Testing banners

[andponomarev]

Hi there,

Guys, how do you test your banners? The thing is I'm running multiple traffic sources, and each of them has 10,000 of placements. And I would need to make sure that my banner is OK, and it's a specific placement that doesn't work. In other words, if I know that my banner is good, and qualitative, I can only think about the traffic quality of the specific placement.

How do you (if ever) deal with that? I used to test my banners on Traffic Junky tubes. I consider that their traffic is the most qualitative (and most expensive). What do you think about it?

 

[andponomarev]

Also: how much do you spend per one placement before culling out? 0.5 x payout? 1x? 2x? I prefer to spend 1x, and I cull it if I don't have a conversion. What's your strategy?

 

[newlxx]

I also like TJ for testing since their traffic is of great quality

3x payout for testing before I cut

 

[andponomarev]

Thanks for your reply. A few questions:

a) Did you see any difference on traffic quality of TJ sites? I mean, RT/YP/PH?
b) You're spending 3x payout even when running RON? I'm not talking about TJ, but RON on other big networks. Don't you think it's too expensive? I mean, do you really think it's worth of testing 3 x payout instead of 1x ?

Thanks in advance for your replies.

 

[Martin smith]

Like your discussion

 

[johnny_r]

1x payout might be not enough for a real test, in my opinion. What offers do you promote and what's the payout?

 

[andponomarev]

Erhm. Well, honestly, I'm getting about 15K impressions for the 1-1.5x offer payout. It's Spain, desktop, and $2.00. In my opinion, that's enough to estimate at least the CTR of the banner that would be low on a bad placement, and high on a good one.

 

[newlxx]

Erhm. Well, honestly, I'm getting about 15K impressions for the 1-1.5x offer payout. It's Spain, desktop, and $2.00. In my opinion, that's enough to estimate at least the CTR of the banner that would be low on a bad placement, and high on a good one.

I would test 1x payout only for higher paying offers (~$10)
Running RON on these traffic sources will need a huge budget.

Instead if I were you, I would pick one placement on premium site and start testing there to limit variables and focus one ad spot - even spying will be easier

 

[TrafficHunt]

Hey,
We have both exclusive placements and RON, hit me up if interested.

 

[lau]

i think it's depend on the type of offer and the payout, so many offers are worthy to spend much money to test,you need to find out the better banners and lps and sites.
i usually test 2x

 

[andponomarev]

I would test 1x payout only for higher paying offers (~$10)

Why? I can't see any significant difference between low and high payout offers. There's always a correlation between traffic costs and the payout. And advertiser could be greedy, and set a low payout while the needed traffic is more expensive, but that's the offer quality issue. In other words, you shouldn't run such an offer at all since the very beginning.

Running RON on these traffic sources will need a huge budget.
Instead if I were you, I would pick one placement on premium site and start testing there to limit variables and focus one ad spot - even spying will be easier

Easier said than done. Do you know premium how quality sites I could use for such testing? For example, I've tested my new banners on Traffic Junkie tubes. I've got average CTR of say 0.20. And I had to pay $0.20 CPM. Are these tubes of high quality? Guess so. Then I found a file hosting website with 1M impressions daily. The CTR raised to 0.70, and CPM was as low as $0.10.

So it's not that simple

 

[newlxx]

Why? I can't see any significant difference between low and high payout offers. There's always a correlation between traffic costs and the payout. And advertiser could be greedy, and set a low payout while the needed traffic is more expensive, but that's the offer quality issue. In other words, you shouldn't run such an offer at all since the very beginning.



Easier said than done. Do you know premium how quality sites I could use for such testing? For example, I've tested my new banners on Traffic Junkie tubes. I've got average CTR of say 0.20. And I had to pay $0.20 CPM. Are these tubes of high quality? Guess so. Then I found a file hosting website with 1M impressions daily. The CTR raised to 0.70, and CPM was as low as $0.10.

So it's not that simple

By testing lower paying offers you can get statistically significant data.
Higher payout offers require larger budget.

I would promote offers that I know others have been promoting successfully first - you can find those out by talking to your AM or spying.

Nothing is simple if u use CTR as measure of quality of traffic

 

[andponomarev]

By testing lower paying offers you can get statistically significant data.
Higher payout offers require larger budget.

Sure as hell, but... Did I say something against that? I just wrote above that there's a correlation between the payout and traffic cost. So you need to spend 1-2x Payout in either case (both HIGH and LOW payout offers). It's just CHEAPER to test low payout offers, but you would also get LESS revenue comparing to the high payout offers. It's quite obvious, but I'm explaining this as I didn't write anything against those facts. I was just indicating that correlation is more or less the same (1-2x), it's just X (payout) that is different.

I would promote offers that I know others have been promoting successfully first - you can find those out by talking to your AM or spying.

That's what I'm doing right now Again - nothing against this.

Nothing is simple if u use CTR as measure of quality of traffic

Guess I didn't explain that well. I don't use CTR as a measure of traffic quality of course, it's EPC that I use. And if you refer to my last reply, you would see that it's almost impossible to find a single site to test your creatives.

Here are my stats for your reference:
TJ:
EPC: 0.04; CPC: 0.06
CTR: 0.20%

Premium site:
EPC: 0.06; CPC: 0.02
CTR: 0.50-0.70

But even at the earlier stage of the campaign I can judge that traffic costs are 3 times cheaper for the premium site comparing to the TJ.

 

[Justin Baker]

TrafficHaus is a great ad network for testing new creatives. You have full control over managing your campaigns including uploading new creatives, disabling ads and setting weights (% of traffic each creative receives). You will see CTR for each ad so you can monetize the campaigns to its full potential. If you are interested, ping me for more info. Skype: dirtybaker and/or E-mail: justin(at)traffichaus.com