Home > Mobile > Follow-along Campaigns

---

Mastering Mobile Pops and Sweepstakes (43)

---

Here's the deal.. I got my start in AM back in 2005. Over the next few years I learned a lot, networked at all the events and had some successes along the way but nothing huge or life changing. After a while I began taking on clients with an agency format where I manage their website and online marketing strategy. I now have a small team of freelancers that work for me in my agency but my passion, ambition and goals have always aligned with what AM offers.

My goal and objective is to transition away from offering marketing services to clients and focus on my own projects, AM and media buying. I had time to jump on here about a year ago, catch myself up on what the industry is like and started launching campaigns. Spent a few thousand and learned a lot but didn't carry through with it as my existing business started to require more of my time.

I've been back on STM and launching campaigns for about a month now and I need guidance - straight up, I want to make this work and I need to figure out what areas are holding me back and learn a real foundation.

Last year I tested various verticals and a few traffic types. Mobile pops and sweepstakes has been the combo I've had the most traction with so far and want to stick with it until I make it profitable, learn a process so that I can replicate the building of a profitable campaign and scale.

My initial goal is $500/day revenue, 50%+ ROI in the next 30 days.

Over the last 3 weeks I've tested 5 campaigns, focusing on tier 2/3 markets and all have been in the sweepstakes vertical.

Blended Stats (over the last 3 weeks from all campaigns / testing)
Total spend: $2,636
Total Rev: $1,086
Balance: -$1,550
ROI: -58.79%


What is happening now...

Currrently I have a mobile pops / sweeps campaign running in 1 geo and focusing on 1 carrier. In my initial testing I identified this carrier to be converting better and chose to narrow my focus and work on making it profitable, then revisit other carriers and wifi traffic.

I started to find certain pockets within this carrier that converted better than others so I broke those into a separate campaigns (one campaign targeting a specific phone brand and another for specific placements). All of which are very small volume on the existing traffic source but I'm hovering between breakeven and 20% ROI on those two.

PopAds is the traffic source I started about but they have very little carrier traffic for the geo/carrier I am targeting. So I set off to find more sources to test until I find one with more carrier traffic for this geo, it's not a tiny geo but it's not a huge one either.

Right now I have 3 campaigns setup on PopAds and several more setup across different traffic sources all of which are low volume. I've identified one source that does have greater volume and after running about $200 through it yesterday, it did convert for me but I was down -75% at a high bid. So it's running today at a much smaller bid and has been at breakeven all day today at a much smaller volume of course.

Here is a screenshot of today's stats so far:




Here's what I've tested so far...

1. I ripped several landing pages from Adplexity, tested them and kept the winner as it was converting well for the carrier that I'm targeting.
2. Tested a few variations of the lander by adding vibrate, audio, etc. Took the winner and have been running with it since.
3. The primary offer I've been running is converting well for me, I've only found 3 other similar offers in the geo and 2 are brokered from the network I'm already running the offer from and the 3rd offer is a much smaller payout that doesn't allow pop traffic.
4. I've tested 6 traffic sources so far, on one I wasn't receiving enough conversions to keep it running. So I've been running several small campaigns just to have some volume coming through. Note: the biggest source I've found does convert at both high and low bid levels but I took significant loss at the high bid yesterday.


What I'm Doing Now...

1. My AM is about to bump my payout, so that will help. Currently pushing 40-60 leads a day.
2. I feel like I should test more landers since I've only been running with the one lander I ripped and modified.
3. I also think I should test more angles but I'm uncertain just how creative I should get with the angles. Right now the standard "Congrats, blah blah blah.." is what I'm running. It seems the general feedback I read is to stay close to what is working (based on what you see floating around the spy tools) yet, what can I do to differentiate my lander and angle from others? Testing my headline and angle may be the most impactful test on CR, which I've yet to fully test.
4. I've been retesting my lander speed and it seems slow when including the Voluum redirect. I'm compressed the images to the max, the size is only 156kb, I'm running on Rackspace Cloud Files, using Google javascript link. See the screenshot below. I'm not sure how to control "First Byte Time" and "Cache Static Content" I thought that would be handled by the CDN. How does DNS speed effect campaigns or what can I do to improve that?


5. The newest traffic source I've been testing does have much more volume for this carrier than all my other sources combined, so if I can improve my CR and combined with a higher payout, I should be able to increase my bids and test with more volume.


Biggest Questions I Have

1. Should I be testing other angles besides the "Congrats...." style? Just how creative should I be thinking here? Or just test variations that are along the thinking of the existing angle.
2. On the smaller campaigns that are just targeting a phone brand or specific placements, if I day part the traffic literally halts during the day when it should be running. Any other options to control the loss during the poor timeframes?
3. I tested one redirect campaign off of a completely new source and it converted well but was expensive and seems to be volatile so I've stopped running that campaign for the time being until I increase ROI much more. Is that the right approach?
4. Also, what are tips to optimize a campaign that has 50+ pages of traffic that are all low volume so far? I've gotten many 1 click 1 conversion placements so far and a group of placements with multiple conversions.

I'm also interested in forming a small group to mastermind with and/or a more experienced mentor.

More updates to come. All feedback is welcome.

---

I may have missed it, have you already mass tested offers with your best lander/geo?

---

@bandito - Thanks for the response.

I've only found 4 offers so far that are similar.

Offer 1 is the one that I'm running.
Offer 2 & 3 are the same offer I'm running but are brokered from the network that has the main Offer 1.
Offer 4 doesn't allow pop traffic.

I have accounts with 12 affiliate networks and have explore all the options for this offer at this time between these accounts.

---

That is a slow lander for pops.

You might want to combine all those photos into 1 photo and use sprites - That should reduce the amount of requests.

I can see that one of your file names are phonespin so I can only assume it's the spin-and-win sweeps lander. I've used that lander before and have gotten the speeds below 500ms.

---

@glennb - Good point. I'll do that now.

I will note that the above speed test includes the Voluum tracking link and not just direct to lander, so some of that time loss is the redirect.

---

Ok, so here is an update for today.

The day before yesterday I started a campaign on a new source that I found has alot of traffic for my targeting settings. I set the bid high and spent about $200 over the course of the day. It did get 32 conversions but an overall loss of -75%.

So at the beginning of yesterday I lowered my bid to a rate that I estimated would make me break even, which it did but as expected, it cut my volume by around 60% in both traffic and conversions for that one campaign.

Here is a snapshot of overall results for Aug 2nd:




Like I mentioned a few posts back, there aren't many offers for this geo/vertical. The 2 offers I did find have just been brokered from the original network that I've been running the offer with and the original network has the highest payout. So my assumption was that no way would these other two networks be comparable in performance because A) they are rebrokered, so more redirects, etc. and B) the original network provides a higher payout (and my AM is about to bump it even more).

However...... I decided to test one of the other networks running the exact same offer with a lower payout and you can see the results below. So that makes me wonder, is the main network I've been running this offer through scrubbing leads? Because this other network seems to be converting much better. I've understand it's wise to test the same offer across multiple networks but you would think that the primary network would be the best. I guess there's no room for assumptions.

Offer Split-Test



I know my load times seem to be slow. From various testings, it loads within 1.5-3 seconds, including the redirect time for the tracking link and testing it with the dotcom secrets tool set to a android phone. Which baffles me because it's fairly small and I have a well rated CDN for this geo being used.

Nonetheless, I have my developer reconfiguring certain elements. Combining everything into a single sprite file, correcting any W3C validation errors and creating a couple variations that function slightly different in order to improve conversions.

That test will go live around later today.

On one traffic source I've been monitoring the adspend and placements. It seems I have to regularly pause certain placements that are getting 50+ impressions and 0 clicks. I've noticed that any placement that has less than a 5% CTR has not been profitable at all. It needs to have 10%+ CTR's to be profitable. So I've been pausing placements that reach 50-100 impressions with 0 clicks unless their spend is just a few pennies.

Questions:

1. I have not done a bot test yet, should I make that a priority on my items to test?
2. Should I mention the difference in CR in comparison to the other network to my affiliate manager? I say this because I'm not quite to the threshold to be paid out yet so I'd like to reach that, he's increasing my payout on the offer but if the CR is half of the other offer, I'm not sure how much that will help.

---

Another update.

As glennb suggested putting all of my images into a sprite because he was able to get this style of lander down to 500ms loading time. I had my developer combine all images into a sprite, compress the image once again, fix all W3C validation errors so it's clean and clear. That did help a huge about but not down to 500ms, I think partly because of the geo I am targeting has very modest mobile/wifi speeds.

Results of further optimizing the lander:
Original lander loading through Voluum link: 3.03 seconds and 21 requests
Updated lander loading through Voluum link: 1.25 seconds and 4 requests




This updated lander is running today and the results will be posted tonight.

---

Running traffic through the faster lander to the best performing offer (not being scrubbed?) - Super interested to see the next round of results!

---

Here's an update:

So I tested the original lander (v4) with the update version (v5) that is optimized very well at this point, in terms of size, speed and valid code.

I also ran just the best converting offer after yesterdays test, which is at a lower payout and from a different network that I was with previously.

The results were a bit weird today, as I didn't get as many conversions as normal and overall ROI is lower. I have noticed that several of my campaign, although from different ad networks, are converting on similar placements / categories. In particular, one campaign has received 46 out of 53 conversions from 4 specific groups of placements, all of which are profitable.

Here are the overall campaign stats:




Here are the stats on the lander test, that I plan to continue gathering data on tomorrow. The test didn't start running until mid morning so I've only included data while both landers were running at the same time.




Here's what I intend to do next:

1. One of my AM's from another network is attempting to get the offer so I can split test it with them.
2. I'm going to continue running the lander test to collect more data.
3. In one campaign, I've identified 4 groups of placements that are generating 90% of the conversions. So I'm setting up a separate campaign to target them specifically and possibly use a more aggressive bid.
4. I've been planning and gathering offers to run tests in more geos.

---

Update for Aug 4th

I took the best converting placements, that were all profitable, and placed them into a campaign of their own. I slightly increased the bid in an effort to try to get more traffic from them. I may try to increase it a little further to see if volume picks up but that campaign turned a 43% ROI today.

The other campaigns are lagging behind and I've dropped all but the top two campaigns going forward. Managing several small campaigns is a waste of time/effort. Even the two larger ones I have are pretty small.


Today's Stats:



Landing Page Stats:

Despite v5 being valid code and much more optimized in general. They both are performing very closely.




Whats Next:

1. Since this campaign is focused on 1 Carrier/Android only and I've found a source with decent volume for this geo, I've launched a few new campaigns. One targeting all of the other carriers and all OS's and a second campaign targeting wifi and all OS's. Revisiting these targeting settings on the new traffic source while using my best lander/offer combo should allow me to guage if there are any other pockets with potential.
2. I'm also preparing to launch sweeps campaigns in two other geo's.
3. When checking my best lander on browserstack, the spin wheel portion doesn't look right on certain phones. I'm not sure if that is just an artifical hiccup from the browserstack emulators or if its an actual issue. I check several landers on Adplexity, most have the same issue but a few do not. So I'm having my developer revisit this.

Questions:

1. How often should I be optimizing and making adjustments to my campaign? I understand to remove any placements that do bring clicks or have very low CTR in comparison to all of the targets that are converting. I just want to make sure I have a clear understand of atleast one approach to optimizing. One method of cutting only the worse metrics, I'm assuming that should be done once a day unless I'm pushing alot of traffic through there. I've also read that cutting OS/Carriers/Models all at the same time but I think that approach was intended for display, seems like its a fair approach for pops as well.
2. Something I've come to realize after reading a few threads over again, I don't necessarily have to get my campaign to break even before cutting placements. On iAmAttila's case study, he mentions only bringing it to around -30% ROI and then doing a few rounds of cutting which brought it to a positive ROI. Again, I assume that really only works if the traffic volume is sufficient for that geo/target settings.

---

That sounds like an attainable goal! Tier 2/3 sounds great, and sweeps are hot right now.



Several thoughts here:


1)Unless there's a TON of traffic for your target geo and carrier, it probably wouldn't be worth your time to set up separate camps to target small target segments - such as a specific phone brand (unless it's samsung or apple). Do keep in mind that camps will require time to manage and optimize - don't let a bunch of tiny camps making a couple of bucks a day each tie up all your valuable time - time that you could be using to test and find promising offers, that you could scale to double/triple/quadruple digits in revenue.


2)When testing a new geo, try to rip as many different-looking landers from adplexity as you can (5-10+). Then use at least 2-3 offers to test these landers to identify a winner, then use this winning lander to test every similar offer you can find. Alternatively, you may choose to mass-test offers from the very beginning, depending on your preferred testing approach.

In your specific case: It sounds like you've tested quite a few landers already, and you said you've tested all 4 offers you can find for that offer type and geo. If you've tested staggered bids (at least 3 - low average high) on a couple of traffic sources, and you're not seeing any offer+lander anywhere near breakeven for at least one major traffic segment (e.g. one of the major mobile carriers/OS/etc.), then it may be a good idea to just test other geos.


3)That new traffic source with the high traffic volume does sound hopeful! The more traffic volume you have to play with, the more optimization options you'll have. For example, even one of the smaller carriers or OSs could result in significant profits when you have access to larger traffic volumes.


4)Testing new lander angles can be a good idea, but keep in mind that with all the spy tools around, your new lander will get ripped in 2 seconds. You'll need to be creative and figure out a way to keep your original lander out of the spy tools as much as you can in order to maintain that edge for a while. I'm not technically-inclined myself so I partnered up with people that are. This thread would be a good start if that's what you're wanting to do:

http://stmforum.com/forum/showthread...s)-Part-1-of-3


5)As for pausing the redirect camp until later after your ROI improves - good call.


6)A loading time of 3 seconds is appaling! You need to cut that down for sure! Caurmen has already proven to us that at around the 2-second mark, visitors will start leaving your page in droves.

Have you read all the threads on how to make pages faster? Browse the links in the thread below, under the section "Try decreasing your page load speed!"

http://stmforum.com/forum/showthread...-AM-FAQ-Thread!


7)Regarding the 50+ pages of placements - this is simply telling you that you won't be able to depend on getting green by cutting placements. If you've tested most of the offers available and are still not seeing promise - move on instead of relying on cutting placements to get profitable.

Pop camps are often short-lived, so try to focus on testing stuff that has the biggest potential impact on your ROI, i.e. testing offers and landers (ESPECIALLY OFFERS!) If your optimization approach takes too long, chances are your camp will die before you can get it green. Test widely, only pick camps that are so promising that they'll get green with minor optimization, then scale as hard and as fast as you can. Repeat as often as possible so you'll continually find new profitable camps to replace dying ones.

Pop has become so saturated with competition, and so volatile, that it's getting harder to harder to make high profits consistently. It's a fantastic traffic type to learn the ropes with, but consider expanding into mobile display, native, or FB when you have a chance. That's the best advice I could give to anyone currently running pop.



Amy

---

Nice discovery regarding split-testing the same offer from different networks! It would be difficult to speculate the reasons why the network that's brokering the offer, is converting better, without making the original network look bad. I can't comment on the possibility of scrubbing, but their infrastructure may not be as good as the other network (which could be resulting in slower redirects for example). But yup - stats don't lie, and assumptions don't have a place when it comes to deciding which offer version will convert better.

As for placements that have really low CTR - yes that can be a sign of fraud traffic. Cutting based on CTR is a bad idea in a vast majority of cases, but when the CTR is unrealistically high or low compared to those of its counterparts, it may be a good idea to cut those placements.

You can do a bot test for sure - and if you do I would suggest to focus on the bigger placements and ignore the smaller ones for efficiency reasons.

However, I don't feel that the bot test will make or break your campaign as it stands. At your current CR and bid you're not getting a lot of traffic volume. You'll need to somehow increase your CR in a big way to make this camp even worthy of pursuing further. You've already tested all the offers you can find. Let's hope that your now much-faster lander will convert significantly better. A pay bump would help as well. Maybe the 2 combined will push you into green - guess we'll see!

You can sometimes get a good bargain by making 2 networks "bid" for your business. However, like you said, if one version of the offer is only converting half as well as the other, then I highly doubt your AM would give you a big enough bump for the offer to have a chance at winning out over the offer from the other network. If cashflow is tight and you really need to reach that payment threshold, consider running at a loss until you reach it, and then switch to the other network.


Amy

---

I'm really impressed with how you were able to bring the lander load time from 3s+ down to 1.2s. However, I'm REALLY surprised that the faster lander did not do significantly better than your original lander!

All your "What's Next" sounds good.

The question "How often should I be optimizing and making adjustments to my campaign?" isn't an easy one to answer. This would depend on way too many factors.

As for cutting placements - there are so many approaches you can take depending on the particular camp and your preferences. If you don't have a lot of traffic to begin with, you really won't want to cut placements while you're still testing offers and landers - in that case it would be better to try to do your reasonable best at finding a good offer and lander before cutting all placements that aren't profitable in spite running the best offer+lander. As per your assumption, If you have a lot of traffic volume, you'll have a lot more leeway when it comes to optimizing the camp, for example cut placements very aggressively and STILL have lots of traffic left after your cutting spree.

Really looking forward to seeing how your new geos will do!


Amy

---

Quick update for Aug 5 results

Spend: $178.14
Revenue: $105.00
ROI: -41%

Things to note:
1. I ran a bot test that cost me $8.50
2. One test campaign spent $8.38 with my best lander/offer combo and make 0 conversions so I paused it.

Thats about $17 of my losses for the day right there.

Today has been the most revenue I've have generated with this particular campaign thus far. The reason I had more volume today was because I relaunched a new campaign targeting all of the segments I previously discarded because now I'm on a source that has more volume for my geo. I set the bid relatively low and a few hours into the campaigns I had 10 conversions all around a few specific areas, so again I discarded the worse performing metrics and let the campaign run the rest of the day with just the best. It produced all 29 conversions and helped the campaign finish at -32% ROI.

Going forward, I've consolidated my efforts into two campaigns, one with the best wifi target settings and the other with my best carrier target settings. This should allow me to generate 20+ leads per campaign daily and help me split test faster.

Today's Results:




Lander Stats for the Day:


One thing to note, v3 is for the campaign with more broad traffic. The other landers, v4 and v5, are slightly different and are tailored to the original segment I have been targeting.


@Amy - As always I very much appreciate your insights and feedback. I have experienced a lot of volatility in pops as well, even on other campaigns that I felt had more potential than this current one. It's likely due to the low barrier of entry for pops and the ease of not needing banners / ripping landers, which is not something I'm afraid of. I've had a desire to jump into mobile display for a while and I have been in discussions with some of my AMs, already grabbed offers and I'm preparing several campaigns to upload over the weekend.

That being said, I intend to continue improving on this campaigns if only for gaining more experience.

What's Next:

1. Out of the new campaign today, I've only cut placements that recieved 100+ impressions and 0 clicks. Going forward I intend to be careful of what I cut and focus on improving my lander until I'm consistently hitting break even or better, then cut more. Just from today there were 3,000+ placements that I received traffic from.
2. One of my AM's got the offer I've been running and was able to match the higher payout I was previously receiving. So I will be able to test if running through that network will have any impact on performance.
3. I plan to launch a mobile display campaign around the segments that are working the best for me on pop, perhaps there's less volatility and I can make it work better.
4. With insights from one of my AM's and with an offer recommendation by her for mobile display, I have ripped a few banners from Adplexity and had my designer create 20 more that are similar. Over the weekend I intend to setup test campaigns in several geo's as this in an international offer. I will also reach out to a few other AM's for recommendations to test.

Stay tuned.

-Andrew

---

Sounds like a plan!

Just wondering: Did your bot test uncover any placements with high percentages of bot traffic?



Amy

---

Good question and I meant to speak about it in my last update.

I did uncover a few very bad placements but I'm unsure of how long I should run the bot test for?

The majority of the placements had a CTR of 40-60%, a good chunk only had 20-30% CTR, a small portion had less than 10% CTR and the overall bot test campaign finished at a 40% CTR.. so that means roughly 60% of my traffic is bots, if I interpret that correctly.

Here's a snapshot of some of the placements:





Questions:

1. Should I run the bot test longer or just expect that the actual price I'm paying for traffic is double my bid due to so much bot traffic? (which would explain the really low CR I have in comparison to other campaigns for this geo/vertical)
2. I did check to see if some of the placements that had 35% or less CTR in the bot test had actually converted yesterday and out of the small sample I did find a few that converted. The majority of placements had not converted but many had only generated .10-.20 cents in costs because of such low bids.

---

Update for Aug 6th

After my bot test on the 5th I initally just removed the most obvious placements that were getting bot ridden, approximately a dozen or so of the highest volume placements. I kept the remaining ones because I have recieved a conversion on a few placements that have 1-2% CTR and 1-3 clicks in total.. but are profitable. I think that helped the ROI a bit as you can see my best lander has improved from yesterday and there was no other changes made at that time.

However, after running more traffic today and reviewing the stats, I decided to change that strategy. I paused a total of 232 placements so far that have received 40+ impressions and 0 clicks. Then I also paused all placements that were less than 3% CTR. After reviewing my stats I noticed that every conversion I've received, except for 3, has come from placements with 10+% CTR and those 3 that did not only had a CTR of less than 2%. I'm bidding very low, which is why I'm probably getting so much bot traffic and instead of spending money to find which bot placements still convert, I have cut them off.

So tomorrow I expect two things, an improvement on my lander CTR, CR and ultimately ROI due to removing those bottom feeders.

Today, I split test my best lander with one with a few changes - obviously my ideas to test completely tanked which was interesting because that particular test has proven to improve most of my other campaigns.

Here are the results: (I switched to my best lander at one point during the day because it was clear of the winner)




That was the only campaign that ran today because I have consolidated my other 3 small campaigns into one and awaiting its approval.

What I'm Doing Now

1. I have a new lander test setup for tomorrow.
2. I have my translator reviewing the copy on my existing best lander to ensure there aren't any mistakes and I have a few variations of the headline that he will tranlsate so I can test.
3. My designer has created 40 general banners, a mix of the 3 most popular sizes, that I will be modifying and using in my mobile display campaign.
4. I did some spying today but having a hard time getting an android emulator that allows me to download and use apps so that I can spy for my mobile display approach. If anyone has any feedback on an straight forward way to set that up, let me know.
5. Finishing up what I need to launch the mobile display campaign.


Questions:

I have an offer that is hot and converting right now via feedback from my AM. I'm using the Main Course approach, I did my research, created 28 angle ideas and have chosen two angles to begin with. I have my ads almost finished and will be running 3 banners for 2 different sizes.

1. The Main Course mentions to run each angle in their own (4) campaigns but I believe I read feedback from Amy mentioning that you can run more than one angle in the same campaign. Whats the best approach to testing many angles at once to determine which is best?

2. Is it wise to also test direct linking if its a app install offer?


Stay tuned!


Andrew

---

In caurmen's thread on how to run a bot test, he suggested 30-50 clicks per placement should be enough:

http://stmforum.com/forum/showthread...n-Any-Campaign

What I do when running a bot test, is I would assign a really low test budget to the camp, run it until the budget runs out, then check placement stats and pause all the ones that have received at least 50 clicks, then allocate another couple/few bucks to the camp and let it run and check and pause placements again, and repeat - until I've gone through all the biggest/bigger placements. Then I would identify the really bot-infested ones, so that when I'm running the camp "for real" I'd know to blacklist those from the start.


Very nice! Looks like you've uncovered a few big placements that are bot-heavy.

Based on my own experience, I've found that some of the most profitable placements can contain up to 50%+ bot traffic. Go figure. I guess the human portion of the traffic must be top quality such that even with all the bot traffic, the placement's overall performance is still good.


I would suggest to just cut the really bot-heavy placements - such as ones with 80%+ bot traffic as suggested by caurmen - and then cut the rest based on performance alone.

At the end of the day, all you care is whether a placement is in profit or not. The bot test is useful for weeding out the worst right off the bat, but after that, you could just cut placements based on whether or not they're in profit/loss.

Thanks for sharing your results!


Amy

---

This kind of observation is so valuable that it's not just gold...it's PLATINUM!

Usually no matter whether you're bidding high or low, you only get one advantage: Either low traffic price, or high conversion rates.

What you're telling us based on your stats and observations is that, by bidding low and THEN eliminating the bot-infested placements, you're getting the best of both worlds!

I'm gonna try to set up some tests to test this further at my earliest convenience. Thanks for the amazing tip!

I've used this android emulator before:

https://www.genymotion.com/

If you ask around, your family or friends may have an old android phone they're no longer using, that you could use for spying/testing purposes.


Either should be fine, but if you're following the mobile cookbook then I'd suggest to follow it to the "T" for now. No reason to deviate from a great recipe, especially when you're cooking for the first time!


Yes!



Amy

---

Just a short and quick update...

I've been testing so much this week it's crazy. I've launched several campaigns on Go2mobi and about a dozen pops campaigns. I don't have the time to organize and post all of the various screenshots but I have a couple campaigns that are showing strong potential. I plan to make a more thorough update in the next day or so with my progress.

I don't think I had a full realization of that being an amazing tip but since your comments I've been rolling it over more in my thoughts. I definitely think it works for certain traffic sources / geos. The particular source I started this campaign on has alot of volume but I discovered (after a more thorough bot test) that this source has a TON of bad or average placements. So the method of bidding low has been great for revealing most of the bad placements early on without breaking the bank. Using this method I have been able to improve the ROI from -83% ROI to -52% ROI and then after testing more offers today I discovered a better converting offer that has brought me to -44% ROI.



Stay tuned for more.



Andrew

---

Another gold tip - although many newbies may not recognize it as such!

Having read so many follow-alongs on this forum, and having spoken to so many people in this industry, I would say that having the willingness to launch many camps to test many offers and angles is one of the main keys for success in this industry. Casting a wide net and optimizing and scaling ONLY what has the most potential is MILES BETTER than taking a single offer and some random creatives and spending a ton of time trying to optimize it to green.

This is such a basic concept, and yet it seems to be such a common blindspot. I do my best in encouraging people to test more offers - but sometimes it feels like I'm pulling teeth. The way how most people would rather analyze and cut thousands of placements than to test a few more offers puzzles me...

Sorry don't mean to come across as complaining - it's always exciting to see someone willing to do massive testing to uncover the gems!



Amy

---

Here's an Interesting Update

@Amy - Thanks for the feedback. I am somewhat of a newb, atleast to mobile AM in general, I've had good success back in the day with Adwords AM like 2005-2009 but never learned to make it consistent, so in some regards I have a strong skills set and knowledge. I usually can pick things up quickly. But I very much appreciate your guidance.

I've been working around the clock these last 2 weeks, literally 16-20 hour days, only because I've had no client meetings to tend to. After this post I am going to grab a few hours of sleep, its 5:50am already here.

Summary of the last few days:
To summarize what I've been doing over the last 5-6 days, testing, launching, testing, launching, launching, testing.. you get the picture. I'm not kidding when I say I've ripped 200+ landers from various geos/verticals and currently have tested atleast 50-60 of them. (see below for some screenshots)

Since jumping into mobile pops I've mostly focused on sweepstakes as it was easy to get conversions but it has been very difficult to get a campaign better than -30% ROI, and even that has been limited volume.

So I asked some of my AMs for a suggestion and one of my newer AMs recommended a few, some were app installs and some where sweeps. I've setup tests for many of them, most were flops or I couldn't get to convert.

I launched a few campaigns on Go2mobi but those didn't show any conversions at this point, however, some of the campaigns I've been testing on pops has started to show real promise.

I have been thinking much more on how I need a real process and not just aimlessly make decisions with no real focus or clear thought to it, most of which is due to lack of confidence in opimization. Maybe you can related, but I know myself and how easy it is to become emotional when you are spending and losing $100-300/day while testing so many things - it wears on you, it adds up quickly and you make choices on your campaigns that actually work against your efforts.

What you're reallying looking for:

Initially I launched a campaign with 10 landers and the same offer on 2 different networks.



The results of that overall campaign were at -83% ROI with the best offer/lander being around -63%, I was going to consider it a flop but with the whole "have a process" echoing in my mind, I've been studying up on my optimization process through STM. I stumbled upon this guide by matuloo, reread Amy's guide and reread other content on here:

How To Optimize a Campaign

Pop Campaign Optimization - Some Considerations

Seeing as the offer my AM recommended is an App Install and fits the type of offer that matuloo said his guide fits, I've set on a course to stick to that optimization plan until I get a better grip. To be honest, most of this information I already know but the actual putting it into practice and remaining consistent, focused and unemotional is what I've needed to do.

First step, I ran a bot test and eliminated those. I took my own approach to the bot test but it fit what I was doing. That brought my ROI up a bit. Next, I found the same offer on two other networks and found one network that converts much better than the other two! That brought my ROI up as well. In the screenshot below, the first row is the newest offer that converts the best, the second row was the previous best offer.



Last night I found 3 other networks that I have an account with who have the offer but I had never run traffic with them. This morning I tested all 3 of those offers with my current best and found one that converted a little better than the previous winning offer. I'm aware that I should have tested all of these at the same time to same money and not do three rounds of offer testing, going forward I will be more strict about that part of my process.



Next, I setup and tested another 9 landers with my current best and found a new 'better' lander/offer combo that brought me to -49% ROI but I was telling myself, after all of this testing why am I not closer to break even!




So I took my best two offers, prepped 30 landers (3 new geos with 10 landers each) and setup the campaigns. Started up one campaign in a new geo, after $50 spend and a couple hours later.. here are the preliminary results that have held true since the campaign launched...




What has me anxious about the initial results of this campaign is that I've only blacklisted 8 placements that are bot-ridden, I have this campaign targeting all Android traffic for the geo I am in.. I've not cut or limited the targeting in any way at this point.

Stay tuned for more.


Andrew

---

You're doing well Andrew, looks like you are getting closer to the final destination

Looks like you were able to track down the best offers, you have 1 winner for now, but I wouldn't let the other one go completely, it's not that far behind.

The game changing point for you seems to be one of the landers that really outperformed the rest, it would be awesome if you could come up with one more with similar performance.

Keep collecting data for some more, let's see if it holds the performance. If it does, you can focus on the targeting and grow the ROI where you want it to be

---

Nicely done Andrew!! I must say you're doing all the right things here!

You have my utmost respect for hustling so hard - but please don't neglect your health. I'm guilty of the same, but lately I've started to rest and exercise more - because I realized 2 things:

1)My health has been on a steady decline since I started in AM and I want to do something about it before it's too late.

2)When I neglect my body, my energy levels would dwindle - which has made it more difficult to focus on my work.

I'm definitely more productive when I'm well-rested, eat and exercise regularly, and take time out to meditate.


I hear you. Having a basic optimization system will take some of the emotion out of your decision-making. And once you have a few profitable camps running, you'll feel a lot better about spending money on testing - you're SO close!

As far as optimization strategy goes, the key would be to focus on the big things first and then tweak the smaller things later, i.e. improve your funnel as much as possible first such that you manage to get a large enough portion of the traffic profitable, and THEN cut whatever is still not profitable while you're running your best offer + lander + targeting.

Which is exactly what you're doing here. You're a true master. Being willing to rip and test 100+ landers commands respect no matter how you look at it. Next: Do the same with offers!

What I would suggest at this point, is to scour all the affiliate networks you're with and compile a list of ALL the offers you can find for the offer type and geos you're running in now, and use the best lander to test them all.

And don't feel that testing a lot of offers will require a big budget either. What I would often see, is that a couple of the offers would make enough leads early on to enable me to cut a lot of the other offers. The better your funnel is, the less cutting it will require to take your campaign to 30%+ ROI or whatever is the minimum ROI you deem will be worth your time in maintaining the camp. Also, your scaling efforts will be a lot more satisfying - the better your funnel is, the more traffic sources you'll potentially get profitable on when you scale.


Anxious? Haha - some of the placements you've identified are big ones that were responsible for a lot of your traffic! You got your money's worth out of that bot test.

Why are you just targeting android? Is it because your offers only accept android traffic?

As for drilling down to traffic segments: The big one would normally be OS, but you're targeting only Android. Others to drill down to would be devices and carriers. And maybe browsers.

After another round of offer-testing, the other major things to monitor and cut would be placements and hours-of-day (dayparting).

Testing different bids again at that time would be beneficial as well.


I definitely agree with you that pop traffic is easier to work with because you don't need to mess with banners. However, display camps can last longer than pop camps, and they're less volatile when it comes to conversion rates.

A good strategy would be to test offers on pop first, then port the successful ones to display. Although that can be tricky because display networks tend to be a lot more strict when it comes to requirements on creatives - they tend to be a lot less tolerant about aggressive/misleading tactics. So some lander adjustments may be necessary.


YW! I dappled in adwords in 2008 as well! It was crazy easy to make money there back then. Them good ole days...


Amy

---

Another Update

So I let the campaign run with just the best offer/lander combo for a few hours while I stepped away to sleep. After spending over $90 is it still in a positive ROI. But with the advice of mutaloo and Amy, I have more testing lined up for the day.

What I Will Be Doing Today:

1. Run more traffic to both my best lander and top two offers.
2. Run more traffic to my second best lander to see how it performs over a longer period of time and volume, it's not too far behind my best lander.
3. I'm speaking with my AMs and researching similar offers that I can setup for a test later today with my best lander.

@matuloo - Thanks for your input, I appreciate it. And your feedback on the optimization thread. I will keep the second offer around, it seems to be reporting slower than my main offer is but if you look at the screenshot below I can see several conversions came in while I was asleep and it kept a solid ROI as well. I will run more traffic to both offers today. Also, if you look at the screenshot you will see my second best lander is around -24% but I still need to run more traffic to it as well to prove it's stability. Is that a close enough second as you were mentioning on your optimization thread? Should I find new landers to test or test variations of my best lander to find a solid 2nd? I'm pretty sure I know why it is converting better than the others.






@Amy - I completely agree in terms of health. I have worked for myself and from my home office for the last 9 years. I'm very accustomed to the toll it can take on your health to work like I have been recently, it's not a schedule I will maintain for long and I definitely will take a break this weekend. I usually hit the gym 3-4 times a week amongst other things. The reason for my intensity as of late, I either need to go grab a couple of new clients soon or get a profitable campaign in the works (I'm trying to transition away from client work). And I echo the statement on being more productive when rested.

I'm really not too concerned about spending money, it's the losing money that hits you in the gut as with all business investments / ventures, I'm accustomed to the risk factor - it's just managing it sometimes takes a toll.

Over the last 30 days my adspend is just over $4k (not including Voluum, stm, translators, developer, etc) and have made $1800 back, so I'm around -55% overall - but that's about to change.

As for being anxious, I meant it in more of being eager and anticipating good things because my targeting settings are very broad and the best combo is already breaking even, in decent size geo.

Because of the offer, I can only target Android so I've been focusing on countries who have mostly Android traffic.

My settings currently look like this:

Android Only
Phone and Tablet
All WIFI / Carrier
RON

My AM originally suggested that I run this offer on mobile display, so after I get further along with this campaign and scale to other sources. I will try to scale it to mobile display as well and then test other geo's.


What's Next:

1. Per matuloo's advice, I'm going to continue testing both offers, test my second best lander more to see how it performs with more traffic and try to find another similar performing lander.
2. Per Amy's advice I've already started to gather a similar offer from 3 networks and I'm working on finding a few more networks with that offer.
3. I'll reconfigure by best lander and setup my flow to test that lander with the new offers.
4. I will test bids out after I've improved my funnel a little more. I've already done some bid testing previously at different levels, and that if I go up even just a little more my traffic grows substantially and I want to ensure my CR / ROI can sustain the higher bid cost.


Questions I Have:

1. @Amy, you stated testing ALL similar offers.. I know of another offer that is performing well that is similar but should I venture even more outside of that offer but same overall category?
2. After review stats this morning, I noticed that all wifi traffic is at a 39% ROI and Carrier traffic averages out to about 8% ROI. After I do more testing, should I consider splitting these into two campaigns to better test / manage?


Update On the Campaign Stats

Since running the best offer/lander combo through the night while I slept the last few hours, it has recouped most of my losses and I'm now sitting at a -8% ROI overall for the campaign with the best combo being at a 14% ROI on over $100 spent with the campaign.


It's rewarding to finally see some positives after putting in so much work. And a special thanks to those who have continuously contributed to my follow along. Stay tuned for more.


Andrew

---

1)As long as you feel that your current landers will be a good match for that offer.

Ideally you'd want to stick to testing offers that are of the same offer type as your current offers. You've tested so many landers and now have a couple of "best" landers. If you test a different sort of offer, your "best" landers may not work as well with this new offer compared to your current offers.

2)Yes, it would definitely be a good idea to target carrier and wifi traffic in separate camps, mainly because carrier traffic is typically considerably more expensive than wifi traffic. You can leave the carrier traffic in there for now since you're needing the traffic to collect data when testing your stuff. But once you're ready to test bids again I would suggest breaking out the carrier traffic into a new campaign, and testing bids for each.


Woot woot! *breaking out the champagne...*

Can't wait to see you improve your funnel further and then scale!



Amy

---

-24% is a bit of compared to the winning one, it's not a totally lost case but since you say you probably know what makes the other lander perform better, I would try to make variations from it.

---

Quick Update


@matuloo - I figured that would be the best approach. Thanks for the input!


Just wanted to provide a quick update. I've been taking a break today and haven't worked too much. I ran a test last night that took my best lander and my best two offers, ran a bunch of traffic to them to see how they perform since the previous test on the second best offer wasn't too far of from the first offer, as matuloo metioned. His advice was accurate - even though they are at different payouts, after spending $20 on each offer, they are only 1% difference in ROI. So both offers are solid.

Then, as Amy had mentioned, I tested my best lander with another similar offer type that I gathered from several networks. And I found 1 of those offers to convert exactly the same as my other best performing offers, which is great.And I just found out that two networks I'm with have this offer set to private, so I will do another round of testing for this offer group on Monday.


Official Results for Launch Day

I've received late conversions for the Day 1 test, up to 18 hours later and the final results are now below in a screenshot. My overall test campaign made it to the green @ 3.5% ROI and the best lander/offer combo did a 31% ROI with over a $100 spent on that test, most of which went to the best offer lander on a phase 2 of that same test campaign.




What I'm Testing Today

1. I'll be testing several variations of my best lander.
2. After reviewing my techonology setup this afternoon, I realized that my files are not pulling from a region that is setup near my campaign geo. That will be changed and I will test that along with my original lander as a control to guage performance difference.


Questions

1. Ok, so now I've found several offers that convert well. Two from one offer type and one so far from another offer type. What should be my gameplan with those? My thought process right now is that I should continue to use and rotate all offers since they perform exactly the same +/- 1% ROI. Monitor their performance, if one starts to tank/gets pulled, I have fall backs and if certain offers are performing better for certain targeting I can maximize that based on data. This also allows me to run traffic to several networks and build up more rapport with each while diversifying my risk of payout with each of these networks (I've just recently joined each of them and trust needs to be developed on both sides). Am I thinking along the right lines?



Subscribe to this thread to stay updated.


Andrew

---

Wow! 31% ROI!!

Of course testing more offers and landers is always a good idea! But it sounds like your best offer+lander is ready to be scaled! Your plans on testing additional offer versions and landers sound good. After another round of testing you should definitely focus your efforts on scaling in the next little while.

You can absolutely run multiple offers at the same time - doing so will delay saturation and burnout, as well as hedge against risks of one offer being pulled like you said. As long as cashflow doesn't become an issue, I can't think of any cons.

Are you running both offer types on the same landers? Or are you using different sets of landers?


Andrew you're doing amazingly well - looks like everything's on the up and up!


Amy

---

It's the exact same lander duplicated with a couple of small changes to better fit the second offer.



Andrew

---

Great numbers.

---

Great job mrpayne

---

Wow! This thread needs to be printed and framed. The level of detail in this follow along from both you and everyone helping you is fantastic. So much respect to Vortex and Matuloo for their golden advice the entire way through. And so much respect to you for your drive! 233 campaigns in 4 months that's beyond inspiring this was such an exciting journey to read thank you to everyone that's contributed I've learned so much!

---

November 2016 Update

So, I haven't made updates as frequently as I wanted to in November because things have gotten busy for me.

But I have committed to sharing my November recap and to complete a full 6 months of recaps I will post a December recap just after the New Year. At which point I will no longer post month to month reviews of my performance but I will still occassionally post on my follow along and respond to those who leave comments.

My income level is still not at the level I want it to be but I have made alot of progress over the months and finished this month with a higher overall profit that previously. I had found two campaigns that were making me profit, one in particular did really well for a short period of time, but I really didn't launch many campaigns in November because I was traveling over the Thanksgiving holiday and was sick for a few days afterwards.

In November, I hit a new record daily profit of $714 which felt great and you can see my profit by day in the graph below. Overall, the success was short lived but it helped me finish the month with some green.

Snapshot of November Results

Overall Profit: $2409.90
Overal ROI: 41.22%




What's Happening Now?

December is off to a strong start and I've been putting alot of my resources and budget into trying to scale one particular campaign that is doing well for me.

Just when I thought my record breaking day in November was awesome, lo and behond, I experience a new record day just 3 weeks later.


New Daily Profit Record


*Results not typical... atleast not yet! Haha


Yes, I was just as blown away as you are right now reading this. 500% ROI for this day with nearly a $1k profit... this day was great and while I've have a good campaign running, the daily results are not this strong. I've been averaging $300-400/day profit so far this month overall.

So close to that coveted $1k/day profit goal but atleast I'm trending in the right direction.

Hope this serves as motivation for you and you are working as hard as I am to finish out the year strong!



Cheers,
Andrew

---

Congrats! Just curious how are you getting that much revenue with so little spend? Is it because of your bids?

---

I just have the perfect storm at the moment.

Solid lander/offer with quality traffic and a well optimized campaign.



Andrew

---

$914 profit day and average 3-400$ a day and you started this 4 months ago. All possible because of your obsessive drive to succeed. This follow along is giving me the fire I need can't thank you enough for the amazing story Mrpayne

---

So close to 4 figures/day in profits! Very nice Andrew!

It's frustrating isn't it, how pop camps are only profitable for such short periods. But I'm assuming that you've already identified bad placements for a variety of geos on the major pop sources by now, and that by focusing (or least prioritizing) on those geos are giving you an edge.

Are you and Brad still working together? He mentioned wanting to expand into FB. I think that would be a wise thing to do.



Amy

---

Hey Amy, thanks for stopping by!

Yea, I have been playing to my strengths with pops and also focusing primarily on geos that I have already made work since I have a fair amount of data for them. Still testing other things and more geos but not as aggressively as normal. But I am planning to expand my data soon due to the custom tools I have.

And yep, Brad and I are still collaborating and our mastermind is going strong. We recently accepted one more member into the group who brings a very different perspective and type of experience with his approach to pops that has been a good addition.

Brad and I both are interested in breaking into FB, that is our goal. We are still working with pops as we are getting prepared to jump into FB. The biggest hurdle we are working through is reliable accounts creation and funding, which is probably the biggest challenge for FB in general, but we are making progress.

I see myself continuing with pops for a while and working on FB - I've made too much progress to quit pops completely at this point.

Just seeing the growth trend of my profit by month since I started in July, it doesn't make sense for me to exit completely. So far this month I'm at $4600 profit, my best month and if my camps continue through the month it could be a really solid finish to my year and prepare me for more aggressively tackling FB.






What are you working on these days? You've been a little quiet lately :P


Andrew

---

$4600 in profit this month so far = $460+/day on average = not too shabby at all!

I get such a thrill hearing about how your mastermind is working out!

The reason why I've been "quiet" for the past week or so was because I was at AWA. Just the flying from Canada to Thailand and back took up 2 whole days. And between tending the STM booth, socializing with friends and colleagues, and business meetings, I hardly had enough time to rest. Good thing we have matuloo on the forums everyday to hold down the fort!

It would be so great to meet you in person at a conference some day. I got to meet Brad at the Toronto meetup. A ton of people were there and we didn't get a lot of time to chat but it was neat meeting in person.



Amy

---

Matuloo does a great job, he seems to be lurking on STM as much as me - almost!

I so badly wanted to go to AWA, one of my affiliate networks was giving away a free trip to it and my AM was so gentle in letting me know that I didn't win haha.. he's like "I know you were really set on going but..." and I told him it's ok, I tried my best during the contest haha.

It just didn't make sense for me budget wise this time around. I am planning to attend ASW and atleast one of the AWA's in 2017, possibly both. We will definitely meet and have a nice chat!

Enjoy the rest of your weekend!



Andrew

---

Happy New Year STM!

Between the holidays and everything work related I've not updated this follow along very frequently. You can expect the occassional update to my progress as I continue to build and pursue those steady $1k, $2k, $5k, and $10k/days!

That being said, I wanted to conclude my monthly recap series by showing a final Decemeber recap and a final 6 month recap that documents my progress from when I first got back into affiliate marketing in July.

To catch up on the previous recaps that posted, you can find them here...

July/August Recap
http://stmforum.com/forum/showthread...l=1#post286536

September Recap
http://stmforum.com/forum/showthread...l=1#post289338

October Recap
http://stmforum.com/forum/showthread...l=1#post293273

November Recap
http://stmforum.com/forum/showthread...l=1#post296971


And here we go....



December 2016 Recap




December has certainly been an amazing finish to the year and after so much hard work and man hours put in, it's nice to see a solid profit. Still, it's far from where I need to be and I need to achieve atleast this level of profit on a regular basis to make affiliate marketing work in the long run.



6 Month Recap - July to Decemeber 2016






As you can see from the chart above, I struggled to make any real profits for months but my overall trend was going up and up along the way... from a negative month in July, to break even in August, to a small profit in September, to practically break even in October, to a small profit in November and finally to a solid profit month in December.

Now, something to be mindful of is that this is stats based around my adspend and not the total costs associated to tools, resources and a full time employee that I now have.

All costs being considered, December finished at about $9k in profit and over the last 6 months I've profited about $10k.


How Does All of This Help You?

Never.. Give.. Up..

Be smart with your money, stop asking questions and start testing for yourself. You will learn so much more. Start a mastermind, learn how to network and befriend people in the right manner, if you show value first then people will see you as a valuable resource and want to continue collaborating with you - that is one way to get ahead and learn. Start a follow along, spill all your beans in the beginning, read and learn from every thread and post on STM.

Be serious, be focused, be relentless. Have a plan and execute it.

Now go forth and conquer 2017!



Cheers to all and thanks for all the support I've received along the way!


Andrew

---

Congratulations on your success!

---

Home > Mobile > Follow-along Campaigns