Mobile app install in BR + my story (16)
Hi to all,
have posted one follow along already, here is another one. Working my ass off, tired a lot of not being successful after so long time. But strong believer and going to make it.
I have a really entrepreneur past with lot of ups and downs. First summer job got at the age of 15 for oil drilling company carrying some seismic sensors and cables in muddy forests for nearly no money at all. I was 15. Shitty job, decided never to go that way. First business started in my 16 (year 2000) selling CDs with some useful data over internet nearly on autopilot. Just to burn a CD and mail it for couple $xx. Running till today, delegated, download with credit card payment. Then tried to create online app, totally crashed it, costed around $5000 + 1500 hours of my time. The worst loss however was my girlfriend at that time (after 8 years together), who was totally against the project. After the split finally closed the project down. Started another project and burned like $25,000 and finally made it slightly profitable ($xxx/month). Had to delegate and withdraw. Never going to make that money back. Great lessons though!
Fast forward 15 years on got little stuck running brand/design/web agency as CEO with small team of creatives. I got so tired of all the clients and their constant requests. Making some decent profits, but I HATE the feel of working for success of others.
So I'm 31, 5 months in AM now. Running agency nearly on autopilot for 5 hours/week. Focusing on the ONE THING. Affiliate marketing. Still feel motivated and going forward. I'm so much willing and working my ass of. And still hoping my success will come soon :-)
So here is the follow along:
Networks
Popads, Decisive.
Tools used
iMobiTrax tracking,
VPS Beyond Hosting (Managed dedicated server - DVS 12GB tariff)
whatrunswhere mobile
STM forum
Offers
Clickdealer (7077) UCBrowser /BR [WAP] *INCREASED PAYOUT* ($0.36)
F5 (4929) [WAP] UC Browser - CPI - BR - Android ($0.36)
Clickdealer (7650) [WAP] DU Speed Booster /BR ($0.70)
Batch A - decisive
About 20 different banners, most of them from the Clickdealer offer (UCBrowser native design)
Results from Imobitrax
Results from Decisive
Most clicks
Conversions:
Batch B - popads.net
Pop-up traffic, max bid 0.01 CPC, targeting non-adult traffic. Mobile+ cellular.
Batch C - decisive following Caurmen's mobile appetizer
Approved 5 minutes ago, letting them run.
Some TODO, to come in next post. I guess I'm missing a lot necessary info to write down, please let me know and will write it. At the moment formulating ideas even what to ask :-)
Good start mate, looking forward to seeing the data roll in!
Keep it up - from what I can tell it's the people who stick to it that succeed, and it seems like you've been able to persevere through a lot in your life!
Day 2 here:
Batch A - decisive
campaign running now, data later today
Batch B - popads
campaign running now, data later today.
Batch C - decisive
- have won 15% of the bids. Would assume it’s not the best quality traffic and there is a lot of competition. Correct?
- average around 2% CTR. Not happy with that. Is this OK?
- had just 2 conversions on site, 0 on app.
- as about the appetizer: would you close this campaign?
- changed the speed of the campaign to run more even
TODO
- find some more offers to split test
- try different batch of designs
- find out if such small conversions are OK
- working with coder on optimized landers (much quicker) to add to split testing. Going to be finished today.
General questions
- is Brazil competitive market for buying traffic (should I move somewhere else like Asia, etc?)
- how about the speed of the loading - would you use CDN?
Brazil is definitely a competitive market.
But so it the US, and wherever it is competitive, people are making money.
For mobile you should definitely be using a CDN for static content unless you are using a self-hosted tracking system and the lander is tiny. You're using iMobiTrax and I'm not sure if they need any PHP code on landers. If they do, probably best to self-host it all and keep the lander size slim.
Lastly, why are you using English banners for a nation which primarily speaks Portuguese?
Thanks Zeno for the feedback. Translating all banners into Portuguese now, so with the landers. Will update you soon...
update?
Have been working a lot last 5 days, have to get things in order properly
Hired portuguese translator and new good programmer for LP's. LP's are doing much better now.
All banners and LP's are in portuguese now.
Here are the stats:
Batch A - decisive
Site. Some banners has 3% CTR which I guess is good value. 6 conversions.
Overall -94% on the campaign
App - no conversion at all. CTR around 1,5% which I guess is good for APP.
Overall -100%
Batch B - Popads.net
Best batch of all so far.
Really like the CTR on LP's - 30% on the best P2. Is this really normal? Offer conversions on about 4.5%. Is this OK?
Best LP is about -60%.
Removed some offers which weren't converting. Also removed 3 worst landing pages.
Batch C - WigetMedia
Really confused from this one.
Some landing pages has really low CTR, going to kill them.
Also totally different offers are converting, as compared to Batch B or A.
CPM is $3.10 which is quite low I guess.
TODO NOW?
- ask traffic manager about higher quality traffic?
- kill worst LP's?
- kill offers which are not converting in Batches A an B?
Overall thinking
$135 spent, $11 revenue. Not a great start I guess.
is this campaign even going to get profitable or should I kill it?
Have seen in other forum that a lot of people are running this offer in BR. So hard to succeed?
Whatever suggestions are welcome.
One TIP for MOBILE: Have made a easy php page for landing pages previews.
Using iframes which allows us to see all the landing pages from folder on server in one place. As mobile is about 500px wide, 3 LPs do fit on one screen standard width. Don't need to go through all the files one by one or open them in browser. And it's great for translators: just give them the filenames we want to translate. Perfect.
See the screenshot, you'll get the idea.
Attachment 4342
So thinking loud here:
LP's with "alert" are doing much better than without it - landing pages 20 and P2. Approximately 2x better.
(Made sure that "alert" click is not counted as another click (2 clicks instead of one). So the data should be really OK.)
So will split-test best LP's by adding "alerts"
TODO:
-split-test best LP's by adding "alerts"
-could use converting offers from one batch as offers for other batches?
-use CDN or speeding up? Imobitrax is self hosted, need a line of PHP for tracking in there. Size of LP is about 4,5kB + some pics (30kB max)
-app converting on "install and open". Should we tell the users to open it?
-is this campaign even going to get profitable or should I kill it?
Batch A - Decisive
- try direct linking?
- completelly kill the app batch?
- remove non-converting offers (from batch B)?
Batch B - popads
-split-test best LP's by adding "alerts"
Batch C - WigetMedia
- ask traffic manager about higher quality traffic?
- kill worst LP's?
- kill offers which are not converting in Batches A an B?
Update: got in contact with the traffic manager of WigetMedia (batch C), my price $3.1CPM was actually highest bid yesterday. So. making the situation even more puzzling for me :-)
Looking at the results for these, I'd be tempted to try a completely different angle.
You're going to need a BIG boost to get these into the black - small tests like alerts probably won't do it.
How much placement culling have you done on these campaigns so far? A bunch of bad placements could make the difference between profit and loss even at these ROIS.
Thanks caurmen.
Haven't done nearly any culling yet. Went through the data but there wasn't any significantly big enought site to be banned. Will have to go through the data again to make sure.
As about the angles: we have about 5-6 different angles in the LP's and I guess in the banners too. Hmmm, will try to add some more I guess.
Originally Posted by maweniaran
I noticed you ripped some stuff that's a good start. You might want to consider how what you ripped works. For example they have a lander, if you try their banner itself it might be bad but their lander really huts home, again the same lander might not work on a different banner. It's that unique banner lander mix.
Next is the method of improvement, what improvement will cause the roi to change greatly? Google Charles ngo and see what he writes on angles
What do you understand about the traffic source? I'll refer to pop ads here because I understand this source slightly better.
Think of the traffic source as a women. They are unique in their own way. There is no way to predict what they have inside until you dig in and take a look. But here are some guidelines
How is the bid algo written? Top bidder gets all traffic? Do they force traffic to the lower bidders even if the top bidder gets all?
What kind of placements are you exposed to on different bid levels? Not just the presence of exposure but also the amount?
I'm going to stop here and say that traffic source is not your biggest issue. Your angle is very weak. My initial tests on decisive and pop ads usually shows up with at least a negative 70 or negative 50% roi for most geos if there is any hope.
Your data is way too little. Either you decrease the number of variables you are testing or you increase your spend. Think about this 7 banners by 7 landers is 49 variations, while 3 banners by 3 landers is 9 variations. The former requires 5x more traffic to test. And shows results slower.
Thanks xavierfok. That's a lot of suggestions to improve, will take me a while, but I will do it. Probably going to have some questions too..
Home > Mobile > Follow-along Campaigns