We’re in the learning phase
Marketers love the “learning phase.” Google, Facebook, doesn’t matter. The learning phase covers a multitude of sins. It’s the ultimate get out of jail free card. It’s a convenient scapegoat for fluctuating KPIs and performance hiccups. And the time table is so subjective. Is the learning phase over after:
50 conversions?
50 conversions in 30 days?
30 conversions?
30 days?
2 weeks?
mystical $X spend?
Who knows…and that’s the delicious part we marketers love. The learning phase loves ambiguity and Overton window moving.
Was our spend crazy inefficient yesterday? Learning phase.
Did we not get enough traffic to the site? Learning phase.
Was our revenue down against forecast? Learning phase.
In this wonderland of excuses, the learning phase becomes a magic wand, waving away accountability and deflecting scrutiny. It's a comforting blanket, wrapping us in the warmth of "things are still figuring themselves out." Yes, the learning phase is an amazing excuse in the excuse tool box. And while I don’t want to completely write it off as an illegitimate excuse, I want us as marketers to understand it better so that we can use it as an excuse when it actually is the reason for KPI fluctuation.
What is the learning phase?
Simply put, the learning phase describes a period of initial data collection and optimization on Google and/or Meta. During this phase, the platform actively tests different targeting options, ad creatives, and delivery settings to determine the most effective way to reach your target audience and achieve your campaign goals for a new campaign or a significant edit to an existing campaign that lack historical data. Google and Meta differ slightly in their learning phases.
Google Ads and Meta Ads utilize different algorithms with unique learning approaches.
The duration of the learning phase can vary depending on the amount of data available. Campaigns with frequent conversions or clicks typically exit the learning phase faster.
Google Ads generally requires specific conversion tracking setups, while Meta Ads often leverages its own user data for basic conversions like website visits.
Manual CPC bidding in Google Ads bypasses the learning phase, while Meta Ads typically learns regardless of the chosen strategy.
Why does it exist?
The learning phase exists for several reasons, as it allows the platform to:
analyze data about your ads, where they're being shown, and how much you should be paying for clicks or conversions.
constantly tweak campaign settings to prioritize placements, audiences, and bids that are most likely to deliver results.
refine its approach and allocate resources more efficiently.
When does it get triggered?
Google and Meta learning phases get triggered when:
new campaigns or accounts are created OR
significant changes are made to existing campaigns
bid strategy adjustments (i.e. switching from clicks to conversions)
budget alterations (~20% or more)
targeting updates
conversion tracking settings modifications
major ad copy or keyword changes (additions, removals, pauses)
Basically, any time you start something new or make a major change, Google and Meta throw the campaign into the learning phase because it cannot rely on the previous campaign performance inputs to guide its algorithm outputs.
How long does the learning phase last?
The answer to this question is…it depends. And that’s why marketers love it. It’s a murky nebula where internal algorithms duke it out with campaign complexities. But, Google and Meta provide guidelines for what we can expect.
Google Ads: Generally exits after gathering sufficient data to optimize bids, often within 7 days for smart bidding strategies. Manual CPC bidding bypasses the learning phase altogether.
Meta Ads: Typically exits after 50 optimization events (i.e. conversions or clicks) occur since the last significant edit, but this can vary based on factors like conversion windows and campaign objectives.
However, there are factors that can impact these learning phase durations:
Conversions: The more conversions you get, the faster Google Ads can gather data and optimize your bids. Low conversion volume extends the learning period.
Conversion cycle duration: If someone typically takes a long time to convert after clicking your ad, the learning phase may take longer.
Click volume: Similar to conversions, higher click volume provides more data and potentially speeds up the learning process.
What should I do when a campaign is in the learning phase?
Being in a learning phase does not give you carte blanche nor absolve you from any downward spiral in performance. If an existing campaign goes back into the learning phase, it’s important to understand why and if this was an unforced error on the marketer’s side. Making frequent and major campaign changes means you’ll constantly be in a learning phase and never letting the platform figure out what it is you actually want. That’s a red flag.
However, if you’re in the learning phase for a new campaign, remember to set realistic budgets and goals. Avoid starting with overly ambitious goals that might not generate enough data. Monitor your performance statistics for signs of stabilization. Once fluctuations decrease and performance plateaus or improves consistently, it's a good indication you're nearing the end of the learning phase. Third, make sure you get ahead of the curve and communicate new campaign launches to the team and to your boss. Explain the learning phase to clients and stakeholders to manage expectations and avoid unnecessary panic.
In essence, change the learning phase from an excuse to a focused period of experimentation and growth. The learning phase is not a magic wand or an excuse for poor performance. It's a natural part of the optimization process, and understanding its purpose and limitations can help you leverage it for better campaign results in the long run.