Home > Adult Traffic (NSFW) > Follow-along Campaigns

---

3rd Adult Dating Campaign. Trying to get profitable... (8)

---

Hey Matuloo!
Thanks for all the advice!
Not just on this thread but read a bunch of your threads and replies on others as well!

In regards to my situation, yeah it is crazy stressful but I'm very stubborn as well if I can't get something to work so I'll probably tough it out.

Testing wise, I'll make sure to test a boatload more. The problem I get in regards to testing is

1) Since this is a low volume geo, I have to wait days before I achieve any statistical significance, so I'm little worried that my funnel can die out by then or am I crazy to think that it'll burn THAT FAST lol?

2) I'm guessing I should always check my stats in a overview basis from all the days I have tested instead of the small stats day by day? Since it's low volume, I'm assuming that 1 day worth of stats will not be significant at all compared to the average? (ex. 10% conversion rate but if I only get 9 clicks to my offer in 1 day with no conversion it's not really an impact since overall CVR is 10%?)

3) Would I compare stats of new rotated banners/landing pages/offers with an existing banner/lp/offer starting from the day I started the new test variations or can I still use the larger data from the original banner/lp/offer to compare with the new variations?

4) TESTING OVERWHELM: I tend to get this noobie overwhelming feeling when I test. Not sure how to organize it. I would rotate a new banner in but keep other banners that need more testing, then I'd put in a new landing page to test, then few days later put in offer variations. I'd get SO CONFUSED with all these different varying stats. (for example, I'd test 4x offerpayout for a banner, Do calculations on calculator for significance, data spread not wide enough, continue testing, add new offer or LP variation, then I'd wait out more another 4x payout for the same banner just for that offer/LP.) I feel like I'm overtesting and getting overwhelmed with new variables/mixed stats coming in. Should I just test 1 part of my funnel at a time? Test banners, then cut losing ones after achieving obvious winner, then start testing new lp/offers, find winning LP/offer variation, then banners again?

Angles:
Been brainstorming new angles. Some I think are good, some are just stupid, some funny.
I guess the breakthrough happens when you achieve a good angle to market to the geo?
I see so many people in adplexity or even manual spying, literally ripping eachother and running same thing.
I guess success happens when you take what's working from there and put your own twist to the angle/ copy/images?

Also what's funny is that the best performing banner right now is literally a completely ripped banner as a control. Makes me pissed off that a ripped banner is doing well and my own twists are shit lol fuck

Thanks again Matuloo
This forum is a godsend!

Will update data on this thread once I achieve something significant in data or a change!

---

Thanks for the kinds words


Funnels can die fast, especially the banners, rest is usually more stable. But don't think that they would die in a day or two. With small GEOs its quite normal to wait for a few days to reach any significance, that's what makes working with them kinda problematic at times.

Yup, with low volume, you need to look at the bigger picture, do not base anything on a bunch of clicks. 9 clicks in a day is nothing, you need to find a way to drive it higher or change the GEO.

You can use the "larger" data to compare things. Just keep in mind that if you made large changes in terms of targeting or other parts of the funnel, it would become irrelevant.

Testing 1 part at a time is always the best approach, however, it's not always possible. It would probably take you months to test anything with that volume ... You need to find a system anyways, otherwise you will get lost in the process. When I work with small GEOs, I find the most traficked spot that exists, or close to it, and test stuff quickly there ... once I have something with some promise, I take it to the smaller spots and give it the time it needs.

That's it exactly, find what works and make it better ... this worked the best for me. You can just copy and hit a good one from time to time, but giving it a personal touch is always better, even if it's just changing color, image or button style.

This is exactly the reason why it's good to start with ripped stuff - you need a benchmark to build on, you need to know what is decent and you need to learn how to make it better

You're welcome, looking forward to your results

---

Quick questions/Update:

Started testing new variations of a banner from the original batch.
Had paused a bunch of banners except 2 from the original ripped stuff.

1) If I notice in the original banner that over the days/weeks, the conversions are dying down and not performing as well as it's used to, even though OVERALL that banner is still in the positive ROI. Is it wise to just pause that original benchmark banner and focus on the variations I made? ex. Banner 1 did 200% ROI for a combination in week 1 but by week 2-3, it's in the loss of profit and ROI is dwindling down.

2) My second original banner that survived the test is starting to pick up very late in the game. I spent 10x the payout for that banner and it over took Banner 1 which I made variations of. Is it wise to now start make variations of banner 2, so I am now would be testing variations of banner 1 and banner 2?

3) I have used the bayesian calculator for pausing and running my banners and it says now a variation of a banner is out performing my original banner 1 and banner 2. However banner 2 is picking up and has no variations. So with given this information above, I would pause Banner 1 since it's dying out and test the variations, and start making variations for banner 2 then do the same for banner 2 once I find a variation that is better?

I am still around -30% roi overall in my campaign, in a typical adult dating campaign. Does a positive ROI consist of a bunch of variations of banners doing positive? I find it hard to believe that around 5-6 banners would ALL be positive everyday. Is it just 1 magical combination that would blow the rest out of the water that will give me most of my profits?

Also if I spent around 1x the payout and I am noticing multiple conversions coming in. Is it safe to assume this banner would be doing well in the future or is it too early?

**It's really weird how some of my banners that has a higher CTR, have no conversions. but this low CTR one (0.10% ctr) is having a good start. and the ONLY thing I changed was the girl picture. besides giving my own touch/new layout/color/CTA to the original ripped banner.

The one big problem I'm getting is that, throughout all this testing I find a **supposedly** a good combinations but it's super overshadowed by all the other testing which is giving me the -30% roi. ON TOP OF THAT, the banner/combination that profits varies DAY TO DAY so I have no idea which would be the winner everyday. which is frustrating on making data decisions on how to make it profitable.

Man is this price for a low volume geo or am I just stupid and don't know how to make data decisions?


P.S. I know there is also a strategy where in the adult lander, if the lead clicks "NO" on the "are you above this age" question, they can be sent to a PPS offer. How effective is this? Is it worth it?

---

Banners in adult can die quite quickly, if you see some past performer losing track and you have a better replacement, it's wise to replace it. Or at least give it higher % of rotation if possible. Keep the good banners at hand though, after some time, you can re-use them.

In some cases, banners take time to show their full performance, that's why it's good to wait for enough data. One of the reasons are day to day fluctuations, banners simply do not perform the same on all days. To answer your question, doesn't matter if it's banner 1 or 2, as long as it shows good results, you should attempt to make variations of it.

Adult is very impulse-driven market, sometimes it's pointless to over-complicate it and try all kinds of scientific methods ... generally speaking, you always want to work with the best performing creatives, but do not throw any decent performer away, you never know when it could come handy. I always try to run with multiple banners, even if some are showing a bit poorer performance. In your case, it would make the most sense to start making variations of banner 2 now, since it showed more promise ... but again, keep banner 1 and it's variations at hand.

Yup, as I said in the previous answer, I always prefer to run SEVERAL banners and even LPs at the same time. This way, I'm more immune to these daily fluctuations - it's hard to have 5-6 or more banners all profitable EVERY day, but it's more likely to reach consistent profits with some variety, compared to running just 1 single banner.

Multiple conversions for 1x payout worth of traffic is always a good sign, BUT keep in mind conversions happen randomly sometimes ... in other words, some users will sign up no matter what banner or LP they've seen, it simply happens. If you base your results on such users, it might be a bad decision ... always wait for at least some level of statistical significance ie... multiple conversions.

This is very typical, higher CTR banners are usually misleading in one way or another, so they convert worse ... on another note, the female on the banner is one of the most important factors in adult dating, if not the most important of them all.

A good rule is to spend only a minor part on testing, exactly because too many tests will eat up all your profits.

Yup, low volume GEOs are known for this, will lower volumes, everything is less stable and more random stuff happens. Do not rush it, give it the time it needs. When I work in small GEOs, sometimes it literally takes days in order to be able to optimize anything. On top of that, performance changes in adult from one day to the next, even during the day .. then the weekends ... you need to learn to factor this into your optimization process.

I never had any special success with this, I tried to send these clicks to a different offers, even a non adult offer, sell the traffic to someone else ... it never really made a difference compared to simply forcing the visit to the original offer anyways.

---

Definitely an improvement, I see some green combinations in there. One question though, is this mobile or desktop traffic? If it's desktop, the LP's have a bit low CTR, you could definitely improve that.

RON on TJ is solid sometimes, but there are also some footer and other low visibility spots that can deliver tons of volume but almost no CTRs. Quality wise, the traffic is ok, but sometimes it's just too expensive traffic due to the low CTR. I would try more spots on TJ if I was you, ntvs, footer, in-videos ... there are multiple and you can also run on several sites at once.

Other english speaking GEOs are all more competitive than the one you run now, maybe Canada might work. "Everyone" targets EN geos, so expect high competition. I'm not sure this is the best idea.

Yup, based on the latest data, you are only about 20% negative, that's definitely something that can be optimized into green. Payout bump, better offer, better LP ... this all can help you get there.

Adult is a lot about testing banners/LPs ... so it's good you're testing new variations. Try to move up the CTR a bit, it's possible to go to 0.2 - 0.25 at TJ without being too misleading, so there is still space for improvement here.

Do it if you really have to, I know it hurts to wait days for data. But my preferred method would be to stay in the GEO and scale to other placements or sources, not changing GEOs now. If you want to change other GEOs, the best pick would be Canada, UK is very competitive and AU has very high bids and not many good offers.

---

I don't daypart ... almost never ... there are a few exceptions, but it's generally something I don't wanna do. It's true that I lose $ during some hours in some campaigns, but then again, I profit at the same time in others. I believe that when there are people interested enough to click my banners, I should be able to conver them.

But your feelings are right, the CTR and also CVR is different on some days... weekends are usually stronger in adult.

Improving ripped stuff is a great decisions, that's what everyone should be doing ... rip for inspiration, but give it your own personal touch. Surfers need variety, so let's give it to them

Test more FEMALES !!! That's the only advice I can give you in order to get better CTR and overall results ... and play with stronf headlines, that's the second most important factor.

You're doing well, don't stop now!

Matej.

---

I agree with Matuloo regarding dayparting.

I dont think you have enough data to do this, and since you're in a small geo it's even more counterproductive to do it.

You're off to a good start though. Best of luck to you

---

When you see a large drop like this, it's usually one of the following :

- banner burnout, people are still clicking it, but there is less buyers in the mix. This usually goes hand in hand with lower LP CTR. I have an article on this here : http://stmforum.com/forum/showthread...-down-and-down

- you've been outbid and hence getting a bit lower quality traffic

- the offer changed something, try to ask your AM whether there are more urls available for the offer and test them

- the advertiser or network is scrubbing your leads - sometimes they do this when the quality of your leads isn't that good , in order to keep your payouts high ... sucks, but that's the way it is sometimes. You can test different offers or another network.

One way or another, the only way to stabilize this is to find MORE traffic - how many spots/sites are you buying from now? IRE is a small country, but it's still 6 or 7 millions of people, there is definitely more traffic to get than couple leads a day.

---

Home > Adult Traffic (NSFW) > Follow-along Campaigns