Home >
The Newbie Zone >
Questions and Answers
There are conversions at start, but after none (11)
06-23-2019 06:01 PM
#1
zakhryapin (Member)
There are conversions at start, but after none
Hello to all!
Sometimes I have this situation: I create a campaign, start up traffic and within a few minutes I get several conversions (3 or 4). And after for the whole day I can not get a single conversion. No campaign changes. I work with Megapush.
Tell me please, what could be the problem?
For example, today's campaign: the first screen 8am, the second now
06-24-2019 03:56 AM
#2
vortex (Senior Moderator)
Apparently it's a common issue with push traffic - I would suggest checking out the last 10-20 posts in this follow-along, where the OP, matuloo and erikgyepes discuss this exact issue:
https://stmforum.com/forum/showthrea...-The-Beginning
Hope that helps!
Amy
06-24-2019 05:50 AM
#3
zakhryapin (Member)
Thanks for your reply! I will check it
06-27-2019 04:44 AM
#4
maynzie (Moderator)
Yeah bloody push traffic can be sporadic at times, you're best not to get too attached between the daily hours, just check back and count statistics at one time each day.
You will notice everyday of the week fluctuates too with push, so having a birdseye view of the campaign over a week is usually the best approach.
At least you're getting some conversions though 
06-27-2019 05:54 AM
#5
erikgyepes (Moderator)
I used to call these lucky conversions 
But as Mr. @maynzie pointed out above there could be these swings during day.
You will have to invest some time and money into this and then you will be able to recover which hours tend work best for your campaigns/geos.
Usually daytime hours are quite good, and the performance starts dropping by evening and night time.
Though there are specific hours (lunch time) when there are drops too.
But these are my observations, could be different in different geos/verticals.
06-27-2019 10:46 AM
#6
matuloo (Legendary Moderator)
I still don't have a clear answer to this problem either, but the more I think about it and the more I look at the data, I think it has to do with the way the algo serves the ads. The news ones are thrown into the testing mode, so the algo knows how high CTR they would have and then serve them based according this results. And l think the initial testing traffic is of high quality, hence the initial sales. As soon as the algo knows the CTR it starts serving the AD in the "normal" mode, as a new one it has decent CTR so it get's a share of the good traffic for a while. But as soon as it starts to settle down, it goes down in the bidding chain and lower quality traffic kicks in. So for the ad to continue to perform well conversion wise, it has to maintain a very high CTR and a healthy bid.
06-27-2019 06:29 PM
#7
diplomat (Member)

Originally Posted by
matuloo
I still don't have a clear answer to this problem either, but the more I think about it and the more I look at the data, I think it has to do with the way the algo serves the ads. The news ones are thrown into the testing mode, so the algo knows how high CTR they would have and then serve them based according this results. And l think the initial testing traffic is of high quality, hence the initial sales. As soon as the algo knows the CTR it starts serving the AD in the "normal" mode, as a new one it has decent CTR so it get's a share of the good traffic for a while. But as soon as it starts to settle down, it goes down in the bidding chain and lower quality traffic kicks in. So for the ad to continue to perform well conversion wise, it has to maintain a very high CTR and a healthy bid.
Actually this is what I'm testing right now. Just running my campaigns to figure out which time of the day works the best and what kind of bid would work the best. Then it would be possible to automate everything and let the bot do it's job (ZP in this case). Unfortunately for me, it will be an ugly red table for a week, but sometimes you gotta do what you gotta do.
06-27-2019 06:49 PM
#8
zakhryapin (Member)

Originally Posted by
maynzie
Yeah bloody push traffic can be sporadic at times, you're best not to get too attached between the daily hours, just check back and count statistics at one time each day.
You will notice everyday of the week fluctuates too with push, so having a birdseye view of the campaign over a week is usually the best approach.
At least you're getting some conversions though

Thanks for your reply! I will test more soi offers than CC to collect more data
06-27-2019 06:52 PM
#9
zakhryapin (Member)

Originally Posted by
erikgyepes
I used to call these lucky conversions
But as Mr. @
maynzie pointed out above there could be these swings during day.
You will have to invest some time and money into this and then you will be able to recover which hours tend work best for your campaigns/geos.
Usually daytime hours are quite good, and the performance starts dropping by evening and night time.
Though there are specific hours (lunch time) when there are drops too.
But these are my observations, could be different in different geos/verticals.
Thanks for your reply! I also began to notice that
the performance starts dropping by evening. But I need more tests
06-27-2019 07:00 PM
#10
zakhryapin (Member)

Originally Posted by
matuloo
I still don't have a clear answer to this problem either, but the more I think about it and the more I look at the data, I think it has to do with the way the algo serves the ads. The news ones are thrown into the testing mode, so the algo knows how high CTR they would have and then serve them based according this results. And l think the initial testing traffic is of high quality, hence the initial sales. As soon as the algo knows the CTR it starts serving the AD in the "normal" mode, as a new one it has decent CTR so it get's a share of the good traffic for a while. But as soon as it starts to settle down, it goes down in the bidding chain and lower quality traffic kicks in. So for the ad to continue to perform well conversion wise, it has to maintain a very high CTR and a healthy bid.
Thanks for your reply! I did not even think that the initial testing traffic could be of better quality! That can be a reason of this problem.
Then I can create many new campaigns and stop them after the first conversions

This is just a joke
06-27-2019 09:44 PM
#11
matuloo (Legendary Moderator)

Originally Posted by
zakhryapin
Then I can create many new campaigns and stop them after the first conversions

This is just a joke

All jokes aside, I was doing exactly this with one network for a few months, couple years ago

Each new campaign would receive a solid amount of quality impressions that always converted great, so I was submitting up to 50 new campaigns per day... until they banned me LOL
Home >
The Newbie Zone >
Questions and Answers