Patience, persistence and perspiration = success (31)
Patience, persistence and perspiration make an unbeatable combination for success.
Napoleon Hill
He must have been an affiliate marketer!
After a couple weeks of banging my sleep deprived head (new born baby caused) against a wall regarding my woeful tracking situation without solution (see here http://stmforum.com/forum/showthread...le-App-Install), I'm going to crack on.
I've decided rather than get frustrated, which I am, I'd rather plug on and try to get the other foundations right. After reading caurmen's 2014 step-by-step guide last night on stats and coincidently reading Dr Ngo's post with free reporting spread sheets, I've decided to nail the campaign analysis element in readiness for my next campaign, by which time the tracking problems are resolved.
If anyone has any benchmarks they use in terms of cutting ads or pivoting based on stats that they are keen to share please do. I will be analysing my stats and update you all on what I think.
NT
Good stuff! I shall look out for updates and offer advice if needed!
Tracking update: still buggered.
So I've put the elements of the mobile cookbook main course (http://stmforum.com/forum/showthread...006#post154006) and the Step by step guide (http://stmforum.com/forum/forumdispl...-By-Step-Guide) together so put what I think is something workable for me.
Would love to get people's feedback on this. It obviously doesn't take into account the creative element but it's a blueprint for me:
1. Choose an offer valid in 3 countries and on 2 affiliate networks
2. Create 10 angles
3. Choose 2 angles
4. Create 5 banners for each angle and translate for the 3 countries
5. Put live with "smart CPA" of 3x the payout
6. Any banner with 1200 and 0 clicks pause.
7. Any banner with 5000 impressions and less than 0.1% CTR pause
8. Replace any banners to keep at least 5 banners in rotation at all times
9. Any banners with over 8k impressions and over 0.1% CTR should be kept running and 3 variations of it should be made
10. Any banners with 10k impressions check maximum possible conversion rate vs minimum viable conversion rate - cut as necessary
11. All banners by this stage should have an acceptable CTR or be paused
12. Create 3 landers to split test
13. Test all remaining campaigns with another network
14. Use binomial confidence intervals to choose remaining ads
15. After at least 5 conversions per lander calculate best lander with >90% probability
16. Create 3 variations of best lander and test.
1-16. Keep daily record of all major KPIs
17. Scale (haha)
Would love to hear what you experts think of this. I think the 3 countries might be running before I can walk.
NT
good job , keep update how the campaigns are going for you!
You can achieve considerably higher CTRs with mobile than with desktop, so I'd up your initial hurdles considerably.
Obviously they do still vary by country and placement, but if you want to take this approach, I'd kill anything that hasn't had any clicks after about 700 impressions, and kill banners with less than 0.2% CTR after 3000 impressions.
Don't make variants of any banner that hasn't passed at least one ROI cut (stage 10). Also, once you're past Stage 10 you don't care what the CTR is - if I have a banner with 0.08% CTR but it's profitable, I'll keep it running.
You don't have a point where you will decide whether to kill campaigns, but you refer to "all remaining campaigns" in Stage 13.
You don't need 5 conversions per lander, just 5 conversions on the most successful one.
And 17 will happen, if you keep at it!
Hope that helps!
Thanks Caurmen.
Will be starting afresh with this as a blueprint when I set up a new campaign this weekend.
I may have dreamt it between the extreme tiredness and the baby crying at 3am but what happened to the thread about hot girls being dumb? It was pure entertainment.
ok so in the name of persistence....
I've decided right or wrongly to go with the same dating app install as before but using ClickDealer rather than F5. I'm also changing the country I'm targeting.
I've created 12 banners. I wouldn't say they are massively different but have slight differences in messaging. My feeling is they need to be completely different. But based on my last trackless campaigns I did get conversions so they must be hitting a good note somewhere.
Waiting for the offers to be approved so I can set up
I've amended my blueprint slightly. How does this look now Caurmen?
1. Choose an offer valid in 2 countries and on 2 affiliate networks
2. Create 10 angles
3. Choose 2 angles
4. Create 5 banners for each angle and translate for the 2 countries
5. Put live with "smart CPA" of 3x the payout
6. Any banner with 700 and 0 clicks pause.
7. Any banner with 5000 impressions and less than 0.2% CTR pause
8. Any banners with over 8k impressions and over 0.1% CTR should be kept running
9. Any banners with 10k impressions check maximum possible conversion rate vs minimum viable conversion rate - cut as necessary
10. All banners by this stage should be profitable or be paused. Variations of profitable banners should be made.
11. Create 3 landers to split test
12. Is the campaign making at least +20% ROI? Yes, keep going. No, kill.
13. Test all remaining campaigns and banners with another network
14. Use binomial confidence intervals to choose remaining ads
15. After at least 5 conversions on a lander calculate best lander with >90% probability
16. Create 3 variations of best lander and test.
17. Keep daily record of all major KPIs
18. Scale (haha)
Looks great - although on point 13, are you planning on testing one network first and another second?
You should split test between networks from the very beginning.
ok yea makes sense Zeno. Will do.
That looks like a good list, keep us updated!
Thanks bb! I've just created new campaigns but the android offer is going to the iOS, so need to contact the AM.
Looks good! My only confusion point is #14 - surely you're already using binomial confidence intervals in #9?
ok yes that does make sense!!
I've set a few campaigns live will update tomorrow.
Question for the pros: how many unsuccessful campaigns did you set live before hitting the "ah ha" campaign? (need some motivation!)
50±49.
Not what you want to hear but I don't think there is any average number!
It depends on far too many factors.
However, I would suggest you set your expectations at having at least 10 unsuccessful campaigns before making much ground.
Whether you will take more or less than this will depend on strategy, offers selected, traffic sources, competition, optimisation procedure, etc.
Just go in knowing that 95% of the things you'll try over your affiliate career will fail, and your goal should be to identify these asap to minimse loss, while taking necessary risks and spending money to get data (== knowledge).
Thanks Zeno that's what I expected and wanted to hear!
Also keen to get everyone's opinion on this mobile lander: http://neotmarketing.com/landing/IT.html
Things I'd like to hear about is the general structure, images, CTA, technicals of the html, load speed and anything else you think is important for a good landing page.
I set a campaign live yesterday with the lander so will update later tonight.
lander looks good, although looking on an iphone 5S it does cut the CTA off in safari. Chrome on IOS is good to go (hope that helps)
So yesterday's stats from Decisive with the new mobile lander:
The only ray of hope is the tracking is working!
This conversion and previous ones that didn't track 100% came from the skout app network. Clearly unhappy skout users want to jump ship. With this in mind I'm going to amend the banners and lander to focus on unhappy Skout'ers and target those through Decisive.
Good idea?
....however, looking at today's stats for the same offer I can't help thinking that either the lander is still a dog or the offer isn't worth pursuing.
Guys am I falling for the offer when i should dump it and move on?!
Lander CTR and how many clicks to the actual offer?
25th - 509 banner clicks >3.3%> 17 offer clicks> 1 conversion
26th - 447 banner clicks >1.3%> 6 offer clicks> 1 conversion
Today - 228 banner clicks >0.4%> 1 offer click> 1 conversion
Honestly that lander CTR is simply prohibitively low.
CTRs aren't critical but you do need to get eyeballs on the page that can generate revenue.
It could be the lander (and/or a tech issue with it), it could be the angle, it could be the traffic.
I would move on to greener pastures right now and put it in the "might revisit later" box.
This current campaign appears overly demotivating and that is bad for business!
Yea I agree I'll move on and possibly revisit.
Originally Posted by nt2000
I've gone a bit silent on posting recently but been working to get the basics right. Angles, banners and landing pages.
After only receiving a couple conversions and as @Zeno says having really bad lander CTR i thought it's time to get the landers' images updated. Along with the headline probably the biggest factor (correct me if I'm wrong!).
So i came up with these angles for a casual dating app install offer. I noticed that the conversions I did get from Decisive were from the Skout and Grindr apps so want to target them specifically when running the next set of banners/landers.
So angles/headlines I've come up with (Im sure this has noob written all over it!):
- Shy guys wanted by HOT girls
- Over 24 years old and tired of SKOUT?
- Italian girls looking for love, maybe more?
- 1 app every single Italian MUST HAVE
- These foreign girls LOVE Italian men
- These girls are desperate, help them but dont tell anyone!
- Treat these girls well and YOU'LL be rewarded
- SKOUT was yesterday, Lovoo is the future!
- Using Skout and still single?
- 1 app all the studs are using
- Chat, flirt, FU....
- Feel her soft lips caress your face. Find her LIPS NOW!
- Be the guy with the hottie! Make your friends jealous.
- Don't be old and single, find a HOT YOUNG girl
- Still a virgin? These girls will take charge and show you!
- 1 trick to get any girl (have i stole that?)
- Like SKOUT but LOADS better!
I used the same lander and used different images. I think the sexy images should work better than the stock app image. The CTA button pulses. Typo in the headlines - arrive. NEED MORE CLICKS ON THE OFFER PAGE!
Create some rough and ready banners:
I've set an Italian campaign live with the landers above and submitted the banners for a UK campaign on Decisive.
Keen to hear what the pros think of my creatives and angles. Am i heading the right direction?
Good work - keep on plugging!
The landers - have all your landers got that much text on them? Too much text is a killer for CTR on most landers - and I speak as a very wordy man here, so believe me when I say I've tested that thoroughly. It's particularly bad for mobile.
Yea I think the landers might be a bit wordy too. With trying to get too many factors in eg heading, bullet points, testimonial, ratings etc its become very wordy indeed. I'll remove the testimonial and see that happens.
Today's lander stats 223 banner clicks only 10 click on the offer so a LP% CTR of 4.5% with 1 conversion. The banners are working in my eyes, I assume this as they are getting the clicks. Admittedly they could be the wrong type of traffic. I haven't hit the right angle for the lander that gets them clicking through.
What did everyone think of the angles and banners?
I'm going to get live the UK campaigns. I'm not hopeful!
Angle-wise at least you are thinking of lots - some times a good way forward is to just bulk test everything you can think of.
Your landers are better, but as Caurmen said, you should kill some of the text.
Too much of it IMO!
Or perhaps reorganise and put the testimonial beneath the CTA and make it more clear that it is in fact a testimonial/social proof.
Something I'd recommend as an exercise for almost anyone learning AM: try a really, really minimal lander. Can you get it down to 5 words and no image? If so, try that! It might work, it might not, but the exercise alone will be worth it.
That's interesting as a lot of what you read suggests having lots of different factors.
Yes I think it's solved the dv1 and dv2 in the offer url was incorrect.
I manually fired the conversions saved as dv2 and the conversion appeared in both
Pretty pleased with myself! 
Home > Mobile > Follow-along Campaigns