Home > Paid Traffic Sources > Mobile

Loss of traffic from Adcash and Startapp (20)

03-30-2016 12:14 PM #1 continental (Member)
Loss of traffic from Adcash and Startapp

Hello! There is a big problem - huge traffic loss, in especially with Adcash. I understand when a loss of 10%, but 20-50 is already too much.

To solve this problem i tested different lp (even tried direct) - losses remain.
Tested different geo (including europe) - losses remain.
Tested different trackers - losses remain.

I thought that i should buy VPS in every country, but i have VPS in Germany and i have losses in europe country. As I understand, traffic sources have bot publishers that click on the ad and immediately terminate the connection, resulting in no tracker transitions. And % of these publishers is constantly changing, so i check every day, how much more / less I l will lose traffic. Do you have such problems? How to fight them?

03-30-2016 01:15 PM #2 manu_adefy (Veteran Member)

Everyone has the same problem and you just have to adjust your prices and expectations.

Something that comes to mind that you can do if the loss is very high is check with support how many impressions/visitors some placements delivered and then seeing if there's a huge one with many losses. When you get >20% clickloss, there's a high chance you get a couple of bad placements just killing you with bots. Once it's lower, all placements have some bots and you won't be able to get rid of them.

Usually, as you optimize the campaign the clickloss should also decrease to <20%. For me, it usually stabilizes at around 15% or less.

Cheers.

03-30-2016 01:48 PM #3 continental (Member)

Quote Originally Posted by manu_adefy View Post
When you get >20% clickloss, there's a high chance you get a couple of bad placements just killing you with bots.
I understand it. Support of Adcash can't ever compare my click log with their click log. So hard to work with them)

03-30-2016 04:10 PM #4 Mobidea (Veteran Member)

This is such a huge problem that we also face every day.

In our case, we see that bot publishers simulate impressions, not clicks. This means that we see a lot of impressions without clicks directly on the DSP.

It’s also important to understand that the spot you choose to work on can be more prone to bot activity. For instance, if you work with Popunders there’s a higher percentage of “lost clicks” than if you work with banners.

We suffer from the same problems as we count unique visits. This is something adnetworks should work on, investing in proper anti-fraud software.

03-30-2016 05:24 PM #5 servandosilva (Member)

Quote Originally Posted by Mobidea View Post
This is such a huge problem that we also face every day.

In our case, we see that bot publishers simulate impressions, not clicks. This means that we see a lot of impressions without clicks directly on the DSP.

It’s also important to understand that the spot you choose to work on can be more prone to bot activity. For instance, if you work with Popunders there’s a higher percentage of “lost clicks” than if you work with banners.

We suffer from the same problems as we count unique visits. This is something adnetworks should work on, investing in proper anti-fraud software.
Additionally, there are some bot placements with super hight CTRs too. Sometimes your average CTR is 20% and some placements get 70-80% with 0 conversions or even above 100% CTR.
As Manu said, if you end up filtering them your clickloss gets reduced.

03-30-2016 06:56 PM #6 matuloo (Legendary Moderator)

As others have already mentioned, there is not much you can do here, unfortunately. Count with the clickloss as a part of the cost of doing biz, cut placements with the highest % of discrepancy, fine tune your hosting so it loads as quickly as possible ... and in case of significant loses due to bot placements, ask for a refund from the traffic source. Thats about it.

03-30-2016 07:31 PM #7 continental (Member)

Quote Originally Posted by Mobidea View Post

In our case, we see that bot publishers simulate impressions, not clicks. This means that we see a lot of impressions without clicks directly on the DSP.
Yes, the problem is that this % is always changing.

03-30-2016 07:32 PM #8 continental (Member)

Quote Originally Posted by servandosilva View Post
As Manu said, if you end up filtering them your clickloss gets reduced.
How to filter them? By ip? Or making white list?

Quote Originally Posted by matuloo View Post
As others have already mentioned, there is not much you can do here, unfortunately.
It's very bad. I understand if clickloss was 10-15%, but 20-30... Startapp made refund, Adcash support says:

"As previously mentioned, there can be discrepancies between third party tracking systems. We cannot comment on why you are not seeing the amount of impressions there as you can see from our system."

03-30-2016 07:51 PM #9 andrea1987 (Member)

Forget about ad cash they have a lot of BOT traffic, i honestly don't recommend it.

03-30-2016 10:34 PM #10 matuloo (Legendary Moderator)

Quote Originally Posted by continental View Post
Startapp made refund, Adcash support says:

"As previously mentioned, there can be discrepancies between third party tracking systems. We cannot comment on why you are not seeing the amount of impressions there as you can see from our system."

Yeah well, some networks are more willing to admit they sold you bots than others ...

You should filter out (block) the placements that send you the most bot traffic, or those that give you the biggest discrepancy between what they claim they sent you and what you registered in your tracking. Also look at IP reports, you might find out some placements send you tons of hits hits from the same IPs, block those too.

03-30-2016 11:33 PM #11 continental (Member)

Quote Originally Posted by matuloo View Post
You should filter out (block) the placements that send you the most bot traffic, or those that give you the biggest discrepancy between what they claim they sent you and what you registered in your tracking.
The problem is that i don't see some placements, that's why i can't block them. Traffic doesn't reach landing page) So, solution is targeting by ip if i understand.

03-31-2016 04:26 AM #12 servandosilva (Member)

Quote Originally Posted by continental View Post
The problem is that i don't see some placements, that's why i can't block them. Traffic doesn't reach landing page) So, solution is targeting by ip if i understand.
What do you mean by "not seeing them"?
If you get a report from your manager from the stats by placements and compare them with the ones you see on your tracker, you will notice that some of them will either have a lot more traffic reported than the one you received or simply you didn't receive traffic but they still appear there.

Adcash is considered more of a network to scale campaigns so you don't lose a lot of money just by blocking websites while testing your campaign IMO.

03-31-2016 05:02 AM #13 deondup (Member)

As far as possible you should always track DeviceID's in your mobile campaigns. If you can't then track ImpressionID's. Its the best way to prove bot traffic and gives you some hard data to go back to the ad network to ask for refunds.

03-31-2016 05:12 AM #14 simon_89 (Member)

Quote Originally Posted by servandosilva View Post
What do you mean by "not seeing them"?
If you get a report from your manager from the stats by placements and compare them with the ones you see on your tracker, you will notice that some of them will either have a lot more traffic reported than the one you received or simply you didn't receive traffic but they still appear there.

Adcash is considered more of a network to scale campaigns so you don't lose a lot of money just by blocking websites while testing your campaign IMO.
This data right here is in a span of 8 days. Is it possible to consider that a particular IP will visit the same lander given the frequency cap settings(1/24hrs)? Based off my stats at Adcash, I think the first IP is a bot since there were many visits without a click. Would you go back to adcash and show them a screenshot of this bot traffic?

03-31-2016 06:27 AM #15 cosmeivan ()

Umm, actually now that OP mentioned this issue, I've been running for a whole year with AdCash and just about 3-5 months ago I noticed the ROI dropped a lot, despite constant optimization.

So I checked my stats versus stats on my AdCash account and I see 20-30% discrepancy in clicks, this is what OP is seeing, lot of clicks not even getting registered in Voluum.

So now I will export some reports from adcash and Voluum and compare them side by side in Excel.

brb

03-31-2016 08:10 AM #16 cosmeivan ()

So I made some tables in Excel and compared side by side. The data is just for a few hours this campaign ran yesterday. I stopped because the discrepancy was already 20%

You can see I marked in red the bad zone ids that appear in Adcash but not in Voluum. To the right there are 3 custom columns I added: 1st col checks if zone ids in the same row match, 2nd col is the difference in impressions, and 3rd col is the $$ amount lost.

Now I will export data for a whole month, this will take me more time because I gotta offset rows manually when there is a mismatch.

03-31-2016 10:29 AM #17 matuloo (Legendary Moderator)

Quote Originally Posted by cosmeivan View Post
So I made some tables in Excel and compared side by side. The data is just for a few hours this campaign ran yesterday. I stopped because the discrepancy was already 20%

You can see I marked in red the bad zone ids that appear in Adcash but not in Voluum. To the right there are 3 custom columns I added: 1st col checks if zone ids in the same row match, 2nd col is the difference in impressions, and 3rd col is the $$ amount lost.

Now I will export data for a whole month, this will take me more time because I gotta offset rows manually when there is a mismatch.
This is what I had in mind too, if adcash reports clicks for some zone and you dont see the zone in Voluum at all or with way lower hits, block it as its most likely bots.

04-01-2016 01:53 AM #18 cosmeivan ()

Hey guys sorry, I felt asleep last night, so I finished the Excel sheet with data from a whole month, just a little camp from AdCash and I'm surprised with what I found.

Voluum Data:

Total Impressions: 648,482
COST: $1,591.06 (I calculate CPM a bid higher)

AdCash Data:
Total Impressions: 828,987
Cost: $2,050

Difference:
Impressions: 180,505
Cost: $537.67


That is 30% more impressions billed than the ones that got to my tracker. Imagine what I will find when I go 6 months back and also for the other camps.
My margin on that camp was exactly 30%, so my profit is none, zero, nill

This is in part my fault for only relying on the data from Voluum, since AdCash timezone is different I always thought the discrepancy was because of that.

This morning I told my manager from Adcash about the issue and he told me "I will be in a meeting, will be back", and never heard from him again lol.

I have attached the complete excel sheet with all the data for the past month, so you can benefit and grab the baddest publishers ( I noticed at least 10+ pubs with overspends over 500%) and make your blacklist.

https://dl.dropboxusercontent.com/u/...arch%2031.xlsx

04-06-2016 03:29 PM #19 matuloo (Legendary Moderator)

Quote Originally Posted by cosmeivan View Post
Hey guys sorry, I felt asleep last night, so I finished the Excel sheet with data from a whole month, just a little camp from AdCash and I'm surprised with what I found.

Voluum Data:
Total Impressions: 648,482
COST: $1,591.06 (I calculate CPM a bid higher)

AdCash Data:
Total Impressions: 828,987
Cost: $2,050

Difference:
Impressions: 180,505
Cost: $537.67


That is 30% more impressions billed than the ones that got to my tracker. Imagine what I will find when I go 6 months back and also for the other camps.
My margin on that camp was exactly 30%, so my profit is none, zero, nill

This is in part my fault for only relying on the data from Voluum, since AdCash timezone is different I always thought the discrepancy was because of that.

This morning I told my manager from Adcash about the issue and he told me "I will be in a meeting, will be back", and never heard from him again lol.

I have attached the complete excel sheet with all the data for the past month, so you can benefit and grab the baddest publishers ( I noticed at least 10+ pubs with overspends over 500%) and make your blacklist.

https://dl.dropboxusercontent.com/u/...arch%2031.xlsx
Never rely just on what you see in voluum or any other tracker. Always double check with the traffic source, its never EVER gonna match.

Even when optimizing and cutting placements, always compare the income with the real costs at the source. In case of smart bidding alogos, the price per click/hit/impression will be different for every placement so when you look at trackers data with the same cost setting for all placements, it simply cannot match. The differences can be pretty big in some cases.

04-13-2016 11:51 PM #20 herefornow (Member)

Lol , typical...

Home > Paid Traffic Sources > Mobile