Network: Above All Offers
Traffic source: SiteScout
Total Budget: $100
Ad format: 300×250 Banners
LP or DL: Direct Link
Welcome to part 2 of my mobile dating follow along campaign. If you missed part 1, here’s the link for you to catch-up- (https://gossips.best/stm/sho…buy-with-J-SiN).
This campaign will only be using site placements that got clicks from the previous campaign and haven’t been disqualified. This should prevent my daily spend from being ‘nickeled and dimed’ by 10,000s of placements that aren’t producing results. Instead, it will go towards testing new banners and finding the best performing placements.
Setting Up My Campaign:
I will still be using Voluum for tracking and SiteScout as my advertising platform. I’ll be using a new campaign in SiteScout, so I can easily bid on only the placements I want instead of the RON setting of the original campaign.
Using the ‘Advanced Campaign’ option (the ‘Mobile Campaign’ has very little inventory) I set up the campaign with the following criteria-
No day parting
1 impression every 24 hours
$5.00 Max CPM Bid
$20 per day budget
Manual added select placements
Mobile Web (optimized) – ‘optimized’ means it will only show up on mobile versions of websites, not when a mobile device browses a non-mobile version of a website.
Mobile Web (not optimized)
Capture Conversions- I added a new audience list called “Love2Meet Conversions”. This feature will add people who’ve converted to my specified list for re targeting purposes. I then set up this audience to be ‘negatively’ targeted so anyone who converts on my offer will not be shown the ad again.
Now I wait for SiteScout to approve me campaign…
Link (Previous phase) – https://gossips.best/stm/sho…buy-with-J-SiN
From:J-SiN Date: 06-18-19
From:lucas Date: 06-19-19
Total spend to date: $15.44
Even though my campaign wasnt approved until 2:30 PM PST, it wasnt a bad first day. I had 35 clicks (.55% CTR) and my Impressions won were at 73%. When I analyzed the data in Voluum, I saw most clicks were on target (from UK and mobile devices), so why am I still not seeing conversions?
I reached out to my affiliate manager and he seemed as perplexed as I was. I sent him my tracking link and he confirmed that the networks tracking domain was not blocked. After further discussion, we decided the offer page was probably the issue, and may not be good enough to work direct-linking. I havent previously had much success creating my own jump-style landing pages, which actually increased overall conversions, but my affiliate manager said he would send me some proven templates to try.
From:J-SiN Date: 06-19-19
Total spend to date: $30.67
Not much action today. My Impressions Won were up slightly compared to the previous day at 77%, and I had 29 clicks (.41% CTR). I still havent seen any conversions though. I think it might be time to actually try split testing a landing page.
From:J-SiN Date: 06-19-19
Total Spend to date: $45.67
Ill start with a quick update on my stats. I had 64 clicks (.80% CTR) and my Impressions Won were at 77%, but still no conversions. I think its time to start working that new landing page.
As promised, my affiliate manger sent me some basic landing page mockups that had been successful for him in the past. These helped with figuring layout of my landing page, but now I needed to figure out how to build it. I did a search in the AFFPlaybook forum for instructions on how to build a mobile landing page and found the Bootstrap tutorial (https://gossips.best/stm/sho…strap+tutorial). I decided to use Bootstrap because I could build a webpage that would easily work on a mobile phone, that work just as well as if it were viewed on a desktop (Responsive).
Instead of reinventing the wheel, I pulled bullet points and testimonial quotes from the offer page. Then I did a search in Google images for UK selfies and choose 9 pics that looked like real self pics. My criteria for a real selfie is bad coloring, pictures out of focus, and girls not looking like models. I edited the images to the right size in paint.net and then uploaded all the landing page files to a newly created Amazon s3 bucket.
The last step was to setup Amazon Cloudfront. This will allow all the content to be served quickly to a visitor no matter where the website was accessed from. This should also help prevent a high bounce rate, because people arent bouncing due to a slow site.
Now that my landing page is ready to go, I logged into Voluum and switched out the direct link for the new landing page. I decided to not split test the landing page alongside direct linking, because I felt that we had run without a landing page for long enough to prove this didnt work.
From:J-SiN Date: 06-20-19
Placement Campaign: $60.72
Mobile Campaign #2- $15.01
Total spend to date- $75.73
Ill start with an update on the new landing page. Voluum reported 29 clicks on our ads, and 9 clicks on the new landing page. Not horrible, but not great either. Also, still no conversions to report.
The campaign has been running for a few days now, and has collected enough data that I was able to start disabling low preforming placements and ads.
First, the ads. I pulled stats for all time and used the ads CTR to determine the winners and losers. Out of 9 ads, I kept 5 because they collectively had double the CTR compared to the other half (4). Even though not statistically relevant, it was enough of a spread that we could at least pick a smaller set of ads to test further.
Now for the placements. Again, I pulled stats for all time and disabled placements that met the following criteria- a total spend of $3 with no clicks or a total spend of $6 (double the payout) with clicks, but had no conversions. This spend was measured across both the original campaign and the newer placements campaign.
I ended up disabling 9 placements, which left me with only 24 running. Of those 9 disabled, 2 were bringing in the majority of the clicks, so consequently my campaign slowed to a crawl.
I went back to Voluum to reanalyze the data to help determine my next move. I noticed that the only clicks (from our landing page) to the offer page were coming from blackberry phone users. I decided to turn back on campaign #2, but modify the campaign to only target blackberry phones. Also, I left it as run of network, so I could harvest new placements, while at the same time slow the spend of that campaign.
At the end of the day Campaign #2 had 20 clicks (.23% CTR) and 84% for Impressions Won. The placement campaign had 17 clicks (.19% CTR) and Impressions Won were at 89%.
From:J-SiN Date: 06-21-19
[To give you a bit of background I buy about 20 million clicks a month on mobile (probably be 22 this month)] I don’t say that to brag or boast, just to let you know I speak from experience (David would vouch for me that I’m a pretty humble guy…nothing compared to him though)
The trick to mobile in general is to find publisher that are sending fraudulent or accidental clicks and weeding them out…as you hinted to before. It might be best to target one big pub at a time so you don’t have 12 clicks from 4,000 crappy ones eating your budget up.
Another big secret that we’ve been fighting for years is that a lot of "UK" traffic is fraudulent traffic from India that is spoofing a UK IP to defraud the networks and you & I down the food-chain.
It is REALLY bad. I’ve seen regularly 30% fraud and often 50%+. There’s not way to know and often the networks don’t either. It looks to be UK. The reason I can tell it is not in the UK is because by the time I have it I have an install on the device and there’s no way for them to fake their locale and network connection once I’m on the device.
Lately it’s gotten so bad that I’m even seeing it in the USA where the dirtbags are faking the IP and using a phone emulator that automatically installs the app.
A super duper (I never say that) BIG tip for mobile is to avoid UK at first. Unless of course you have a reputable publisher with a decent app or mobile website…which is often not available to us buying the sub-prime traffic. The premium traffic often goes to the big boys for a premium cost.
You could run $20 per day campaigns for the next year before you find something that works. i know, I’ve blown $12,000 last week on a premium publisher (Pandora) before I finally had to pull traffic. They are a good publisher with a good name. Just tons and tons of accidental clicks. Did I say TONS?
In this game you’re fishing. Unfortunately, you picked a pond full of scum suckers (no trout here). Move to a different pond would be my suggestion.
If I didn’t know what I do now, my first suggestion would be to try another offer…but knowing what I know about UK mobile I suggest try a different geo.
From:KickassMarketer Date: 06-21-19
I’m not a Sitescout expert, but usually I split a mobile campaigns into 4 sub campaigns, using the targeting combinations below:
Site + wifi
App + wifi
Site + carrier
App + carrier
If you set up just one campaign that includes all targeting possibilities, then usually most of the budget is allocated to high-traffic placements (visited via wifi connections). In the current campaign setup, a (large) part of the budget seems to be allocated to non-mobile optimized sources as well.
I’m pretty sure you’d get more mobile optimized traffic when using 320x50px and 300x50px banners. Actually, I think a 300x250px banner will only get you a limited amount of tablet traffic on mobile optimized sources, whereas it unlocks all mobile traffic on non-mobile optimized sources. However, non-mobile optimized sources usually cause a bad user experience on many mobile devices, decreasing the quality of ad clicks dramatically, so personally I’d stay away from it when testing a mobile campaign.
Re. fraudulent traffic: I’ve seen this issue before (.. traffic from India while exclusively targeting Uk or US). Usually it’s possible to weed out these bad sources by looking up ‘country + placement ID’ reports in your tracker, and then block the suspicious placements accordingly.
From:jorgos Date: 06-22-19
At ad:tech NY I went to a presntation which claimed 30% of mobile impressions are fake and on a remnant platform like Sitescout it’s probably worse, see –
On a practical level, try checking the referrers on all these clicks you’re getting. They may be blanked, but they may reveal that the clicks are coming from junk sites completely different to the ones you are targeting.
From:fallonp Date: 06-22-19
Placement Campaign- $71.54
Mobile Campaign #2- $25.82
Total spend to date- $97.36
Activity on both campaigns was pretty slow. Campaign #2 had 8 clicks (.15% CTR), but was at 96% for ‘Impressions Won’. The placement campaign had 19 clicks (.32%) and ‘Impressions Won’ was at 90%. However, Voluum showed only 6 clicks to the offer page (from our landing page), not good.
Voluum was reporting that all clicks were still only coming from blackberry phones. I decided to make sure the landing page was working correctly and tested it using an online mobile web emulator, as well as on my android phone. It was working and looked right, so what else could it be?
I went back to Above All Offers and re-reviewed the offer. I hadn’t realized before that the offer page was not utilizing a responsive page design. Instead, it was relying on browser detection to determine the format to serve, and I know from experience this can be problematic. Still, no conclusions can be drawn at this time.
As you can see from our total spend to date, we reached our phase 2 budget of $100 and thus phase 2 has now come to an end.
Between both campaigns, there were a total of 217 clicks and 10 clicks to the offer page. This tells me that if I were to continue, and 1 conversion came in after re-funding the campaign, at most I’d see a conversion rate of.46%, and our landing page would convert at 10% max. It would be hard to make this profitable given these ratios.
But, why didn’t this campaign work? Was SiteScout the issue, or was it a bad offer? I’m not sure. However, after receiving advice from fellow forum members (thank you KickassMarketer and jorgos) and watching the Adsimilis Mobile Webinar (https://gossips.best/stm/sho…ighlight=tapit), here’s what I would change if I decide to continue running the offer.
Enable targeting by language in SiteScout – choosing a specific language to target was missed when setting up the campaign
Exclude web app traffic – I would wait to add this later due to the high volume of accidental clicks and junk traffic. Also, there is a rumor that there is bot traffic.
Exclude Wi-Fi traffic – Again, to keep out the junk traffic.
Using different banner sizes – As jorgos pointed out, employing different size banners might open up more targeted traffic opportunities.
Only target phones (no tablets) – This would help keep spend down
Change CPM bid to $1.01 on Mobile Campaign #2 – While this would keep us from bidding on premium inventory, it would also help keep spend down.
From:J-SiN Date: 06-22-19