Rapid Growth
When running test ads, it is important to grow slowly. By growing too fast you miss relevant data points. When LLN (Law of Large Numbers) is applied in a financial or business growth scenario; LLN suggests that a large entity that may have rapid growth may possibly not be able to maintain that speed ongoing.
There are data points from the beginning to the end.
Beginning ----- Increase ----- Decline ----- Burnout
The beginning may be quite unstable. As results increase, data points you can collect are how consumers are responding to specific content groups.
For example if you have 5 content groups and somewhere between the beginning of the ad campaign and the increase you start to see based on the sample size how many users are responding to each content group per ad.
If content group B starts to scale on day 3. If you make an assumption at that point it may be too early to add additional budget to this ad campaign. I usually prefer to wait at least 30 days consistently prior to adding budget and increasing the sample size.
By increasing the budget without adequately testing you risk reaching the burn out point without having sufficient data to make informed decisions.
If there are increases and then the decline, you can observe influence between both points prior to reaching the burnout phase. To potentially pivot.