from vast datasets to predict outcomes. Tweaking campaigns may reset this process and trigger a learning period.


by Peter Coish

Latest from the Blog

Display Advertising: Direct Buy or DSP – Which is Right for You?

Bypassing the middleman usually means a lower price for the buyer. But when it comes to display advertising, this truism ain’t, umm, true.

Google Downranks AI Content. But Google Is Paying Publishers to Create AI Content

Google, what the heck?

The Everything App Will Amount to Nothing

Elon’s cringey press release about X as the “everything app” is a case of a billionaire smelling his own farts for too long.

How Can AI Improve Your SEO?

This article is only 35% written by ChatGPT!

9 Reasons We’d Pay to Watch a Musk vs. Zuckerberg Fight

To the Vegas octagon!

Google’s ubiquitous pay-per-click (PPC) advertising platform, Google Ads, is underpinned by a sophisticated algorithm designed to harness a massive quantity of data points to predict future outcomes.

It’s an unfortunate fact of ad tech life that this algorithm take time to learn how to optimize campaign performance, and that making changes to a campaign can send the algo back into a learning period that can, at times, be infuriatingly long.

What is the Google Ads Algorithm?

Google Ads’ automated bidding algorithm employs machine learning to enhance real-time performance based on real-world data, thereby minimizing the discrepancy between predicted and actual outcomes.  As you might imagine, Google hoovers up a vast quantity of data as users surf the web, and for that reason it possesses a deep understanding of behaviours and journeys. It’s aware of a user’s position in the sales funnel, tracking their interactions with competing websites, the number of pages they visit, and even arcane data points like their YouTube watch history.

Google leverages a myriad of signals, including demographic factors, browsing habits, seasonality, and even meat-space events like weather changes and elections, to predict optimal bids and, consequently, auction outcomes. These data points are collected, processed, and implemented in real time for each search query on Google’s search engine.  Accomplishing this immense task of processing such big data in milliseconds is a feat only achievable through machine learning and automation. And the exact details about how it works is a secret Google will never reveal.

You Still Have Control Over Your Campaigns

Even though automated bidding in Google Ads takes care of the bid management, it doesn’t mean you lose all control over your campaign.  There are still several significant aspects of your Google Ads campaign that remain within your control:

  • Campaign Goals: You define what you want to achieve with your campaign, be it website visits, conversions, brand awareness, or sales
  • Budget: While Google’s algorithms determine individual bid amounts, you still set your daily or campaign budget
  • Ad Creative: You’re in charge of creating your ad copy, including headlines, descriptions, and display URLs. It’s also up to you to design and upload images or videos for display and video campaigns
  • Keywords and Negative Keywords: For search campaigns, you still decide which keywords trigger your ads, as well as specify negative keywords to prevent your ads from showing up for irrelevant searches.
  • Audience Targeting: You can specify who sees your ads based on demographics, interests, search behavior, and more. You also control retargeting lists for remarketing campaigns
  • Ad Extensions: You can enhance your ads with additional information using ad extensions. These might include phone numbers, location details, additional site links, structured snippets, and more
  • Ad Scheduling: You can control when your ads are shown by day and hour
  • Location Targeting: You can specify the geographic areas where your ads will show
  • Campaign Settings: You have control over various campaign settings like language, networks (Google search, search partners, Display Network), and devices

When you make changes to these campaign parameters, you could find yourself in the dreaded “learning period”. The learning period serves as an analytical phase, during which Google’s algorithm digests the campaign changes and amasses performance data about your ads to fine-tune them for optimal performance. The significance of the learning period cannot be overstated.

And it cannot be avoided.

Perhaps the hardest part of the learning period is that it requires us to tread a fine line between patience and proactivity. Easier said than done when you are under pressure to deliver results. It’s critical not to make knee jerk campaign changes, as they can disrupt the learning algorithm, extending the learning process and potentially undermining the campaign’s effectiveness.

How Long is the Learning Period?

Again, Google is secretive. Their official line is that the duration of the learning period is determined by these factors:

  • The number of conversions that your campaigns, ad groups, keywords, or products obtain
  • The duration of your conversion cycles. For example, the amount of time it takes for a click to result in a conversion
  • The bid strategy (Enhanced CPC, Target CPA, Target Return on Ad Spend, Maximize Conversions)
  • The type and extent of the campaign changes.  Minor copy changes or additions are unlikely to trigger it. A significant change to daily budgets will

Our experience is that significant Google Ads campaign changes result in a learning period that is anywhere from 7 to 21 days.

And those days can be excruciatingly long if you need results today.