Losing my Mobile Virginity (32)

I've just started trying to get to grips with mobile, and as I've had a Jumptap account for a few weeks and done nothing with it, I thought I'd best pop my mobile cherry.

Let's set the scene:

Traffic: Jumptap
Offer: Free iPhone4 offer on Neverblue
Country: UK
Tracking: Using CPV Lab

After a couple of rejections from Jumptap (daily budget too low, no more than 3 cycles on animated gifs), I finally got some creatives approved and turned on the traffic this morning (UK time).

I'd setup the tracking using CPV Lab, so I could use c1,c2,c3 variables for the Jumptap tracking parameters. I set this up as shown below:



That seems to be working okay, as I can see the campaign/handset/network stats in my CPVLab reports:



Now, the major problem I'm having (well, apart from no conversions yet and not having figured out conversion tracking), is a major click discrepancy between JT, CPVLab, and Neverblue.

Here are my JT stats at the moment:



Seems to be fairly good CTR wise (and is especially strong in certain categories when I drill down in the reports).

Here's a snapshot of stats in CPV Lab:



Mmmm.... 154 clicks in JT, but only 93 showing in CPVLab.

But wait - it gets worse!



WTFFF!!! Only 25 clicks have made it through to Neverblue!!? Out of 154 clicks in JT!


Anyone got any suggestions on what I'm doing wrong? I know we all get click discrepancies, but this is taking the piss.

Is it because I'm using CPV Lab? Should I just put the direct Neverblue link into Jumptap? But if I do this, how can I track my different JT parameters, because NB only support one subid right now?

Heeeelp!

Direct link, you can use commas and then break out the subids with excel, or just use NB's advanced reporting

NB's reporting is delayed a little

Cheers Andy.

So my Neverblue URL would look something like this?

http://qdmil.com/click/?s=123456&c=123456&subid=JT_CAMPAIGN,JT_ADBUNDLE,JT_HANDSET,JT_OPERATOR

Yes exactly like that. A server side redirect is another hop but a 301 is the fastest way of doing it.

Thanks guys. I've copied the ad creatives and set it up with a 301 redirect as suggested. I've appended 'subid=jt-campaign,jt-adbundle,jt-handset,jt-operator' to my Neverblue URL inside the php script for the 301 redirect, and selected to pass the tracking parameters from the Jumptap interface. Hopefully that will work.

Just waiting for the new ad to be approved now.

Couple of things. JumpTap isn't very strong for international inventory try Admoda perhaps.

Another thing is the click discrepancies they are going to happen but can be reduced based on what you're targeting. For example I find targeting apps that I get more discrepancy then web or feature more than smart phones.

Lastly I'm not sure what you're running but I'm guessing dating because thats what everyone seems to think is the way to go. Just remember if its the most obvious thing to run then obviously theres tons of people already running it.

The offer is a 'Free iPhone4' one on Neverblue. Payout is $6.50.

I reached out to Jumptap since my last post, and I've just been assigned a rep, so will try and speak with her later. Interesting what you say about their international inventory. That's one of the things I want to talk to her about.

Good info so far keep us posted.....

and I would def eliminate cpvlab...due mainly to the fact it relies on cookies to work right....and alot of phones don't support that well....

Originally Posted by polarbacon

Good info so far keep us posted.....

and I would def eliminate cpvlab...due mainly to the fact it relies on cookies to work right....and alot of phones don't support that well....

I've tested a p202 and cpvlab a number of times on a few different hosts and I would eliminate cpvlab. The new p202 does doe better but still has some issues. Either look into hasoffers or directlink.

hasoffers or diretlink

p202 and cpvlabs ARE NOT meant for mobile for many reason that ANGRY RUSSIAN has covered and myself in posts about mobile.

if your tracking is off... then what are you measuring ?!

You can also sign up to Beta MobAff Tracker and use that when it opens

http://www.mobafftracker.com

Alex, your tracker looks awesome. I signed up a while ago.

Can you tell when will it be available?

Originally Posted by ipomea

Alex, your tracker looks awesome. I signed up a while ago.

Can you tell when will it be available?

I wish I had a date. I'm working with Tyler of Beyond Hosting now to get our servers up to par to handle multiple users and millions of clicks as well as spikes in clicks without overloading the server. Also, I'm not happy with reporting as its slow and clunky.

I'm shooting for February.

Will be following this thread, keep it up

Will be awaiting the new tracker.

Alex, have you thought in hosting it in the amazon cloud? It may solve the speed problems.

I don't believe in clouds. There is no rush things will get done and get done right.

My two cents

Switch out the JT_CAMPAIGN variable, and instead use JT_PUBLISHER

You can typically tell from Jumptap reporting (or your subid) which campaign is getting the clicks/conversions - but one of the biggest keys to success on Jumptap is being able to block or double up on specific publishers.

Thanks.

Progress is slow right now (as in zero progress), cos I'm tied up with other Christmas related things. Will get it running again soon, but gonna switch the offer to a US based one, based on a chat with my JT rep.

I prefer JT_SITE to JT_PUBLISHER because a publisher could be something like 1 exchange that has tons of inventory. By blocking it you can be severely limiting your potential volume when in fact maybe it was only 1 site within the bucket of publisher that was killing off performance.

re: Russian

We just found that out to be true actually - after several wishy washy explanations from our rep, we figured out we're getting exchanges and multiple many-to-many relationships between pub and site.

We pull in both so it's not an issue for us, not sure if CPV Lab limits you to three tokens

Originally Posted by bmcadvertising

re: Russian
We pull in both so it's not an issue for us, not sure if CPV Lab limits you to three tokens

This has been slow progress, what with Christmas coming up, but I've managed to do a little more testing over the past few days.

In summary:

- I dropped the UK iPhone offer, as my JT rep suggested targeting the US would be better, as there isn't much inventory in the UK on Jumptap. I'm currently testing a US 'free' iPad offer on Neverblue. Payout is $1.10.

- After many hours of hair pulling, I've finally got my tracking sorted. I'm using HasOffers, and I'm hugely indebted to Leber026 for his extremely useful tutorials he's posted regarding tracking for mobile. Conversions are showing up in HO, Neverblue, and Jumptap.

- Nothing profitable yet, and my rep is saying I need to spend at least $200-300 before beginning optimising a campaign. She must be on commission.


Here's my data from yesterday. I only ran it for a couple of hours.



I separated App and Web traffic and ran in the same campaign? Should I have done this, or should I just focus on Web traffic for now?

Analysing my data in Excel, Sprint had the greatest conversion rate, but even that wasn't brilliant.

There's a really useful tutorial on pivot tables here. Can't remember if it was already posted on this forum or not.

Pivot Tables tutorial

I think moving forward, I'm gonna scrap testing the App traffic for now. Sure there's a lot of inventory, but it doesn't seem to convert well at all.

I'll re-launch a new campaign for this offer:

- only Web traffic
- Android
- Sprint, Verizon, AT&T, T-Mobile
- distribute spend evenly (I've been delivering it as fast as possible up until now)
- keep bids at £0.07 ($0.11)
- 300x50 and 216x36 ad sizes (separated out into their own ad bundles)

and we'll see what data I get back from that. I'm happy now my tracking is sorted, but I need to hone down a process to get more methodical with my testing.

I'll also pick a couple more offers and run some tests using the same criteria above.

I agree with your new approach. Don't target app traffic to start as it's going to be hell trying to make it profitable given the number of misclicks and such a low payout. Speaking of low payout, I don't think you're going to get the offer profitable anyway because it pays $1.10 and you're bidding $.11.

Testing is your best friend so I could be completely wrong. Goodluck!

Looking good nick!!

Hows the click discrepancy looking? Would be good to see some stats on how it has improved/

Originally Posted by tijn

Looking good nick!!

Hows the click discrepancy looking? Would be good to see some stats on how it has improved/

Keep it up Nick you're almost there.

Remember that the price floor is NOT .10 in JumpTap... hit up your rep and you can set your bids to whatever you want!

How does bidding lower effect the traffic on Jumptap? If most are bidding at .10 and your rep lowers the bid to .5 do you still get much traffic? Or do they have enough to go around?

I haven't given up on this... just had an extended Christmas break whilst my kids were off school.

So, I created 6 banners, in both 300x50 and 216x36 sizes, and ran some more traffic yesterday.

To recap, here are the criteria I am using:

- web traffic only (wifi also blocked)
- Android
- Sprint, Verizon, AT&T, T-Mobile
- distribute spend evenly (campaign budget set at £40)
- bidding at £0.07

Results for the end of yesterday:

spend - £40
revenue - $27.50 (about £18)

Here are my stats in Jumptap. Only 19 of my 25 conversions registered for some reason.



Now, based on this, my gut reaction is to pause all the ads except B03 & B04 (in 300x50 size), because they are getting most of the traffic, but also performing best in terms of CVR.

Looking at my data in HasOffers, I have the following:

- 18 of 25 conversions were from Sprint. 4 Verizon, 2 AT&T, 1 T-Mobile.
- 19 of 25 conversions came from two sites (split fairly evenly).

So, what's the best thing to do now?

1. Just let it run for another day (capped at £40) to get more data.
2. Pause all the ads except the two mentioned above, and let it run.
3. Kill all carriers except Sprint.
4. Something else....?

I'm pretty sure my JT rep will say it's too early to restrict any sites.

I'm going to set up another camp with another offer, and will report back. Sooner, this time.