Home > Paid Traffic Sources > Adult Traffic (NSFW)

Question regarding banner burnout and when to cut banners (10)

01-05-2017 10:22 PM #1 jabong82 (Member)
Question regarding banner burnout and when to cut banners

Hey guys so I think a problem I'm having is not recognizing when I should be uploading new banners. I think this is probably eating in to my profits as I don't know when I should cut/upload new banners.

So lets take a real life example:

Day 1-5:

Banner A: +25% ROI

Day 6:

Banner A: +5% ROI


So in the above scenario we have a banner that has been running well for 5 days but on day 6 it takes a nose-dive to about break even / slightly profitable. At this point am I cutting this banner? Or do we chalk it up to maybe a bad day and let it ride to a day 7?

I guess my question is do we look at banner performance on a "day by day" basis, so even if a banner has been a big winner for several days, as soon as it drops we are already tossing it?

Thank you in advance.

01-06-2017 02:55 AM #2 vortex (Senior Moderator)

Instead of trying to identify when exactly you should replace an old banner with a new one, I would suggest to rotate in new banners on a regular basis and cutting banners based on performance.

The spreadsheet calculator here is actually perfect for this kind of thing:

http://stmforum.com/forum/showthread...Banners-Part-2

You can use banner stats from the most recent couple/few days - plug them into the spreadsheet to see which ones are likely to remain profitable, and cut the ones that aren't.

So you can keep creating and rotating batches of banners, and only keep the ones that are likely to remain profitable. That way, you can minimize the effects of decreased banner performance on your campaign overall.

And by keeping more banners in rotation, blindness will set in less quickly as well.

Also - blindness doesn't usually set in suddenly over the course of one day after being a big winner for several days. If you see a nose-dive in performance, it's likely caused by other reasons.

What I would do is take that winner and create lots of variations on it - same angle but different headlines, different colors/fonts, different layouts, different image, etc.

Another thing you can do to combat blindness, is to rotate two sets of banners over alternate days. Any way you choose to rotate them, it's better to have more banners in rotation than just one.



Amy

01-06-2017 05:22 AM #3 jabong82 (Member)

Quote Originally Posted by vortex View Post
Instead of trying to identify when exactly you should replace an old banner with a new one, I would suggest to rotate in new banners on a regular basis and cutting banners based on performance.

The spreadsheet calculator here is actually perfect for this kind of thing:

http://stmforum.com/forum/showthread...Banners-Part-2

You can use banner stats from the most recent couple/few days - plug them into the spreadsheet to see which ones are likely to remain profitable, and cut the ones that aren't.

So you can keep creating and rotating batches of banners, and only keep the ones that are likely to remain profitable. That way, you can minimize the effects of decreased banner performance on your campaign overall.

And by keeping more banners in rotation, blindness will set in less quickly as well.

Also - blindness doesn't usually set in suddenly over the course of one day after being a big winner for several days. If you see a nose-dive in performance, it's likely caused by other reasons.

What I would do is take that winner and create lots of variations on it - same angle but different headlines, different colors/fonts, different layouts, different image, etc.

Another thing you can do to combat blindness, is to rotate two sets of banners over alternate days. Any way you choose to rotate them, it's better to have more banners in rotation than just one.



Amy
Hello Amy thanks for your response.

Yeah I think the strategy I've been using was wrong.

So when you say keep uploading banners, are you talkin about on like a daily basis? Say for example I'm on a large tube site (ie pornhub, redtube, youporn etc), I launch 5 banners and let it run for a day. Next day I see 2 are profitable and the other 3 are duds. So at this point do I stop the 3 losers and upload another 3 banners to replace the ones I stopped? (or some similar strategy)

Curious to your thoughts. Thank you.

01-06-2017 10:36 AM #4 caurmen (Administrator)

As soon as you see that a banner is statistically significantly not generating a positive ROI (on an otherwise optimised campaign) you should kill it. No point in killing it before statistical significance, and no point in leaving it running after stat.sig either.

01-06-2017 12:07 PM #5 jabong82 (Member)

Quote Originally Posted by caurmen View Post
As soon as you see that a banner is statistically significantly not generating a positive ROI (on an otherwise optimised campaign) you should kill it. No point in killing it before statistical significance, and no point in leaving it running after stat.sig either.
Thanks Caurmen. I know this is a "how long is a piece of string" type of question, but when do you consider a banner statistically significant?

Also do you have a certain set time you do a banner check (say every 500 clicks etc)?

Thank you in advance.

01-06-2017 12:23 PM #6 caurmen (Administrator)

It's not a "how long is a piece of string" question at all - in fact there are very precise formulae you can use to tell exactly when it's hit statistical significance.

Vortex's spreadsheet above will calculate it for you - http://stmforum.com/forum/showthread...Banners-Part-2 - or you can read about how to do that yourself at http://stmforum.com/forum/showthread...ats-Calculator

As for how often to test - that depends a lot more on how fast you're spending. As a rough rule, I'd check about the time a banner hits 2x payout spend, and every 1-2x payout therafter.

01-06-2017 01:03 PM #7 jabong82 (Member)

Quote Originally Posted by caurmen View Post
It's not a "how long is a piece of string" question at all - in fact there are very precise formulae you can use to tell exactly when it's hit statistical significance.

Vortex's spreadsheet above will calculate it for you - http://stmforum.com/forum/showthread...Banners-Part-2 - or you can read about how to do that yourself at http://stmforum.com/forum/showthread...ats-Calculator

As for how often to test - that depends a lot more on how fast you're spending. As a rough rule, I'd check about the time a banner hits 2x payout spend, and every 1-2x payout therafter.
Thank you I will definitely checkout and tinker with these spreadsheets!

Appreciate all the help!

01-06-2017 01:26 PM #8 vortex (Senior Moderator)

Quote Originally Posted by jabong82 View Post
So when you say keep uploading banners, are you talkin about on like a daily basis? Say for example I'm on a large tube site (ie pornhub, redtube, youporn etc), I launch 5 banners and let it run for a day. Next day I see 2 are profitable and the other 3 are duds. So at this point do I stop the 3 losers and upload another 3 banners to replace the ones I stopped? (or some similar strategy)
Yup this would be as good a strategy as any! As to how often to upload new banners, that would depend on how much volume you're running to those banners, and from how many sites. Obviously the more volume you're running from a single target site, the faster a banner will reach burnout. And the fewer banners you have in rotation, the faster they'll reach burnout as a group. So you can play with the "5" in your example and try different numbers.

Try to test different angles once in a while, identify the best, then write a ton of headlines and find a ton of corresponding images for each good angle, and produce a shit ton of headline+image combinations from them. Then look for the good combinations and produce variations on font style/color, background color, banner layout, subtext, etc. Having to constantly produce new banners is definitely one of the most labor-intensive parts of running adult.

Another thing to try, would be to re-introduce some of the old winners from a few months back. Depending on the target site's visitor turnover and other factors, you can sometimes get more life out of them.




Amy

01-09-2017 08:50 PM #9 matuloo (Legendary Moderator)

Let me chime in too since you mentioned adult sources

If possible, always test new banners, just don't dedicate all the traffic to testing in case you have a profitable set running already - keep majority of the traffic to turn profits, test with smaller %.

You can also use a strategy where you have a go-to placement for testing purposes - so anytime you need to test a new batch, take it there and scale the best ones to the other placements you are using. Having such a testing ground will help you test faster, because you will get to know that placements so good that you will know what to expect from it when you hit a good banner. I'm personally doing this very often.

One more tip for you : in adult, some days simply suck. If you happen to test new creatives on such days, they will fail, no matter what you do. So in case you see a whole batch of banners, or a large part of it, perform decently but still loosing money ... give it a few days and retest them.

You need to develop your own banner testing strategy so to speak, do not try to find exact numbers or a method to go by 100% - your goal is to always have something usable at hand, ready to replace the old dying ones, but at the same time keep the good banners running and making profits - balance is the key

01-13-2017 04:38 AM #10 johnaff (AMC Alumnus)

How big is your audience you're targeting? Is there any metric that the site owner gives you in regards to how many monthly unique users they receive?

Home > Paid Traffic Sources > Adult Traffic (NSFW)