Home > POP / PPV / Redirect >

---

Mobile Pops/Redirect Campaign That Made me $150/day average (32)

---

People that know me, know that I am big time into mobile. I don't cloak facebook and run muscle, skin or diet; sure doing rebills and free trials is what will get you 200-500% ROI but you also have to deal with accounts going down constantly, tons of new credit cards for the new accounts so it definitely is too much for someone just starting affiliate marketing.

I have a real agency [with employees - see post here], and at many traffic sources we enjoy special perks like once a month billing (Golden nugget: you too can get that once you hit xxxx per day revenue on a traffic source)

Anyway, I like running green and establishing long term relationships with AMs + Traffic Sources. It's not what you know, its who you know - I'm sure you heard this many times before, well its true for affiliate marketing as well.

Typically many newbies will start on a traffic source and they'll expect VIP treatment; HUGE MISCONCEPTION - traffic sources have interest so they won't really kiss your ass until you show them you mean business. After you get to to the xxxx per day level, doors will open and premium placements will come knocking plus net30 billing and many other perks that come with being a high volume affiliate.

To get you started on your journey what I am going to do now, is give you something I was really hesitant about. You don't have to pay me a cent for it, you owe me nothing. What comes around goes around, I'm a firm believer so here is a campaign that you can run now and bank.

This campaign made me on average 150 after getting it off the ground with little maintenance. It was pretty fun too, combining current events/trends with economical realities that everyone's broke and wants to save money.

So lets go!




Campaign Vertical: Sweepstakes (Mobile)
Geo: Canada
Traffic source: It is a mobile POP source, i won't give this to you though :P
Offer: Gift Certificate to Starbucks
Payout: $1.75
Accepted traffic: ALL



Introduction


Before getting into anything, I wanted to go broad - because in my experience and here's a golden nugget - broad appeal campaigns, and angles are always the best - because if you want to scale, you can - if you go niche, you won't have volume and as much scaling ability.

Neways, when I ran this campaign I picked starbucks because everyone I know in Canada always goes on starbucks runs, so its a no brainer and fou0nd a starbucks gift card at one of my preferred networks (if you are on my mailing list you know which CPA Networks I use)

I was going to make a lander which asks the visitor a few question about a very hot trend and then gives them a reward. How did I identify trends? I went on to google trends of course! Selected canada, then selected celebrities since everyone is obsessed about celebrities. Ended up Picking PAUL WALKER because a lot of buzz was going around at the time - may he rest in peace.

Now that I had my trend picked out, It was time to make a lander. I used the design services of bannerslanders and asked them to make a simple survey lander. Here it is:




As you can see it has a few important elements

1) Headline getting the attention of the person who lands on this lander
2) Sub heading that tells them about the reward
3) A picture that helps even the dumbest idiots who didn't know who paul walker is know immediately
4) A block of text with the question telling them that there's 3 questions, so its easy
5) Yes and No buttons
6) A Background showing the PRIZE (this often gets forgotten by affiliates, but in my tests putting an image of the prize in the background is like subliminal messaging and it works)


I split tested 5 angles on the top header

Angle1 - Are you a true paul walker fan?
Angle2 - Paul Walker passed away, but his legacy lives on!
Angle3 - Do you know Paul Walker?
Angle4 - How well do you know PAUL WALKER?
Angle5 - Are you smart enough? 7 out of 10 of your friends answered correctly!

NOTE: One of these angles performed really well, I won't tell you which though, but if you can connect the dots - it's obvious

Stages of running this offer

#1) I spent $100 to collect data, so thats like $20 per lander. This helped me establish one thing, which lander was the best out of the gate.

I was bidding high, I looked at what the top bid is and added 30% because I wanted data fast, and to spend my initial budget of $100!

#2) Next up, I paused all the landers that weren't so good, I was looking at CV in volume.

Now I had a winning angle, NOTE how I still didn't touch the placements, or did anything else.

#3) More split testing was in order.

I had B&L make some more variations of this lander, different background colour instead of yellow, different font colour, adding border, moving stuff around, ie the question above the picture, the picture on the top - other randomness. Basically everything I can think off.

There was slight increase in CR, then I did the next best thing and added testimonials on the bottom under the YES AND NO Buttons and this increased my CR by 50%. It was people holding the gift card (yea I had it photoshopped) and smiling and giving some BS. I also paid some woman on Fiverr to give me a video testimonial holding up a gift card which I made her print out - yea it cost me 10 gigs but whatever...

I like to split test a ton of stuff before going in and looking at placements. I don't care if a placement is draining my profits and funds my belief is that if you find the best mix after split testing like a mofo u can make a ton of shitty placements at least break even. And in my experience how traffic sources work, the more placements you buy from the more volume u get overall. So that's another golden tip right there.

After running another $100 bucks on my new multivariate tests, I found 3 landers that were at a good target ROI for me, and i paused the rest.

#4) Now I ran another $100 bucks by this time i was only around -32% in the whole overall and 300$ spend.

Next it was finally time to look at placements, my golden rule is offer payout x3 at these lil payout offers, since I had my landers ultra optimized only through the last few days, I looked at the data during the last 3 days in Voluum, not prior, because then my numbers would've been off.

So, based on the past 3 days, I started pausing placements that were under my cost, and minimum offer payout x3 in spend. I did this quickly because I simply exporting all the placements from volume to a excel CSV, then sorted by the criteria using the rules in open office spreadsheets IF placement sent x visits, and EPV is < realCPC (which I calculated manually btw, pm me if u wanna know how) and gave it to my worker to go into the traffic source and one by one block everything, since this traff source doesn't allow mass exclude/pause.

Day after we were at break even, day after that we were at 30% roi, and again looking at the placements and moving up my exclusion treshold from offer payout x3 to offer payout x5 we did another exclusion and after that ended up at 50% ROI or around 147$ per day on auto.


Where to go from here? How to turn the $150/day into xxxx per day!?

There's two roads I could've taken here, road one is to scale it horizontal and road two is to scale it vertical.

Scaling it horizontal that is when we'd take this to other redirect/pop sources - this is my favorite and the fastest you just really have to know each traffic source you scale it to because its NOT JUST CUT N PASTING IT. What no one tells you when they say you really gotta know the traffic source before you jump around means, is that every traffic source responds differently to change, for example if you change the bid on 1 source, it might kill your campaign because it restarts you in the chain of bidders; whereas on the other traffic source, if you remove a targeting option it might kill your campaign, or if you decrease and increase budgets - so yea before you go crazy scaling horizontal, put in $500-1000 and focus on getting to know that 1 traffic source.


Scaling it vertical is if we'd take all the top 10 most popular celebs in Canada for example, and make a series of questions and lander for each, then run them all and rinse and repeat. It is like how you'd take a global mobile offer and scale it into other geos. Heck, that reminds me you could scale this to other geos and just swich up the lander/reward text to be relevant to that country. TIP: I know in some eastern eu geos, grocery
vouchers ie TESCO are killing it right now... so u could technically use the same method.

That's it for now, hope you find this case study helpful.

iAmAttila shares.

---

I like your approach - make highly influential changes on the front end first, optimise there, and only they start culling placements.

Definitely the best way to not sacrifice volume unnecessarily though clearly you need to trust the process and have a converting offer ;D

---

Excellent post!

How many variations did you make at stage 3?

---

Awesome insight, thanks for sharing!

---

Another awesome case study. Thanks for sharing.

---

20
I always test a creative at around 5 bux each when its a low payout

---

How much did you end up spending for the landers / banners?

Looks like you dropped another $400 there.

---

Absolutely amazing post. I see a lot of the places I went wrong when I attempted sweeps. (mainly not split testing enough questions)

Would you have to hide the referrer of this LP from the affiliate network because the text "Answer correctly and get a $100 gift card!" infers that they "won" after they answer all 3 questions? If yes, would https be an acceptable way to do it for a mobile campaign like this one?

---

awesome share iAmAttila

Thank you

---

Nice campaign.

There's SO many angles to split against Fast and the Furious. Going even broader would help take a larger interest of the target market. Targeting the survey to specific OS's etc. Looking good, lots of scale room here.

---

Great case study. thank you Attila!

---

One of the best studies i have ever read here on STM. Nice job !

---

Great case study. Couple of questions:
1. Do you keep an eye on placements that are sucking up all your budget? Like cut them after 3x or 5x spend?
2. How do you know if you are not given the bad placements initially. I believe on some networks, the more placements you cut, the more they give you. So if you are not cutting from the beginning, you may be getting the good stuff.

---

1. Of course
2. You test, - you won't know till you try -

---

Great post!

I also have some questions:

1. How did you get to pick this offer specifically? I can imagine it was recommended by an AM but did you have some testing procedure beforehand before going with this offer?
I'm currently testing email submit offers and I'm wondering how you test these. Do you just split test a ton of offers by direct linking them?

2. Once you have collect enough data on the 130% average bid, how do you go about finding the right bid price?

---

"So, based on the past 3 days, I started pausing placements that were under my cost, and minimum offer payout x3 in spend. I did this quickly because I simply exporting all the placements from volume to a excel CSV, then sorted by the criteria using the rules in open office spreadsheets IF placement sent x visits, and EPV is < realCPC (which I calculated manually btw, pm me if u wanna know how) and gave it to my worker to go into the traffic source and one by one block everything, since this traff source doesn't allow mass exclude/pause.

Day after we were at break even, day after that we were at 30% roi, and again looking at the placements and moving up my exclusion treshold from offer payout x3 to offer payout x5 we did another exclusion"

Can you explain this a little more. If you exluded x3 payout, wouldnt that have excluded x5 payout placements, as well? Also, what is "under my cost?"

Thanks

---

VERY interesting approach. Not culling placements first would give me the fear, but I can see that it's a valuable technique to concentrate your data and make sure that when you do cut placements, you've got enough information.

Major food for thought...

What about bot traffic, btw? Weren't you worried that bots would eat a whole load of cash whilst you were initially testing, because you're not culling placements at all?

---

1. I tested a bunch not just starbucks gift card. I wanted something that is broad, and many people know.

2. Right bid price is based on what you are comfortable with, if 20% roi works for you, you can bid higher than the guy who wants to make 50% ROI, etc.




we block placements that spend offer payout x (ie 3) and the epv in Voluum is less than the real cpc we're paying.
sometimes it might be offer payout 5x, sometimes 3x... it depends on how strong the overall campaign is when running RON..

if we gotta be very aggressive, we block earlier if most of the placements are good, then we let not so good placement spend more of the offer payout

hope that makes sense


initially we always test for best angle, and just peak at placements to see if there's any really horrible ones eating up the entire test budget

so if a placement sucks, and is bot traffic we will notice it when we take occasional quick peaks the placements

---

Are you guys worried about getting sued for using a Celebrities picture in your landers? OR even their names on banners?

---

I wouldn't recommend using celebrities or an authority figure without their permission.

With that said, from experience if you're not running a lot of volume, you're unlikely to run into any problems.

---

Great case study Attila!

It's funny to see how the same stuff keeps working after many years.

I started 7-8 years ago running survey landers on adwords and the same concept works perfectly these days.

The traffic sources keep evolving but the basics stay the same.

---

But isnt this lander using a celebrity?

---

Having a long term relationship with advertisers is valuable as well. However, you'll have a hard time doing this with these types of landers and creatives.


For any of you guys diving in now:

You are at a serious risk to not get paid, get kicked off an offer, get kicked off an ad network, get sued, get an advertiser sued, have mediocre intercourse while banking $20k daily, etc. if you run non-compliant creatives and landers in this vertical.

---

I reckon it was smth like, "Thank you for your opinion, as a thank you you are selected for a chance to win an xxxxx" as a js pop.

---

Following a few procedures of this case study.
Thanks for it! I'm a bit skeptical though that you didn't cut placements already in the beginning!
When would you decide that the campaign is a fail and try another angle? (Another angle, not another headline)

---

interesting stuff!

Do these sort of offers work with regular display traffic as well?

---

^ Yes they do def work

---

Great info, thanks.

---

Nice, it's still work on 2016 ?

---

Amazing!
So detailed
Thanks!

---

Great info thanks.

---

Again, this is a very old post so the info is not relevant from a large part. Just a headsup for anyone reading this and not noticing the post date

---

Home > POP / PPV / Redirect >