When a new campaign starts running on AI DASH, it enters a learning phase.
During this period, the system collects performance signals and learns how to deliver your campaign more effectively toward the goal you selected. Some variation in performance during this stage is normal.
What is the learning phase?
The learning phase is the initial period when AI DASH gathers delivery and performance data for your campaign.
As more data becomes available, the system can better understand which opportunities are more likely to support your campaign goal and optimize delivery accordingly. Because the system is still learning, results during the first few days may not reflect the campaign's longer-term performance.
How long does it last?
The learning phase typically takes 10–20 days.
The exact time may vary depending on factors such as campaign delivery, available traffic, and the amount of performance data collected.
We recommend allowing the campaign enough time to run before making conclusions based on early results.
What to expect during learning
You may notice more day-to-day variation than usual during this period.
For example:
Performance may fluctuate. Metrics such as CPM, Results, and Cost per Result may fluctuate as delivery patterns, auction conditions, and available performance data change.
Delivery may take time to stabilize. Campaign pacing can vary during the early stages.
Early results may not represent long-term performance. A few days of data are usually not enough to evaluate the campaign as a whole.
Instead of focusing on individual daily changes, look at the overall trend as the campaign continues to run.
Help your campaign learn effectively
The most important step happens before launch: make sure your campaign is set up around the right objective. Choose the appropriate campaign goal, confirm your tracking setup, and use targeting and creatives that match what you want the campaign to achieve.
Once the campaign is running, give it enough time to collect meaningful data. Avoid stopping and recreating a campaign based only on short-term fluctuations unless there is an issue with the campaign setup. After the learning period, you can evaluate performance using a broader range of data and make more informed decisions about your next campaign.