Home > Mobile > Follow-along Campaigns

Newbie Mobile follow along (58)

03-10-2014 05:54 AM #1 dnguyen (Member)
Newbie Mobile follow along

Hey guys. Im new to affiliate marketing and I finally ran a live campaign last week. I seen how supportive the stm community is (special thanks to caurmen and zeno) when it comes to helping others so I thought I'd give it a shot and post a follow along. I expect to lose money for this campaign and I don't mind! I'm just hoping to learn how this game works.

Dating offer
Payout : $1.75
Traffic : Trafficjunky
Country : Canada (offer is also for united states and uk)
Budget : Im not sure how much I should put into this. Whats the general rule of thumb? I remember I read it somewhere but couldn't find which thread anymore
Tracking : STM tracker

Banners : 300x250
I made 7. Same image/gif with different headlines. Whats better to weed out? Image or headline?

Landing Page - I got 4 landing pages and just like my banners, I got the same images but just different headlines. This is definitely the hardest part of the job for me. Being creative is one thing but to actually execute my idea using html was brutal.

I let my campaign run during the weekend and these are my stats

TrafficJunky
Impressions : 23,705
Clicks : 111
CTR : .464

Prosper
Clicks : 156
Leads : 1
S/U - 0.64% (I'm not sure what this stat means)
Spent - $12.17
ROI - (-85%)

I don't even know where to begin when looking at my data. After 23k impressions should I start cutting out some banners? My best banner has 10 more clicks then my worse banner so I'm thinking its not a big enough difference to cut out. For my landing pages, is the CTR the stat that I look at for how well its doing? If so, what is considered a good ctr and when should I cut a non-performing lander?

I was really hoping to get more money spent over the whole weekend and I didn't know if I should have upped my bid or not. The top bidder was getting more then 75% of the traffic coming in while the second and third (me) was getting just a little over 4%. I was thinking, should i just bid close to what the top bidder was bidding to collect data faster even if it meant I wasn't going to be profitable?

Also would it be smart for me to run the same banners on a different site at the same time? For example, on trafficjunky I had the choice of redtube.com,youporn.com, spankwire.com, etc.. Should I be testing out my same banners for all the sites?

Thanks for the read guys!

03-10-2014 11:07 AM #2 caurmen (Administrator)

Looks solid so far!

How many clicks has your worst banner had? If it's not had any over 3,000 impressions, might be worth cutting. If it has had a few, I'd keep running it for now until you get a sense of what the average CTR is for that placement.

Landing pages: Cut them on conversions, not clickthroughs - see http://stmforum.com/forum/showthread...y-Don-t-Matter

I'd say at the moment you're getting a decent amount of data for the amount of things you're testing. Concentrate on upping your ROI a bit unless you have a large budget to hand, in which case you might as well up your bid to get more info quicker.

I would strongly recommend treating each placement as a single campaign. Feel free to start new campaigns on other placements, but don't try and optimise them all in the same data set - different ads and LPs (and offers) will perform differently on different sites.

03-13-2014 04:08 AM #3 dnguyen (Member)

Alright trafficjunky finally fixed a bug that wouldn't let me see how my banners did individually. There are so many variables when it comes to testing that I got a bit overwhelmed and I didnt know where to start but yesterday after reading through some threads on stm, it seems like the general idea is to find a high ctr for banners first and worry about the landing page after.

Update for today
Clicks -50
Leads - 0
ROI - (-100%)

So I made up 8 banners. 2 different headlines and 4 different images. After letting them hit 3k impressions, I decided to keep 2 of them because they had a ctr of %0.5+. I'll try to do a couple things to them to see if I can get the ctr up.

Goals - Bang out ten banners everyday until I get close to .9% - 1% ctr

03-13-2014 04:36 AM #4 zeno (Administrator)

What range of average lander CTRs are you seeing? Adult can be brutal, you will want to spend much of your initial efforts testing lots of banners and landing pages - you can always do better! While CTR isn't the most important metric, your really want to get a decent % of people from banner to offer to get conversion data.

03-13-2014 05:55 AM #5 dnguyen (Member)

Quote Originally Posted by zeno View Post
What range of average lander CTRs are you seeing? Adult can be brutal, you will want to spend much of your initial efforts testing lots of banners and landing pages - you can always do better! While CTR isn't the most important metric, your really want to get a decent % of people from banner to offer to get conversion data.
Im assuming to find out the average would be to add up all the clicks divided by number of banners? Using that formula I would say the average clicks would be 11 clicks per 3k impressions. I'm not sure how to answer the average ctr part lol I guess it would be about .25% - .27% as half of them hit that range while 2 of my banners hit a little over .5%.

Yeah definitely! This is going to be my first actual set of data that I am going to try to be analyzing. I want to try to create some kind of system that will be efficient for testing for future campaigns so I'm not so lost on where to start.

03-13-2014 06:33 AM #6 zeno (Administrator)

Banner CTR is simply clicks/impressions for each banner.
Lander CTR is (clicks arriving on lander) / (number of people who clicked your outgoing link).

The banner CTR you get at the traffic source, 0.25-0.5% sounds right - though not sure if this is high or low by TrafficJunky standards, likely on the low side.

As for lander CTR, go into the STM p202 tracker and go to analyze > landing pages. You can also see landing page CTR data in Overview > Group Overview.

You'll definitely want to keep an eye on those stats so you have an idea of what's working :-)

03-13-2014 07:02 AM #7 dnguyen (Member)

Quote Originally Posted by zeno View Post
Banner CTR is simply clicks/impressions for each banner.
Lander CTR is (clicks arriving on lander) / (number of people who clicked your outgoing link).

The banner CTR you get at the traffic source, 0.25-0.5% sounds right - though not sure if this is high or low by TrafficJunky standards, likely on the low side.

As for lander CTR, go into the STM p202 tracker and go to analyze > landing pages. You can also see landing page CTR data in Overview > Group Overview.

You'll definitely want to keep an eye on those stats so you have an idea of what's working :-)
Oh nice that group overview stat is a new one I didn't check yet.

I had a banner that actually converted twice after like 7k impressions. What would be the steps you would take with that banner? Would you change something major (image or ad copy) or would you just change the little things (cta, color , borders etc.) first?

03-13-2014 12:14 PM #8 caurmen (Administrator)

Do the two conversions outweigh the cost of the 7k impressions? If so, I'd recommend leaving that banner running just how it is now, and seeing whether it's still profitable at about 15k impressions.

03-13-2014 08:43 PM #9 dnguyen (Member)

Quote Originally Posted by caurmen View Post
Do the two conversions outweigh the cost of the 7k impressions? If so, I'd recommend leaving that banner running just how it is now, and seeing whether it's still profitable at about 15k impressions.
Ah I never thought about that. Alright then looks like im gonna run that banner again.

03-14-2014 09:03 PM #10 dnguyen (Member)

Quick little update!

Impressions : 23k
Clicks : 94
Cost : $8
Leads : 1
ROI : (-78%)

So I tested out a somewhat winning image and headline and decided to play around with the colors. I tested out 5 colors on the ad copy and cta and I actually got a conversion on one of them. CTR wasn't to great on them however.

I also tested out 5 banners with a new image with a new headline i have never tested yet. All of them had the same image and headline but I played around with background color for this set of 5. I didn't get any conversions however these are the highest ctr I got, .65% being the highest. Now I have to test to see what made these banners more appealing.

03-15-2014 04:24 AM #11 caurmen (Administrator)

If you can, push your spend up a bit. That'll help you split-test landers and offers - which you should definitely be doing - faster.

03-15-2014 04:24 AM #12 zeno (Administrator)

What range of CTR are you seeing with your ads? Changing headlines, background colours and so on are very minor changes. Same with landers. It's usually better to test completely different styles of banners initially. Give someone a choice between a rose, royal gala, eve and granny smith apple and they might say... meh, but I don't like apples. Give them the choice between an apple, banana, feijoia and nectarine.... now we're talkin!

03-15-2014 08:37 PM #13 dnguyen (Member)

Quote Originally Posted by caurmen View Post
If you can, push your spend up a bit. That'll help you split-test landers and offers - which you should definitely be doing - faster.
Will do. I was getting about 16% of the total traffic with my bid which was the highest I had. I haven't really tinkered around with my lander yet though. How would you suggest I test them? Should i use my highest ctr banners on em? Or make new banners with them?

03-15-2014 08:39 PM #14 dnguyen (Member)

Quote Originally Posted by zeno View Post
What range of CTR are you seeing with your ads? Changing headlines, background colours and so on are very minor changes. Same with landers. It's usually better to test completely different styles of banners initially. Give someone a choice between a rose, royal gala, eve and granny smith apple and they might say... meh, but I don't like apples. Give them the choice between an apple, banana, feijoia and nectarine.... now we're talkin!
I would say my average would be .35 to .45 for ctr. I thought changing the headline was a big change?! I know what you mean though, back to the drawing board I go. I definitely have to change my lander now.

03-15-2014 11:21 PM #15 zeno (Administrator)

Changing headlines can have a big effect, but it is not considered a big change to a lander/banner design. It pays to test several drastically different angles/designs and eliminate those that bomb > move toward the better performing landers > then split test variations of those (e.g. headlines). While doing that you can also have a % of your traffic going to completely different designs - i.e. send most of the traffic to what works but spare some for continued testing of alternatives.

03-17-2014 05:35 AM #16 dnguyen (Member)

Quote Originally Posted by zeno View Post
Changing headlines can have a big effect, but it is not considered a big change to a lander/banner design. It pays to test several drastically different angles/designs and eliminate those that bomb > move toward the better performing landers > then split test variations of those (e.g. headlines). While doing that you can also have a % of your traffic going to completely different designs - i.e. send most of the traffic to what works but spare some for continued testing of alternatives.
OK after doing a little bit of digging I found a thread from mr.green about angles. So angles are to be done on the lp right? New goal : test different angles!

03-17-2014 06:29 AM #17 zeno (Administrator)

Angle is a holistic term - your adverts, targeting, site placements, landers, and even the offer lander may all be a part of your angle. It is a creative approach, a campaign strategy.

Angles are not something you simply 'deploy' on a creative such as a landing page.

Simply put, your angle should start at your ads, flow through to your lander and ideally be supported by the offer landing page, e.g. ad with Asian babe on it > lander also has Asian women peppered all over it > offer is AsianBeauties or some appropriate lander from a broader offer. Your ad copy would then obviously involve dating Asian women, or why Asian's do it better, blah blah you get the drift?

03-17-2014 07:46 AM #18 dnguyen (Member)

Quote Originally Posted by zeno View Post
Angle is a holistic term - your adverts, targeting, site placements, landers, and even the offer lander may all be a part of your angle. It is a creative approach, a campaign strategy.

Angles are not something you simply 'deploy' on a creative such as a landing page.

Simply put, your angle should start at your ads, flow through to your lander and ideally be supported by the offer landing page, e.g. ad with Asian babe on it > lander also has Asian women peppered all over it > offer is AsianBeauties or some appropriate lander from a broader offer. Your ad copy would then obviously involve dating Asian women, or why Asian's do it better, blah blah you get the drift?
Oh damn obviously lol seriiously don't know why that didn't click in my head till I read it.

Gameplan

Start a mindmap and brainstorm different angles, create at least 5 ads to test with each angle, have three differently designed lp

Quick newbie question, I read in another follow along, and the person who started the thread said he was launching 2-3 campaigns daily. What is a campaign? I'm assuming hes just testing new things on the same offer, correct?
Also should I be finding other offers that are related to the offer I am promoting and split testing them? Or was I supposed to find the same offer on diffferent networks and then split test like that?

03-17-2014 09:10 AM #19 zeno (Administrator)

A campaign is a nebulous term - it doesn't mean anything in particular. Affiliate networks for example might consider an entire offer a campaign - across all geos and traffic sources. A media buyer might consider each placement/site a campaign in itself. In Facebook, campaigns are just packages containing ads that have some budget attached.

So, when someone says they are launching 2-3 campaigns daily, they could mean 2-3 new campaigns on the traffic source but with different targeting or angles. Perhaps with completely different offers as well. They could also mean different angles but heading to the same rotation of landers and offers. It's vague but the general point is that you should always be testing NEW things, whether it be new ads, landers, angles (which influence ad copy/banners/landers/etc.), targeting, demographics, offers and even traffic sources.

Should you be finding other, different but related offers? Absolutely. Split-testing separate offers (regardless of the network they are with) is the best offer split testing you can do. If you know an offer is working, or your angle is suitable only for a particular offer, then split testing between networks is ideal (though not so much for cashflow if funds are limited).

03-20-2014 06:45 AM #20 dnguyen (Member)

After a little bit of procrastination, i got a little update. Split testing 4 different offers now and focusing on angles. I tried to test two angles today and my ctr's were looking decent (for me at least), i was getting .47 - .67 however I wasnt getting any conversions. I later realized I messed up something during tracking because in cake it says I got a lead. I think in prosper i had my link with s1 and in cake i used the s2 token..

After testing ten banners, I took the top 4 and these are my stats.

Spend: $50
Revenue: 0
ROI: (-100%)

Hmm I'm not to sure where to go from here with these set of banners. Im thinking of just changing the picture of the two best banners out of this batch and try to run it again. I definitely have to learn how to read data properly.

Im waiting on approval for another set of banners I made. I got two new different angles. When will I know when an angle is a winner? I read a post from mr.green a little while back and he said something like he would test a new angle until it at least broke even before optimization.

Anyways I feel like I'm doing something wrong here but can't put my finger on it. Tracking should be good after i change my tokens, I'm testing new angles, split-testing different offers and got a couple different lps up. Lps could definitely use some work though. Im so html challenged im thinking of going on fiverr and finding someone to do a page for me.

03-21-2014 07:22 PM #21 dnguyen (Member)

Clicks : 155
Spend : 45
Conversions : 0
ROI: (-100)

Tested two new angles yesterday and got 0 conversions after 44k impressions. I'm going to test a couple new ones today but also I am going to re-visit my old banners that did get me a lead and try to make variations of those.

I haven't touched my lander page at all so that could definitely be a reason why I am not getting conversions. HTML is my worse nightmare but I'll try to mess around with it. I tried to email a couple of people on fiverr to see if I could get the pages created properly but got shot down by both people. I'll send a couple more emails today and hopefully have some luck.

03-22-2014 04:31 AM #22 zeno (Administrator)

Did you change traffic sources or countries or something? Your click prices seem to have jumped up several fold - $100 down and not a single conversion is a bad sign - likely means your angle sucks, or the offer(s), or the LPs. Without ANY conversions it's also impossible to select the best offer... How are the lander CTRs looking over the past several days of data?

03-22-2014 07:18 AM #23 dnguyen (Member)

Quote Originally Posted by zeno View Post
Did you change traffic sources or countries or something? Your click prices seem to have jumped up several fold - $100 down and not a single conversion is a bad sign - likely means your angle sucks, or the offer(s), or the LPs. Without ANY conversions it's also impossible to select the best offer... How are the lander CTRs looking over the past several days of data?
No I kept everything the same. I would check at 3k impressions and keep the three highest ctr, which btw weren't very high the last two times i tested. I upped my bid to collect data faster and would then let it run for a whole day. I think my spend jumped up so high because maybe I didnt pause my other ads properly or something. Still trying to learn TJ.

I dont think the offer sucks because those were the ones my am recommended I do and I did get conversions here and there when I first started. Its got to be my lps, angles, and maybe I'm not tracking properly?

How do you suggest I move forward? Should I stop all tests until I can make better lps? Or keep testing new angles for now still and worry about the conversions after? What would you say a good ctr% for the lp be? My highest one is at 23.91%.

03-22-2014 12:43 PM #24 caurmen (Administrator)

I would check at 3k impressions and keep the three highest ctr,
Is this on banners? If so, I wouldn't recommend that approach at all - cutting on CTR for banners is basically just eliminating complete no-hopers. You should be keeping all banners with a potentially profitable CTR and then cutting on ROI - low-CTR banners can give high CVR.

03-22-2014 05:00 PM #25 dnguyen (Member)

Quote Originally Posted by caurmen View Post
Is this on banners? If so, I wouldn't recommend that approach at all - cutting on CTR for banners is basically just eliminating complete no-hopers. You should be keeping all banners with a potentially profitable CTR and then cutting on ROI - low-CTR banners can give high CVR.
I remember reading one of your posts saying something along the lines of 3k impressions is a good time to start cutting. Oh come to think of it you meant only the banners with no clicks at all. Damn lol should I test all my banners again?

03-23-2014 01:24 AM #26 zeno (Administrator)

Perhaps... but did any of them convert at all? I'm seeing a lot of clicks and ad spend but few conversions listed in your stats. Might be worth taking a step back, looking at all the data and seeing if the entire campaign is worth continuing.

03-28-2014 12:11 AM #27 dnguyen (Member)

Jeez theres a bad case of the flu going around in vancouver and I got hit bad from it. I been bed-ridden since sunday and I still don't feel good yet. Really unproductive week..

The bright side of this is I had a lot of time to think about if I should continue on this offer and after talking to my am, I'm going to give it another shot. I'm going to start from scratch with a slightly different gameplan/system. I'm going to make sure my tracking is working properly this time all the way through and be a lot more organized. Hopfully I can get some new updates soon. Cheers!

03-28-2014 01:01 PM #28 caurmen (Administrator)

Don't force yourself back into work until you're well! Looking forward to hearing your next updates.

04-07-2014 10:10 PM #29 dnguyen (Member)

Woo after two weeks of being sick and procrastinating, I'm finally getting back to the grind! So I decided to pretty much gonna take another crack at this offer and start from scratch. This time I am going to focus on being organized with my stats and really focus on trying to analyze these numbers.

Banners : 6
Lp : 3 (I really focused on having three different types/style of landers)
Offers : 4
Impressions : 130,574
Spend : $35.11
Leads : 7
ROI : (-65%)


TJ had some kind of bug so none of my ctr's for my banners were recorded down. I think I found a winner for my landing page. Lander 2 generated 5 leads while lander1+lander3 generated 1 each, however lander3 had the highest ctr% at 27.10%.

As for my banners, I had 4 banners which generated leads so I wrote down similarites in them and I am going to try to recreate similar banners of them.

Questions

Is this a decent amount of data for first rounds of test?
Is the "LP CTR" stat significant for the banners? If so what is considered a decent %?
Is it safe to cut out Lander1 and keep lander3 because of the ctr?
When I re-create new banners, should I do half of them with the same headline with different pictures and the other half with the same images I used for this round of tests and try new angles?

04-10-2014 04:02 AM #30 dnguyen (Member)

Quick update,

Impressions: 55k
Spend: $15
Clicks: 158
Conversions: 0
ROI: (-100)

Terrible day of testing, back to the drawing board. My ctr for my banners are garbage. I am going to go through the creating process differently now. I noticed I haven't been testing one variable at a time when it comes to banners so I am going to go back to the basics instead of just throwing garbage on the wall hoping it would stick.

04-14-2014 04:59 AM #31 dnguyen (Member)

Impressions: 127k
Spend: $35
Clicks: 406
Conversions: 6
ROI: (-70%)

10 banners tested with a new landing page.

Had 3 landers ran on this test. All of them hovered around 22% ctr. 2 of them got conversions so Im going to cut the lp that didnt get any conversions.

Banners averaged 0.35% ctr, with 4 of them generating leads.

I noticed that traffic has been getting slow on the placement I been testing on for the last week and a half. Bids have gone down a fair amount also, about .08 cents. Any thoughts about that?

04-14-2014 05:24 AM #32 davidwikes81 (Member)

Bids will generally go down when your CTR is good/stable and ad running for longer durations.

Now, did you find winner Landing Page? Find winner and run traffic to it for sometime. Check, what mertric performing below average. The Ad CTR? ( looks good to me ), Landing page CTR? ( can be improved, exit popups or not ), offer CVR ( might split test same offer from different network or diff offer in similar niche ), imporving funnel/flow ( making landing page simple )

But your ultimate aim is ROI...

This is my optimisation progress: http://i.imgur.com/M7uj4QY.png

You can see day1, day2, day3..... stats.

Check, how i am killing towards better ROI.

04-24-2014 09:02 PM #33 dnguyen (Member)

Quote Originally Posted by davidwikes81 View Post
Bids will generally go down when your CTR is good/stable and ad running for longer durations.

Now, did you find winner Landing Page? Find winner and run traffic to it for sometime. Check, what mertric performing below average. The Ad CTR? ( looks good to me ), Landing page CTR? ( can be improved, exit popups or not ), offer CVR ( might split test same offer from different network or diff offer in similar niche ), imporving funnel/flow ( making landing page simple )

But your ultimate aim is ROI...

This is my optimisation progress: http://i.imgur.com/M7uj4QY.png

You can see day1, day2, day3..... stats.

Check, how i am killing towards better ROI.
Its been a a lil bit since I shared some updates, but its not fun to post up -100% ROI up haha

Thanks for the post david! I definitely took what you said and tried to apply. I focused on getting my lp ctr up and i got it up to 27%.

As for my ads, I am testing a lot of different styles. Took me a little bit to understand that changing the colors and pictures is not really changing the style lol shoutout to finch for posting up his list of ad variations in an older thread.

What I am doing differently is I am keeping my campaigns small. I am running 2-3 ads at a time and tweaking just one variable. I feel like keeping it small makes it easier for me to analyze data.

Anyways here are my stats for my best test so far since I started AM.

Impressions: 54k
Clicks: 102
Spend: $17.38
Leads: 4
ROI: (-59%)

Im still in the red, but its close to -50% so I feel like I'm heading in the right direction! Is the sample size to small? I still can't figure out what you guys mean when u say 2x payout. If somebody can clarify that for me it would be much appreciated.

Because the AB split test calculator isn't showing a 95% chance of one being better than the other, I should let this run longer correct?

Is this a good time to start tweaking out the smaller variables (colors,font,cta's) or should I still be focusing all my time on angles still?

04-25-2014 02:16 PM #34 caurmen (Administrator)

"2x payout" means "a spend equal to twice the payout you recieve for a successful conversion". So, for example, if your offer pays out $2 on a successful conversion, 2x payout is $4.

04-28-2014 05:24 AM #35 dnguyen (Member)

Spend: $30
Conversions: 3
ROI-80%)

I tried to optimize my ads and lp and got a disappointing result.

I changed the pictures on my landing pages and the ctr plummeted. The obvious answer was the image performed worse, but I was thinking maybe its because of the weekend? I was getting traffic at a slower rate I noticed on the weekend. Im thinking i should have let my last campaign run through the weekend to compare weekdays to weekends. Also I can't figure out which style of lander is better. Everytime I put it into the split test calculator, they usually are pretty close to even. Any suggestions on how to get enough significant data quicker for them? Until then I am still going to be rotating them both for every campaign.

This is the calculator I use when analyzing both my banners and landing pages. http://www.peakconversion.com/2012/0...al-calculator/
When it asks for "trials" I punch in the number of clicks and for "successes" I punch in leads, correct? What is a good percentage to know which one is better?

I know you wrote a guide for this already caurmen but I was kind of confused still after reading it lol

04-28-2014 06:35 AM #36 zeno (Administrator)

Were you split testing the original lander in parallel? Did it's CTR also drop?

For the calculator you really want trials = clicks to lander and successes = leads, i.e. the EPC of the lander. Lander CTR is an intermediate metric so better to make decisions based on how much money the lander makes, rather than how likely people are to click through to the offer.

The confidence interval is up to you but I would go for something like 90% or higher.

04-28-2014 03:00 PM #37 caurmen (Administrator)

Yep, you're using the calculator correctly! Trials are clicks to lander, as Zeno says.

Make sure you're testing both landers at the same time - otherwise it's very easy for other factors to skew the results.

04-28-2014 11:18 PM #38 dnguyen (Member)

Quote Originally Posted by zeno View Post
Were you split testing the original lander in parallel? Did it's CTR also drop?

For the calculator you really want trials = clicks to lander and successes = leads, i.e. the EPC of the lander. Lander CTR is an intermediate metric so better to make decisions based on how much money the lander makes, rather than how likely people are to click through to the offer.

The confidence interval is up to you but I would go for something like 90% or higher.
Oh damn I didn't even think about running the original in parallel.. Definitely going to do that

Hey does it matter if I run say for example 10 banners with 4 lp in one campaign? Or would it be better to split it up, into like 5 small campaigns with 2 banners+2lp each? Or does it make a difference at all? It should be easier to see which funnel works better when its broken up smaller right?

And prosper has a new update, if I update it will it mess up my stm mobile tracker at all?

04-29-2014 08:10 AM #39 zeno (Administrator)

This depends on your tracking. There's no point of splitting it into separate campaigns if your tracking can rotate 4 banners fine and you have keywords identifying each banner.

04-29-2014 11:46 AM #40 caurmen (Administrator)

Don't update to the latest version of Prosper if you're running the STM tracker - bad things will happen!

04-29-2014 06:48 PM #41 trafficjunky (Member)

Hi dnguyen,
This post is very interesting. At TrafficJunky we make our best to help out our advertisers.
If you have questions or doubts (like the CTR question you had earlier), do not hesitate to contact our support:
http://marketplace.trafficjunky.net/contact-us

05-02-2014 08:17 PM #42 dnguyen (Member)

So this week I have been testing a new angle every day. For each angle I would do 4 banners with 2 landing pages at 10$ a banner.

This is my most promising campaign

Impressions: 109k
Clicks: 241
Leads 10
Spend: $33.51
Revenue: $17.50
ROI: (-48%)

I need to get my cpc down. Banner ctr's were kinda low (.217%) so I could improve on that. Should I be playing around with my bids at this point? Whenever I bid, I usually just bid a little more then the top bidder.

I tested two landers and at one point one of them had a 88% probablity of being better and I was going to cut the other one, however my internet went down so I couldn't get around to it. I checked my stats again this morning and the lander I was going to cut caught up. Its at 66% vs 38% right now. What would be a good number of trials before cutting?

Judging by these stats, would you guys say the angle is decent enough to turn a profit? I'll still keep testing new angles regardless though.

05-02-2014 10:16 PM #43 caurmen (Administrator)

That's quite a big spend per banner - what's the payout on the offers you're testing?

05-03-2014 03:47 AM #44 dnguyen (Member)

Quote Originally Posted by caurmen View Post
That's quite a big spend per banner - what's the payout on the offers you're testing?
1.75/lead

How much would you suggest I spend? I thought about 2x the payout but I figured I needed to spend more to give my landers enough data.

05-03-2014 12:07 PM #45 caurmen (Administrator)

OK, CTRs will vary by placement on TJ, but based on the CTR you list I'd recommend the following rules for banners:

- Cut anything that hasn't had a click after 1,200 impressions.
- Cut anything that's below 0.1% CTR after 5,000 impressions.

After that, cut on ROI - follow Step 10 of the Getting Started Guide: http://stmforum.com/forum/showthread...TION-AND-INDEX

Unless you're planning to link one lander to one banner only, you can keep getting your landers data whilst optimising your banners - just pause banners as needed, add new ones, and keep running traffic to your landers.

Does that make sense?

05-05-2014 06:01 AM #46 dnguyen (Member)

Quote Originally Posted by caurmen View Post
OK, CTRs will vary by placement on TJ, but based on the CTR you list I'd recommend the following rules for banners:

- Cut anything that hasn't had a click after 1,200 impressions.
- Cut anything that's below 0.1% CTR after 5,000 impressions.

After that, cut on ROI - follow Step 10 of the Getting Started Guide: http://stmforum.com/forum/showthread...TION-AND-INDEX

Unless you're planning to link one lander to one banner only, you can keep getting your landers data whilst optimising your banners - just pause banners as needed, add new ones, and keep running traffic to your landers.

Does that make sense?
That sounds like a really efficient way of doing things lol I'll start doing it like that from now on. I think I understand the step 10 of the Getting Started Guide. I'll start applying that as soon as TJ can get their analytics fixed.

Anyways here are some stats from the weekend

Impressions:140k
Clicks: 374
Leads: 23
Spend: $62.35
Revenue: $40.25
ROI: (-36%)

Seems like the little optimizing I did do improved my campaign! The CTR on my banners are up from .217% to .265%.

By the way, do you know what timezone prosper202 runs on?

05-05-2014 12:45 PM #47 caurmen (Administrator)

Prosper runs on the timezone that's set in your php.ini file by default, as I recall. There's a function to change the timezone in the admin section of prosper, but some people have reported problems with it in the past: you may be better to just change it in PHP.ini.

One trick here: you can go to the Spy view, click on one of your landers yourself, then look at the latest clicks and see when they're reported as having come in. That'll tell you what time Prosper thinks it is, as opposed to when you think it is!

05-12-2014 06:15 AM #48 dnguyen (Member)

Hey guys! Its been a little while since I updated. Its been a busy weekend cause of the crossfit regionals. Shoutout to all the crossfitters on stm!

Anyways, this past week I've been testing some new angles and got a bunch of duds. Before the weekend hit I had three campaigns ready to run through the weekend because TJ doesn't approve ads on weekends. I didn't really check my stats during the weekend to much except just to monitor how much each campaign is spending until it hit x3 payout on each banner. I bidded on the same spot for 3 campaigns so I was getting like 10-20% of the traffic on each campaign for most of this test. After almost 2 days of running, I ended up with 2 campaigns that failed and my first camapign in the green!

Impressions:53k
Clicks:137
Leads:12
Spend:$15.70
Revenue:$21
ROI: (+%33)

I had 3 ads with 2 lp for this campaign. Ads were at %0.25 CTR and all of them had pretty much the same number of leads.

Between my landing pages, I finally had a clear cut winner. %99 probabilty!

I'll let this campaign run longer to get more significant data.

Questions

I ran this campaign for about 2 days and it only managed to spend $15. How do I get more traffic without jacking up the bid prices? I'm on redtube (the second highest traffic at 4mil/day) but should I go to pornhub (most traffic on TJ at 6.3mil/day)? My other 2 campaigns I ran parralel to this was able to spend more though so that could have been a reason.

I read a post from Tom Fang and he was talking about how he suggest to anybody in the adult space to put a frequency on their campaigns. I tried to add a frequency of 3 on a cpl campaigns and noticed I was barely getting any traffic at all. Should I even worry about frequency right now and play around with it when I know for sure I got a winning campaign?

05-12-2014 06:56 AM #49 robh (Member)

Congrats on being in the green though, I have given up several times on TJ because of scalability issues. Admire the persistence.

05-12-2014 06:32 PM #50 trafficjunky (Member)

@OP
Congratulations indeed!
As a matter of fact, you are right: PornHub will yield more impressions and the audience is quite similar to the RedTube one, you should not have major changes in the way users react to your ad or convert.
In other words: the same ads should be the ones performing, same thing for the landing page.
Only thing to watch out for if you plan to extend to other sites (my comments also work for Youporn vs Redtube) is the CPM bid, you might need to make some adjustments in order to get a similar visibility.

If you have questions, do not hesitate to contact our support, they will be happy to help!

@Robh, would you mind getting in touch with us in order to explain your scalability issue? As we said on another thread, we are gathering user feedback in order to improve our system

05-26-2014 10:31 PM #51 dnguyen (Member)

Hey guys, its been a while since i posted anything. I haven't been able to hit green again since the last time i posted. I've been hovering around -30% and can't seem to get out of it. Part of it is because of my inability to analyze data properly and a lot of it has to do with my work ethic. I've been going out a little to much recently and I need to stop procrastinating.

Anyways, after reading one of finch's threads about ad placements, I decided to try 305x99 banners but I just don't know how I should test this. Should I keep my banner exactly the same as my 300x250 banners and run both of them at the same time at the same bid to see which placement performs better? And how many banners tested would be significant?

05-27-2014 05:28 AM #52 zeno (Administrator)

Keep the same general theme/layout but obviously edit the source to make it look decent.

Split testing different banner sizes is inherently an unfair test since the placements and banner layout have to change!

05-27-2014 01:27 PM #53 trafficjunky (Member)

Quote Originally Posted by dnguyen View Post
Hey guys, its been a while since i posted anything. I haven't been able to hit green again since the last time i posted. I've been hovering around -30% and can't seem to get out of it. Part of it is because of my inability to analyze data properly and a lot of it has to do with my work ethic. I've been going out a little to much recently and I need to stop procrastinating.
Hi dnguyen, I'm just checking here: do you use our conversion tracking tool?
If you don't, you might want to give it a try, it should help you analyze the data since it will show you right in the interface, which campaigns / banners are actually converting.

Keep up the good job!

05-27-2014 06:45 PM #54 dnguyen (Member)

Quote Originally Posted by trafficjunky View Post
Hi dnguyen, I'm just checking here: do you use our conversion tracking tool?
If you don't, you might want to give it a try, it should help you analyze the data since it will show you right in the interface, which campaigns / banners are actually converting.

Keep up the good job!
No I haven't. I didn't even know you guys had one. I'll check that out!

05-27-2014 06:47 PM #55 dnguyen (Member)

Quote Originally Posted by zeno View Post
Keep the same general theme/layout but obviously edit the source to make it look decent.

Split testing different banner sizes is inherently an unfair test since the placements and banner layout have to change!
Hmm so how should I go about this? Should I focus on one placement at a time then?

05-28-2014 02:38 AM #56 zeno (Administrator)

No, just test as you normally would. Focusing on one placement would be an ideal test but can narrow your vision.

So, for a specific angle/banner, make that banner in all available sizes and then split test them.

The banner size is a variable you are split testing.

The banner design is approximately the same - you can't test everything perfectly so just consider them as having the same design. The 'design' is a nebulous thing anyway.

Placements will invariably change with banner size so just consider this as a combined variable - i.e. site 1 with 320x50 banner = placement A, site 1 with 320x250 banner = placement B, site 1 with 120x600 banner = placement C, etc. Evaluate the performance and profitability of these placements independently but remember they all come under the same 'angle' so can be grouped and culled on the basis of the angle not performing.

05-28-2014 07:42 PM #57 dnguyen (Member)

Quote Originally Posted by zeno View Post
No, just test as you normally would. Focusing on one placement would be an ideal test but can narrow your vision.

So, for a specific angle/banner, make that banner in all available sizes and then split test them.

The banner size is a variable you are split testing.

The banner design is approximately the same - you can't test everything perfectly so just consider them as having the same design. The 'design' is a nebulous thing anyway.

Placements will invariably change with banner size so just consider this as a combined variable - i.e. site 1 with 320x50 banner = placement A, site 1 with 320x250 banner = placement B, site 1 with 120x600 banner = placement C, etc. Evaluate the performance and profitability of these placements independently but remember they all come under the same 'angle' so can be grouped and culled on the basis of the angle not performing.
Ahh alright thanks for clarifying that zeno!

06-02-2014 10:14 PM #58 dnguyen (Member)

Alright stm brothers, my creative juices are exhausted on this offer. I think it's time to call it quits and test other offers out. I definitely learned a lot which was the main goal anyways. I've talked to my am and he signed me up for a couple other offers. I think a fresh start is what I need! Thanks to all who provided me with information. I'll start a new follow along as soon as I get set up and hopefully I can get this done right!

Home > Mobile > Follow-along Campaigns