The learning phase is the period in which the algorithm works out who to show a new or significantly edited ad set to, when and in which placement. Costs are volatile during this period and the results do not represent steady-state performance. The phase closes once enough conversions have accumulated.
The most common mistake is panicking mid-learning and changing budget, targeting or creative; every meaningful edit restarts learning and leaves the set permanently volatile. The second is splitting the same budget across many sets so that none of them reaches the threshold. The practical rule is to leave the set alone for a few days after launch and make the decision on weekly totals. On businesses with low conversion volume, moving the optimization event up from purchase makes the learning phase completable.
The most common mistake is panicking mid-learning and changing budget, targeting or creative; every meaningful edit restarts learning and leaves the set permanently volatile. The second is splitting the same budget across many sets so that none of them reaches the threshold. The practical rule is to leave the set alone for a few days after launch and make the decision on weekly totals. On businesses with low conversion volume, moving the optimization event up from purchase makes the learning phase completable.
The most common mistake is panicking mid-learning and changing budget, targeting or creative; every meaningful edit restarts learning and leaves the set permanently volatile. The second is splitting the same budget across many sets so that none of them reaches the threshold. The practical rule is to leave the set alone…
Knowing the term is not enough; you need to know which lever to pull. In a free 30-minute call we look at your store's actual numbers.
Get a free analysis