Nirvana's POP Follow Along (29)
Hello all and welcome to my FA.
After some learning, setting up, ripping, etc.. Starting my 2017 with a campaign.
Offer : AV App install - Battery Plus
GEO : US
Traffic: Popads
Payout: $0.7
I have 4 landers setup. I lack technical knowledge to code them so I have decided to go with 4 of the simplest ones I could find on Adplexity. Made up for my lack of tech knowledge with some well thought angles, I hope.
3 campaigns setup ($30 each)
LB - $3
MB - $4
HB - $5
Sent for approval. I will update when budget runs out. Thanks
Hi guys the results are in
LB campaign got no traffic.
My budget burned out during my geos sleeping hours. The pic below shows stats before any edits. I am pleased to see some conversions.
Landing page stats
HB campaign - $5
MB campaign - $4
My thoughts
I personally feel that $30 was too little and I do not have enough data.
Questions to ask
1. Should I pause LB Campaign or wait for the budget to eventually run out?
2. How much more traffic should I run?
3. What would you do in this case?
4. Any edits I should be making at this stage?
5. Would you like to see specific data from the campaigns above?
Don't run the LB campaign anymore.
How many offers are you testing? Just 1 from 1 network or?
Campaign HB and MB are off to a good start. Here's what you should do. Run twice as much traffic as you just ran to each campaign. then look at all of the data and keep the best lander (or best 2 landers if they are +/- 10% in ROI) for each campaign.
Then, add in the same offer or similar offer from as many networks that you are signed up with. Run another $20-30 spend to each campaign (using the 1-2 best landers).
Review the stats and see if 1 or 2 offers stand out really well, if not, run another $20-30 to each campaign. Keep the best 1-2 landers and the best 1-2 offers if they are performing similar.
(*revised) Then run another $20-30 to each campaign and go in and cut all the bad placements that aren't converting. Run another $20-30 to each campaign, review stats and cut all placements that are performing worse that -25% ROI. If you can adjust the bids on each placement, then reduce all bids on placements that are between -25% and 10% ROI, reduce their bid by 25%.
Run another $20-30 of traffic to each campaign, repeat the above step.
Obviously, you can look more closely at the stats and see if carrier/wifi is performing really bad then more options can be implemented.
Cheers to a good start to your follow-along.
Andrew
@nirvana - I made a revision to my post, please read again.
I would do pretty much what mrpayne said above. Great start if those are your first campaigns!
1. Should I pause LB Campaign or wait for the budget to eventually run out?
- I would throw in a little bit more traffic just to see if it converts at all.
2. How much more traffic should I run?
- That $30/day should be enough for you to gauge initial performance. The higher your budget the faster you can get data, though don't overdo it, no point in overspending. You can already see how much you need to spend to get a handful of conversions, use that as a ballpark number to test more landers and (similar)offers, and possibly gradually increase it based on however many more variables you may have (landers+offers).
3. What would you do in this case?
- Get some more landers and similar offers in the mix, or even the same offer from different networks (the difference here can sometimes be loser to winner due to a number of factors like having a brokered offer vs a direct deal).
4. Any edits I should be making at this stage?
- How's your LP loading speed looking? It's pretty crucial to have fast loading landers on pops.
- Another big downside of pops is all the bot traffic that can be eating up your budget, definitely test for bots and drop the bad placements. This can be done early on, too. You can do this with various scripts or even Afflow (a search here on STM will show you a bunch of ways).
- Test different bids too High/Medium/Low etc., while a high bid might get you better and more traffic you never know if a medium bid might get on the green side.
5. Would you like to see specific data from the campaigns above?
- Once you've gone over a bunch of landers and offers and cherry picked the most promising combos I would only then dig into the traffic source and attempt to optimize there. I.e. pretty much go over all of the ways you can target a campaign and optimise for them. However, try not to over-optimise either - if your targeting suddenly becomes too narrow you may find yourself having worked so hard for a measly $5/day campaign because that's all the traffic you can get because of that hyper-mega optimised campaign.
That's about all I can figure out off the top of my head. Stay persistent and this stuff will work out for you.
You are getting solid advice and I'm really pleased that you got pretty solid performance straight off the bat!
Based on what you showed us so far, I would suggest a few more steps :
- Looks like MB was the best hit so far, I would try whether there is a level between LB and MB that could perform even better.
- You don't have enough data to optimize everything, but you can already start looking for some VERY bad targets or placements. There are pretty much always some placements or targets that you can cut really early.
Good going so far, looking forward to your results 
I ran another $30 each on HB and MB.
Campaign shot
Middle bid
High bid
Taking out terrible placements and running another $50 to each campaign before making any edits. I feel I still do not have enough data to go with any specific landers.
Cut based on statistical significance using this too: https://win-vector.shinyapps.io/CampaignPlanner_v3/
Go to the Evaluate campaign tab on the left. Anything with a 90% or better rating is a winner.
Andrew
Originally Posted by shishev
2. Planning on increasing budget and test a few more versions of the landers.. if all looks well, I'll look into playing with the bid.
3. Will definitely split test between networks. A new offer as well I want to play with. Too soon to jump on that one I think.
4. Page speed and all looks fine. I removed bot placements manually after my 2nd test. Looked for unusual activity and cut it. I am currently testing 3 different bids.
5. After 2nd update, all I have done is blacklist placements. I am waiting for more data to cut landers.
Thanks for your very helpful advice. Progress report coming shortly.
Originally Posted by mrpayne
One offer from one network. In talks with different networks. Chinese new year puts me 3 days behind.
I am running another $50 to each after blacklisting bot/horrible placements. Will look at landers after that.
Getting more offers and networks on board shortly.
(*revised) Then run another $20-30 to each campaign and go in and cut all the bad placements that aren't converting. Run another $20-30 to each campaign, review stats and cut all placements that are performing worse that -25% ROI. If you can adjust the bids on each placement, then reduce all bids on placements that are between -25% and 10% ROI, reduce their bid by 25%.
^ Will apply this as I get more data. Many thanks.
Originally Posted by matuloo
Based on what you showed us so far, I would suggest a few more steps :
- Looks like MB was the best hit so far, I would try whether there is a level between LB and MB that could perform even better.
- You don't have enough data to optimize everything, but you can already start looking for some VERY bad targets or placements. There are pretty much always some placements or targets that you can cut really early.
Good going so far, looking forward to your results
I just removed terrible placements and am running $50 of traffic to each camp (MB & HB). Waiting for results.
Thanks Matuloo
Originally Posted by mrpayne
Originally Posted by nirvana
Hi everybody,
After spending $50 more on each HB & LB, here are some results. I did also remove some bot placements (after my 2nd update) before running this traffic.
Campaigns
HB
MB
Applying statistical significance
HB - 97% chance of L3 being winner. Only keeping L3.
MB - No winner defined. L3 is a loser. Cutting it. Leaving the other 3.
My plan now
1. I will be running $50 more traffic to HB and MB before making placements cuts based on profitability. Hope to see a clear lander winner for MB.
2. Creating more versions of existing landers before running the next $50 traffic to HB and MB. I feel I can do better. Will add 4 more landers with fresh angles and visual edits.
Questions
1. Any other cuts or optimizations I should be looking to do?
2. Did I miss anything?
3. Is my CTR normal? 8% seems low to me.
4. Is $50 for each campaign a good number for more data at this stage?
5. When I create new versions of the lander, do I throw them in with the current campaign landers or should I start a new one?
Cool fact
Originally Posted by matuloo
Currently running more traffic. Will update once I see solid stats.
Following. Looking forward to the update. You'll get there
I have been pushing different offers as well. Tried to optimize landers, split tested between different variations, cut placements, etc.
After all that, my AM suggests after coming back from holiday today (appflood) that I should not push these offers as they cut payments. Basically steering me away from the utility apps.
Well, I tried to promote a battery app with a virus angle. Not sure if that was a good start anyways. Lol, I'd love to hear your comments on this one.
I am looking for direction at this point. I have made my mind that I want to stick with av offers with pop traffic.
A few questions
1. Should I explore pin subs?
2. Did I make a mistake by picking USA as my first geo?
3. Any geos that are good for starters?
4. If not av, what offers should I look into?
I invested nearly $500 in 3 days. Lost most of it. I got a ton of data and I feel I understood some of the basics for a cheap cost. I want to take a moment and thank everyone who has given any piece of advice on any of my threads.
I am no expert in optimisation, but some of us are looking for such results as yours at our start
Please show some more details on your campaigns.
By now you should be able to determine which Carrier/ISP is performing better, which OS, which placements should be cut down etc....
@platinum - Thanks for your comment. Regarding details, when I optimize to fine details.. roughly 9k traffic remains. Its not worth running. Btw all that at $7.5 cpm.
I have also posted a JV thread. In case you are interested, click here.
Pins are still good to go...AV Pins, Adult Pins, Pin Sweeps (only in a few GEOs)
US is pretty competitive and expensive, not the best GEO to get started. Talk to your account manager on traffic source and affiliate network about GEOs that are currently hot or picking up. Then choose 3-5 and focus on them.
Next time kill the camp sooner, spending $500 at -70% ROI is too much. This will kill your ability to find winning camps in the long run. Testing 10 camps at this rate way will already cost you 5k.
You need to test lots of camps to find the golden GEO/Traffic Source / Offer combination. Rather test 50 at $50 each.
Originally Posted by sebastian_r
I agree. The money was invested before it hit me.
Will definitely look into some teir 2 and 3 geos now. I like the $50. Now is that $50 all together or just initial test?
Nirvana, basically you want to have a process that you follow for all of your campaign launches.
One campaign test for me looks like this... 2-3 landers, 4-5 offers and 1-2 sources per test in a single geo/vertical. That would be round 1 and it may require more or less than $50, that depends on the competition and the offer payout. My goal is for the best converting segment, such as best offer/lander and either Android or iOS, to be performing really well.
If I see positive signs from that first round of testing, I usually only keep the best target settings (Android or iOS, sometimes just a single carrier or ISP depending on overall traffic volume) and continue testing by narrowing it down to 1 lander in the initial test, adding in as many offers as I can find and then testing until I find the best 1-2 performing offers by reaching statistical significance.
Next, I take the best 1-2 offers and test as many unique landers that I can find or come up with on my own in some cases. Run that test until I reach statistical significance. By this point, you should be in the green or very close to it.
Optimize the campaign by blacklisting poor performing placements, etc. And profit. Then I setup a new campaign to test the segments that I previously cut early on and target them with just my best performing offer/lander, retest it all and optimize if there's potential.
Andrew
Originally Posted by mrpayne
One campaign test for me looks like this... 2-3 landers, 4-5 offers and 1-2 sources per test in a single geo/vertical. That would be round 1 and it may require more or less than $50, that depends on the competition and the offer payout. My goal is for the best converting segment, such as best offer/lander and either Android or iOS, to be performing really well. - Understood
If I see positive signs from that first round of testing, I usually only keep the best target settings (Android or iOS, sometimes just a single carrier or ISP depending on overall traffic volume) and continue testing by narrowing it down to 1 lander in the initial test, adding in as many offers as I can find and then testing until I find the best 1-2 performing offers by reaching statistical significance. - So you find the top lander (s), add more offers to the bunch, and then finally os targeting, got it.
Next, I take the best 1-2 offers and test as many unique landers that I can find or come up with on my own in some cases. Run that test until I reach statistical significance. By this point, you should be in the green or very close to it. - Clarifying, at this stage.. I throw new unique (different design, text, etc) landers into the campaign with the existing top lander correct?
Optimize the campaign by blacklisting poor performing placements, etc. And profit. Then I setup a new campaign to test the segments that I previously cut early on and target them with just my best performing offer/lander, retest it all and optimize if there's potential.. - Got it. Great tip.
In reference to your question, yes.
But to address point #2 above - read the below statement. I just look for what is converting the best right away, focus on optimizing and improving my campaign around that segment of traffic and then later revisit the targeting that I cut early on to retest and check if it works now that I have the best offer/lander identified.
The idea is to take 2-3 landers that I've had success with in other geos or similar vertical, translate those and use that in the beginning. Of those 3, quickly find the winner and focus on mass testing more offers to find the best 1-2 offers. Then revisit the testing more landers that are completely unique to each other and run them against your best 1-2 offers.
By that point you are isolating the best offer/lander. It's also helpful to testing 2-3 sources if things feel like they have potential and you want to be thorough with your testing.
Andrew
Originally Posted by mrpayne
Thanks Andrew.
No system for preparing landers, just grunt work. Rip, clean, launch. Once you've tested enough campaigns, you can keep the best ones organized for future use and faster launching.
Andrew
Originally Posted by mrpayne
Home > Mobile > Follow-along Campaigns