Home > Paid Traffic Sources > Adult Traffic (NSFW)

---

Any tips on being more consistent with banners? (13)

---

So I'm finding that my earnings are really inconsistent with my campaigns, and I feel like banners are the issue for me.

I can get a campaign profitable by finding a winning banner(s), and then I'll ride it out for a few days, but then after the banner dies I piss away all the profits I just earned trying to find winning banners again.

I've tried creating variations of the winning banner (ie similar images, same angle, designs etc) but it just feels like I'm throwing shit at a wall hoping something will stick. Results are inconsistent at best.

For example on a mobile campaign I was running, I found a banner that was running 50%+ ROI for several days, but after it dies out, I can't revive the campaign. I've tried using the same image on other angles, used the angle and tested new images etc. but nothing seems to work and Im running today at -30% lol. I know eventually I'll find a winner again but the whole process will repeat and I'll be back at square one again.

With a lot of my campaigns it seems like I'll crush it for a few days, then the few days after that I'll piss away the profits, and then I'll start crushing again etc. It's like an endless cycle lol

Any tips on getting more consistent with finding winning banners?

---

The only way to combat this, that I could come up with, was to always aim for running multiple banners and continuously testing at least some new ones along with the winning ones. Whenever I land a campaign that runs on a single banner, I end up exactly as you described ... it runs for a while (sometimes quite long, even months in rare cases), then dies and I have to find new ones. But it doesn't take all my previous profit away, just a part of it, so you should probably work on speeding up the process of finding new banners.

How many are you testing when finding new ones? Lately, I've been doing this quite aggressively and testing lots of banners across different campaigns at once.

---

Thanks for the input Matuloo.

For me I run sets of 3-5 banners at a time hoping that at lease 1 or 2 will have decent ROI. This can be a hit or miss because sometimes all the banners bomb.

I guess for me I'm having a hard time pinning down the "why" of something works. I need a way to be more consistent instead of just throwing a bunch of banners up and praying something is profitable.

---

I test more at once, 5 in a campaign at least, sometimes 10 ... and I test in multiple campaigns at once. So sometimes I have like 25-30 banners running as tests at once easily. But when running such bigger tests, I cut pretty aggressively, if it doesn't perform at least decently straight of the bat, I just pause them and add more.

---

Testing more banners at a time and casting a wider net is something I was thinking about. I guess my concern was it would cost more to reach statistical significance?

Or are you banking on the fact that in that 10 banner set you'll at least have 1-2 decent performers that you can ride out for a while after that?

---

I'm sure you already noticed, as you've tested 100s of banners by now, that the best performers pretty much perform well immediately. So that's what I'm after with these larger tests, I simply cast the net wide as you said, and look for those that stand out from the start pretty much. In these cases I ignore the significance so to speak. Maybe it's not the most scientifically correct approach, but it speeds up the process a lot

---

Yeah the best performers usually perform well pretty quick.

Okay I will test this new strategy now, thanks again for the advice! lol

---

You're welcome, let me know whether it worked for you.

---

This!

I know I'm always touting the importance of waiting for statistical significance - but banners burn out faster than you can reach stat sig sometimes.

Also: Have you tried taking your best winners from months ago and giving them another try?

What it boils down to, really, is just this: The speed at which you're testing new banners is not keeping up with the burn rate. The more banners you test, the more gems you'll find. Basically keep testing bunches and bunches of banners, and just keep cutting the worst ones. If still not profitable, either move your bar up for the "worst banners" (e.g. if you were cutting at 3x payout without conversion, cut at 2x payout; or if you were cutting at <-70% ROI after 5x payout spent, cut at <-50% ROI or <-30% ROI), or create and test more banners, or both.

And of course, how effective your angles+images+text are plays a big role as well. But based on my experience with adult, it's not always easy to tell what angles would work best - like you mentioned it's hard to figure out what's making the good banners work. So throwing-shit-on-the-wall-to-see-what-sticks may be the way to go.

And if you're seeing some sort of pattern, such as banners of a specific angle seem to never perform well, then stop generating further banners using that angle.

Also: Are you niching down in terms of targeting? That can be a good way to laser-target your ads to your audience. And there are so many niches in adult. The divide and conquer approach may take considerably more effort, but there's less competition, camps last longer, and conversion rates are better.



Amy

---

This is also a great tip! I'm doing it all the time, quite often the winners from months ago can perform well again.

---

From my small experience, I can confirm that it has worked for me several times

---

Works for many sources, not just adult I love the old campaign vault

---

This depends on so many factors..... like... where are you testing, what cap u r using, what niche is it, what bid, how much volume there is, how many banners u r testing, did u set the order of appearance etc etc etc...... after all "winning banner" can mean different thing to different buyers.

Sent from my SM-N9600 using STM Forums mobile app

---

Home > Paid Traffic Sources > Adult Traffic (NSFW)