Good start and then crash (11)
Hi guys,
I have a weird problem with my campaigns (not always but many times).
In many cases I noticed situation where first day of campaign was okey (small negative / break-even / small profit) with a few conversions. Then, next day campaign after removing a few bad placements (mainly these with 0-5% CTR on LP and 30-50 visits or something like that where normally I have 20-30%) is so much worse. Like from 5-6 conversions on 1st day it's just 1 on another. Day after that it's again 1 or even 0.
Do you maybe noticed situation like that in the past?
Is that means that I should move to next offer and give up? Start was promising but in long term it should be better, not worse. And it looks exactly the opposite. I understand that push campaigns could be very short-term but I don't think as that short.
Any ideas?
Thanks!
Dawid
Such stuff can happen all the time.
You start a campaign and get some early conversions so that the campaign looks very good.
When then no more conversions come in for some time it seems that performance drops a lot.
But a campaign often needs time.
To get a good impression of the overall campaign performance a campaign has to run for a longer timeframe.
When you don't have enough traffic or only very few conversions then every single conversion has a huge impact on the campaign performance.
Such swings can happen between morning and evening, week day and weekend or other factors.
The campaign had conversions so it seems to have potential.
Give it some more time so that you get a better impression of the overall performance.
It also would help when you could post more info.
How much traffic did you run, how mud did you spend, how many offers and landers did you test, how long did the campaign run and so on
You'll get better at this as you keep doing it 
I'm no pro, but...
For SOI Push, if I throw $10-$20 on it on a proven source with a proven lander and 0-1 conversions come in: then I know it's a shitty offer (for me). It could be a shitty lander or something. But it's honestly easier to move on and try another offer than it is to try to squeeze it for more potential.
For CC submits, it's not much different. But you do have to have a bit more patience and tolerance for the loss. Depending on the payout, I'd blow $40-100 before giving up. Remember with CC submits, all it would take is one good hour to make a huge difference.
Aside from ZeroPark, blacklisting zones (on push) after 1 day is a bit early. I'd let it shake out for a week or so (depending on the payout) before touching any placements. Push swings are way greater than pops, so I would hold off on blacklisting.
I don't really have good data right now because these situations happened to me some time ago. But right now I run campaign and situation is similar:
First day of campaign:
Spent: $20.36
Rev: $16.50 (6 conv)
ROI: -16%
Second day (for now):
Spent: $18.26
Rev: $5.54 (2 conv)
ROI: -69%
But as you say this kind of flactudations are probably normal. But it's not the first time I see this situation, keep it for a day or two more and it's worse and worse. Let's see how it will goes this time.
About offer, landing and creatives:
It's dating campaign which I try to scale. I got lander from spy but changed it A LOT. I also tested a lot of creatives and now I just scale. So I don't really test other options because when I tried in the past they were always worse than my current. So it seems that my only leverage in this case is: removing bad placements / trying different bids. And when I start with -16% I think it's possible but with -69% it's different thing I guess.
About BL:
I usually don't blacklist too much but when I see that some placement has like 50 visits and 0 clicks to offer (when normally I have 20-30%) I think it's good time to remove it. Or am I wrong and I should keep it anyway for a while?
Edit: Maybe I just found my main problem - I don't give my campaigns too much time, heh.
You gotta be patient with pushtraffic.
Push requires a lot of data to optimize. Remove all bad sources, but let them run for a while.
You need atleast 100 visits per source before you can determinate if its bad or not in my opinion.
So guessing you're promoting soi offers, I would say go with 3x the payout on each source (Maybe you have a list from previous campaigns which sources didnt work. Start without those)
Once you've found the profitable sources, start testing those who didnt work again while scaling the profitable ones.
Lot of good points here. But one thing I would add that seems to be missing is that keep in mind that when you run push traffic you (most likely) bid CPC. But from an ad network standpoint they want to maximize CPM which means they want to find the best balance of high CPC vs high CTR.
So when you submit a new ads the ad exchange knows your CPC but not your ad's CTR so they have to run it and compare with other advertizers. Only way to do that would be to send your traffic to a variety of sources which likely include high quality ones. As such the initial run usually means you could end up with higher quality traffic than your ad actually deserves but this quickly goes away if they realize your ad CTR is only average and if your bid isn't high enough then you get muscled out by your competitors. (I also suspect new ads get favored over old ones since I can imagine an ad exchange would, rightly so, assume an ad running for 20 days would perform worse than a new ad most of the time).
Best way to spot this is to look at your top performing sources and see if you have consistently been getting traffic from them even after the drop off in performance. If you aren't then that would be your answer, but if you are then it is more likely to be a volatility issue.
Originally Posted by sushiparlour
So let's say that you guys have good funnel - what would be your max negative ROI after 1 day / 3 days to keep campaign alive while removing bad placements and hoping to get it positive? Again, I know that every campaign is different, everyone does it differently but let's say campaign with -50% ROI on push is still worth running? Because removing not working placements is a bit different than for example on native (where we have many more bots), right?
Right now I think that actually maybe I should do follow along but the topic suit to how this campaign looks anyway.
So here are stats after 3 and half days. All I did was removing bad placements which had 0-5% CTR to the offer and 2 x payout and no conversions.
Actually I'm glad that this campaign went this way. Because a lot of my campaigns looks like that and more than often I just pause it after 3 days and move on.
And the question is if I do something wrong and just throwing money away? Or this is kind of campaign which is hard to optimize to profit (or at least break-even)?
As I wrote on the beginning - this is scaling campaign so I had proven creatives, landing, offer. That's why I didn't test more on the beginning.
Btw. in other GEOs my lander CTR is usually 20-30%, here is much lower. So in my opinion it somehow shows the quality of that specific GEO, right?
Did you check the placements for every day? Quite often you'll see certain placements taking majority of the budget once you run longer, unfortunately most of the time those placements eatings the budget are the ones which don't really convert, so you sometimes end up in situation where 50%+ ( easily 50% ) are from a handful of placements only which drag the whole performance down.
Alternative way around this is whitelisting if your source allows it, basically having a testing campaign running, see which placements seem to work and increase the bid in a whitelist campaign for those.
You have conversions, so it’s not impossible. Not sure what geo that is, but your CPC bid seems high based on those stats. Have you tried adjusting that bid?
@fjk87 and @jaybot thanks a lot for advices. I think it's good to summary this topic somehow.
Why often start is good and then I see crash in my campaigns? Probably as @sushiparlour said - on the beginning I get the best possible placements but when creatives and offer is not strong enough it will be worse later on. Blacklisting don't really help (at least in my case and for short period of time - maybe if I'd keep it for week, two I could manage it to profit - don't know). Anyway blacklisting when your ROI is around -50% is not the best option I think. Because there are easier options sometimes.
I tried advices from fjk87 and jaybot. I created 2 new campaigns. Whitelist (but to being honest I had limited data, around 20 conversions) and campaign with lower bid (with blacklisted placements from original campaign). Whitelist didn't work and from my previous experience it rarely works. Probably because when whitelist is small you get 2nd, 3rd clicks from specific user in that day (at least that is my explanation).
But solution was so easy. I just lowered bid as jaybot advised. And I can still easily get decent amount of traffic. I was scary that I can't get the traffic with lower bid or quality will be even worse. CR is even better (okey, it's just 1st day but still). ROI is +33% Sometimes I think that AM is overcomplicated because we did it that way. I least I do. Thanks again guys. I will remember to keep it simple stupid next time.
Home > Paid Traffic Sources > Push Traffic