Traffic fallen off a cliff on first profitable campaign, please help figure out why! (10)
Traffic fallen off a cliff on first profitable campaign, please help figure out why!
The campaign:
Targeting Brazil, mobile, site. Targeting 1 specific site in Brazil. Targeting Claro carrier as that was vastly outperforming others, and had no traffic issues before. All of this is on Decisive.
I have been away for 10 days or so. I left this campaign running in the meantime. Up until about the 11th/12th May, it was getting 100k+ bids a day, sometimes a lot more.
After 11th/12th, it has seen a consistent drop in bids. It is now only getting about 3k bids.
What we know:
I have run up individual reports for before and after the 11th May.
2 big things leapt out.
- Before 11th May, there were 23 pages of handsets. After, there were only 12 pages of handsets.
- Before 11th May, most of my traffic was coming from 'undefined' (what it said on the report). After, that traffic wasn't there any more. (I had forgotten that during the week the campaign actually got changed from broader targeting to being set up to target 1 site specifically).
With this information in hand, I have done the following:
[*]Edited the campaign to also have it target 'nopageurlspecified' in sites/campaign set up. This is the closest thing I could find in the list of targetable sites that sounded like it might be whatever 'undefined' is - with all the traffic - on my report.[*]Set another campaign I had paused over my time away running again. This one is a lot more broadly targeted.[/list]
I expected both campaigns' bid amounts to leap up from the get go. Campaign #1 because I had told it to target 'nopageurlspecified' which I assume is 'undefined' on the reports (but I may well be wrong about this), and campaign #2 because before I went away it was getting lots of bids and hasn't been told to target 1 site. Previously it got bids really fast.
So far, the bid amounts for both have not leapt up. They are still crawling up.
So, please help me figure out: why am I suddenly getting a lot less bids than before in Brazil?
This is same with most mobile ad networks. You can take your campaign to diff network.
I have continued to play about with this and have some breakthroughs.
I checked live reports for campaign #1 for a custom date range of 1st-11th May and looked at sites. That live report showed 33 pages of sites. The 'generated' report (i.e. the other sort of report) in Decisive only has 1 page of sites for that date range.
This tells me that there are 32 extra pages of sites in the live report that aren't in the generated report.
It seems a safe assumption, then, that those 32 pages of sites in the live report are what the generated report calls 'undefined', and it has lumped them all into 1 glorious 'undefined' category.
Who knows why. Probably because individually they only get a small amount of traffic, but it all adds up I guess.
Given this info, I removed all site targeting in campaign #1's edit screen on Decisive (but not changing previously blacklisted sites). I am now waiting to see if its bids number shoots up.
IF I AM RIGHT AND THE BIDS WERE ALL PREVIOUSLY COMING FROM THESE 32 PAGES OF SMALL SITES, THEN I HAVE A FEW CHOICES.
- I can keep the campaign as not targeting any specific sites and allow it to target all these small sites. As I get data, I can then blacklist sites that are clearly unprofitable and try to hit a bid number/traffic 'sweet spot' where I am still getting a lot of traffic on a range of sites. The difficult thing here will be making banners + LPs that work across a broad range of sites (ideas welcome!)
- I can set the campaign to target its best performing site. This will put the bids number back to very low as that one site doesn't seem to have much traffic (altohugh in Decisive's planner it shows as having a lot more, but I don't know if that's maybe on the other carrier I'm not currently targeting). The plus here is I can try to make banners that work with that specific site, which might be easier than banners that appeal to 20+ sites.
- I could have multiple small campaigns, each targeting sites that have the largest amounts of traffic individually (but not that much when compared to all traffic available). This will have the benefit of trying to make banners specific to each site, that will work for them. It will have the downside of a lot of extra maintenance, plus smaller/slower campaigns as they have less traffic individually.
Maybe there are other ideas too. Thoughts welcome!
What would you do, given all this traffic stuff?
It should be noted that although there is a little bit of wiggle room, the campaign had been set to target 1 site because that was vastly outperforming any other site (about $0.33 CPA for a $0.51 payout). I guess if I go for targeting a bunch more sites to get traffic, I'll have to hit a sweet spot that combines more traffic with still profitable. (Right?)
I saw similar issues with Brazil traffic on Decisive recently. I had a nice profitable campaign that was targeting a few specific sites/apps on Decisive. I was able to spend $400-$500 per day with these placements. Last week the traffic the traffic completely dried up (no, I was not being outbid). It was suddenly a struggle to even spend $100 per day with the same placements.
Traffic has started to pick up a bit over the last few days but is still not where it was. Does anyone know what causes this?
edit* - also, wtf is up with placements on Decisive's planner that show millions of impressions per day (iTube BR for example) yet no matter what I don't ever get a single bid. It's not that I'm not bidding high enough to win the bid... there are no bids at all. So where the hell are they getting these numbers for their planner stats?
Placements come and go fairly often. I don't have specific numbers, but I remember some countries in the planner changing daily in traffic terms of about 20-50%. Might be your placements just got removed.
Beside that, take the numbers in the planner with a piece of salt. On the frontage of the planner, India shows 145 mio impressions daily right now. At OS, it shows about 60 mio.
Have now been running the two campaigns without site targeting for a few hours. While traffic has picked up a bit, it's not at all what I'd expect.
Starting to look into other traffic sources as stephan suggested - thanks dude (still new so don't have other traffic sources set up yet).
fjk87 - if placements come and go, could I anticipate them coming back? I.e. I'm wondering about leaving these for a few days to see if it picks up (currently feeling rather down about all this: the traffic looks like it's been lacking for about a week now, but these were my best looking campaigns to date so it's all a bit of a downer...)
doppelganger - thanks for input - sorry to hear you're having problems too, but it is comforting to hear it's not just me. Have you noticed these problems with other countries or are you just experiencing this in Brazil?
Email Decisive support and ask if there has been a large turnover in placements.
It could also be that one of the ad networks which provided a lot of BR traffic got removed or stopped delivering for whatever reason.
Decisive plugs into a bunch of exchanges. And since it's a real time bidding platform anyone else who plugs into those exchanges can take traffic away decisive traffic if their bids are high enough.
I think most of your cases will be to do with being outbid. But like Zeno said, email decisive, or PM them on STM.
Hi Guys!
Please do email or pm us.
So for the planner-- country vs OS numbers (or other targetings) we only add them to those categories if we know explicitly that they are a match. If they are not a perfect match we don't add them. So for something like country we have exact numbers. For something like a specific device if it doesn't match the categories (like SAMSUNG-SGH-I537) perfectly we don't put it under another device name.
Does that make sense?
Also at Doppleganger, I checked out the placement you mentioned. It seems to be sending pretty stable traffic. Please do let us know if it drops off precipitously, but as mentioned above some fluctuation is normal.
looks like average winning bid prices on claro in brazil started sloping up pretty heavily around the 11th on the main supply of that inventory (may 11th avg winning bid on claro brazil was 0.17cpm and yesterday it peaked at about 1.69cpm). If this was a cpc campaign there might be some logic in the decisive bidder throttling back bids if it felt your bid price wasn't high enough to compete. I have no idea how their system works but some bidders will look at bids vs floors or historic prices to figure if they can make money and if not they wont bid to save resources.
itube new for BR sends around 8.5 million bid requests out a day across all carriers+wifi and has been steady out of the supply source as Andrea mentioned. It normally averages at around 0.15-0.20 winning bid price but it looks like between may 8th-10th it spiked and was at a range of 0.60-1.50. If you ran into your situations on those days it may explain it but the no bids would indicate something else is going on.
Home > Paid Traffic Sources > Mobile