- Jan 2, 2023
- 818
- 786
Hello everyone in the BHW community!
Recently, when deploying Facebook Ads campaigns, I encountered a fairly common situation: campaigns kept getting stuck in "Incomplete stage" status or not being delivered despite being approved.
Specific scenarios:
Ads approved
Account clean, payment method stable
But campaign not delivered or stuck in learning mode > 48h
Some cases report "Not delivered", "Limited learning", but the reason is unclear
Some reasons I noted:
-The audience is too narrow (less than 1,000 people)
-The number of conversions is not enough for the algorithm to learn (less than 50 conversions/ad set/week)
-The budget is too low compared to the goal
-There are many changes in the ad set or creative, causing the system to constantly reset the learning process
-The pixel does not have enough data or is not firing properly
How I solved it (and it worked in some cases):
- Increase budget slightly or switch to CPM/manual bidding if you need to control delivery
- Expand audience, limit over-targeting
- Keep campaign stable for at least 48 hours, avoid multiple edits
- Optimize creative to increase initial engagement, support algorithm with data
- Re-verify/properly verify domain, ensure events are recorded
I suspect many of you here have encountered or dealt with similar issues. If you have any other tips or insights on how Facebook's machine learning algorithms work, I'd love to hear from you!
Recently, when deploying Facebook Ads campaigns, I encountered a fairly common situation: campaigns kept getting stuck in "Incomplete stage" status or not being delivered despite being approved.
Specific scenarios:
Ads approved
Account clean, payment method stable
But campaign not delivered or stuck in learning mode > 48h
Some cases report "Not delivered", "Limited learning", but the reason is unclear
Some reasons I noted:
-The audience is too narrow (less than 1,000 people)
-The number of conversions is not enough for the algorithm to learn (less than 50 conversions/ad set/week)
-The budget is too low compared to the goal
-There are many changes in the ad set or creative, causing the system to constantly reset the learning process
-The pixel does not have enough data or is not firing properly
How I solved it (and it worked in some cases):
- Increase budget slightly or switch to CPM/manual bidding if you need to control delivery
- Expand audience, limit over-targeting
- Keep campaign stable for at least 48 hours, avoid multiple edits
- Optimize creative to increase initial engagement, support algorithm with data
- Re-verify/properly verify domain, ensure events are recorded
I suspect many of you here have encountered or dealt with similar issues. If you have any other tips or insights on how Facebook's machine learning algorithms work, I'd love to hear from you!