Home > Mobile > Follow-along Campaigns

---

my journey to the optimization land with Zeropark (24)

---

Hello, congrats on launching your first campaign!

-85% looks ugly, but you got some leads, which is always better than seeing nothing. This tells us that either the LP or the offer has some potential and now it's about testing more of both. In the current climate, you need to test a ton of offers and LPs to find something that's gonna work.

Did you already analyze your data somehow? Try to break it down into small pieces, to see whether the 2 conversions had something in common : specific device for example. Take a look at it, so we can plan the next step.

---

Like matuloo said - you need to test more landers than just one.

Right now you're bidding really, really low. This means you're not getting traffic from a lot of the placements (i.e. ZP targets) that convert well (and there lots of people are bidding higher for, to get a bigger piece). Therefore you're not even able to gauge the earnings potential of the offer accurately.

HOWEVER, bidding low can help you save money while you're testing offers and landers - and this is exactly what you should be doing now.

Rip 5-10+ landers that look considerably different from each other and test them. Use this method to cut down to a winner:

https://stmforum.com/forum/showthrea...Banners-Part-1

Once you have winner, take that and test a lot of offers.

Once you have a good offer, test bids.

There are 3 main factors that contribute to a profitable camp. A good offer + a good lander + good traffic. So you need to have a strategy to lock them down one by one.

Good start and please feel free to ask us lots of questions!



Amy

---

ALL I NEEDED TO READ! this is the "SYSTEM" I need... "Lock um' down 1 by 1"

---

Haha glad you found that useful!

Having said THAT - your current offer isn't converting very well, so ideally you'd want to test a few more offers to see if you could find one that converts better.

This way, you can use an offer that converts better, to test and cut landers.

To summarize: You need an offer to use for testing and cutting landers, and the better this offer converts, the less you'll need to spend to cut landers down to a winner.



Amy

---

TO get more ideas about systems that you could use, check out this post of mine too, it's about a system I have used tons of times with great results : https://stmforum.com/forum/showthrea...-of-my-methods

---

Thanks for the help, the current LP t hat i ran had 87~90% CTR - is it still worth it to test more landers?

about bidding in Zeropark ive actually notived that i was top bidder mostly, its a 3 tier country so traffic relatively cheap.

how do should i approach bidding and test them correctly ?

---

Don't judge a lander based on CTR!

I've seen plenty of landers that have lower CTR than others, and yet convert the best out of the bunch. It doesn't help to just get visitors on the offer page - the lander needs to presell/qualify the visitor properly so that when they DO land on the offer page, they'll convert.

(The only exception here is when you see a really, really low CTR - so low that the math would never work out to be profitable. In that case you can cut the lander based on CTR.)

So the answer is that YES, you should DEFINITELY test more landers!

As for bidding - I would normally suggest to test landers and offers at average bid, just to make sure the traffic quality is OK (e.g. so that if you don't get conversions, you can be relatively confident that it's your offers/landers that aren't good enough, not the traffic). That way you'll give the offers and landers a chance to convert. Once you find a decent lander+offer, it would good to test bids at that point. In your particular case though, you were lucky and still saw conversions at a very small bid.

You could try the staggered bids approach: Set up several identical camps and set them to different bids - higher and lower than the average you were using, see which one(s) gives you positive ROI or near-breakeven ROI, and keep those running and cut more placements for each camp until they're green.

Or, you could increase the bid once a day and see the resulting profits/day. If the profits/day increases as you increase the bid, then that's good, and you should test an even higher bid and see if the profits/day will grow. If so, raise the bid further. If not, revert to the previous bid.

There are many ways to test/optimize bids. Those were just a couple. Best of luck!



Amy

---

A one size fits all approach is not really the best for most campaigns.

Normally I recommend bid baskets.

With zeropark however you have the great opportunity to bid based on the EPV of the placements.

Identify the placements that are doing most revenue / profit and dominate them with high bids matched to the unique EPV.

Means you use different bids for different placements based on the EPV.

---

Sorry for not updating, i was off past week now im back to work.

as for previous campaigns i wiped them clean, just got few new offers from Mobidea.

First problem for me is i see mobile content offers that i have no idea how to make pre lander for them, offer only says 'download now' with mobile number form.

How should i approach it?

---

small update, for now im running a bot test on Zeropark & Popads

using: https://stmforum.com/forum/showthrea...highlight=bots


my question is, should i run this test for every single country? or?

---

Guys, i think landers isn't the solution with conversions right now, i think for best conversion you need to give high bid for to get unique users....

---

UPDATE:

Ive spent bit of time to improve my bot detection game and handling a traffic source (whitelist / blacklist).

with help of few STM members i learned how to handle bot blacklists: https://stmforum.com/forum/showthrea...on-PPV-network

Also joined Adsimilis / Kimia to have more selection of offers.

ive just launched a campaign:

Vertical: Mobile Content
Geo: Africa
Payout: 1.62$
Traffic source: Zeropark

ripped 2 landers for now for split test, still waiting for approval from ad network side.

ill keep this thread updated as i have plenty of time now to work.

---

"Mobile content offers" is such a general category - and can include a lot of different types of offers.

The "download now" type offers you're talking about - these posts (from the same thread) may be of interest:

http://stmforum.com/forum/showthread...l=1#post270834

https://stmforum.com/forum/showthrea...l=1#post271219

Some mobile content offers have 1/2-click flow. Those you can try to direct-link without using landers. If the offer page is attractive, coupled with the simple conversion flow that should be all that's required to get conversions.

As for other mobile content offers - you'll need to show me examples in order for me to provide suggestions on how to promote them. You can also do a search on Adplexity to see if any landers turn up.


Yes, because the traffic from each placement will be from different geos.

If you're targeting carrier traffic, then no need to do a bot test - this is only necessary if you're targeting wifi.


@al0red Testing bids and using landers are 2 completely different things. Getting unique users is a third thing. All 3 things are completely separate from one another.

Please tell me why you said what you did - and I'll try to clear up some things for you.


2 landers are not enough - you'd want to test at least 5-10+ to lock down a good one. 1 offer isn't enough either.

In the first part of this post here, I've outlined a couple of test approaches for offers and landers, that you could try if you please:

https://stmforum.com/forum/showthrea...l=1#post317405

Spend time on testing landers and offers, and focus on that until you find something promising, and THEN optimize your camp further. Don't take an offer and a couple of landers and spend a ton of money trying to optimize it to green. It's the expensive way and you'll have little traffic left in the end - just not worth the time.



Amy

---

is there any strategy for frequency capping?

I assume that less frequency = more quality (less "banner blindness")

---

I optimize all my camps on 1/24hr frequency. Once I have something that works well & see some consistency, I will loosen the frequency capping to 2/24 -- give it an entire day, and then compare profit. If it's greater, then you can do this again (to 3/24 etc). At one point, you should notice profit decreases, so that's where you would stop, and revert to the 'tighter' capping used in the previous day.

Just remember to not change anything else during the days you do this, in order to get accurate data. This is just one strategy, of course, and it's not perfect due to the potential daily performance variation.

---

That is correct!

The method explained by Chris is good.

Basically the higher the frequency, the lower the ROI, but the higher the traffic volume. These 2 variables combined can result in an overall increase or decrease in profits, and this is what you're aiming to find out through testing.

This is also why Chris suggested to optimize the camp first, before playing with frequency. Since the ROI will likely fall when you increase the frequency, there would be little point in testing higher frequencies when the camp is still red - higher frequencies would only make it go into more negative ROIs. Personally I wouldn't even bother testing frequencies unless my ROI is at LEAST 50%+, preferably higher (obviously the higher the ROI, the greater range of frequencies you can test).



Amy

---

Thanks for the replies, it helps to see the more clear picture piece by piece


I ran few campaigns yesterday with average bid and noticed that for a short moment i was exposed to few placements that converted like crazy.

now it got me thinking

1) should I edit bid or maybe its better just to duplicate campaign with new bid?
2) is there any formula to calculate maximum CPM I can go with to stay profitable?
3) in case i got good lander + converting offer, where do i want to be in the bidding position? top bidder, 2nd place? (i mean to be able to receive quality converting traffic)



im still bit confused in steps for the optimization itself, running campaign for iOS with only 3G traffic and got few conversions around -40 ROI.

how should i break down my optimization steps?

tested winning lander using peakconversion tool + started cutting placements using Amy's whitelist calculator which makes it much easier - but still i feel like im working in a messy way.

anyone can advise how to systematize the optimization process

Thanks,
Max

---

Forgot to mention in my last post - regarding testing frequencies: You can test lower bids with higher frequencies to try to offset the decreased ROI with decreased costs.

You can do either.

You can clone the camp 3-5+ times and assign different bids to them, to see the different placements they each get traffic from, and cut placements from each camp individually (or, cut placements based on their performance across all camps).

Or, you can cut unprofitable placements at your current bid, then increase it to trigger traffic from higher-quality placements, cut another batch, and increase the bid again - until campaign profits starts to dwindle after another batch of cutting placements. Then revert to the previous bid and unpause the most recent batch of cut placements.

There are so many bidding approaches you can take. Feel free to experiment with whatever makes sense to you, and observer the results - see which approach seems to work better for you.


That would be difficult. Any such formula would assume the other variables to remain the same as the CPM changes, which is often not the case. Often, bidding higher will mean higher-quality traffic = higher conversion rates. If this could be calculated, then there wouldn't be a need to test different bids.


Hard to say. You'll need to test different bids and see which one will maximize your profits. If being the top bidder means having to run at a loss, then obviously you wouldn't want to stay there.


I don't have a concrete optimization plan to suggest, but here are some general thoughts.

Test enough offers until you're close enough to green or better, and then optimize further. The better your offer is, the less expensive optimization will be for you, and the more satisfying your scaling efforts will be. If you try to optimize a bad offer, you'll need to spend a lot of money on cutting lots of stuff, and be left with little traffic - and if you try to scale, you'll be repeating that for each traffic source you scale to.

Focus on the big traffic segments, i.e. placements/OSs/browsers/carriers/devices/whatever else, that are giving you enough traffic volume to be worth your bother in analyzing it. Don't spend hours looking through stats for small placements to make an extra buck or 2 a day. You should be spending that time setting up other tests, or scaling profitable camps, or seeking out new traffic sources or good offers or whatever.

Cut the worst-performing segments first, then wait a bit before cutting further, because when you cut one bad segment, some of the other segments will improve. (e.g. When you cut a few major placements, OSs and browsers etc. etc. will perform better. So after cutting the worst segments, wait and see how the rest of the segments will do, before cutting further.)

Don't over-optimize past the point of diminishing returns. This goes back to what I was saying about focusing on big traffic segments.

Also - that whitelist calculator spreadsheet of mine is overkill when it comes to pop placements - it takes too damn long to arrive at a verdict, which given the short lifespan of typical pop camps is not worth the time. I would suggest to just blacklist placements using rules of thumb, e.g. cut if the placement is in loss by 2x payout or more.


You're making good progress! Looking forward to seeing stats from your next camps! Not every camp will be profitable, but the more you test, the better your chances.




Amy

---

Thanks for the input, the 2x rule of thumb ill use it from now sounds much faster,

I actually managed to hit my first money pocket with Zeropark with 242% ROI - only problem is the volume is so small that during the night i made like 1.70$ haha

I will take your advice and focus on the bigger volume segments, problem was that so far most of them been losers and then only placements that actually converted had tiny volume.

in Zeropark its easy to see that where i can live bid on every single target but on sources like Propeller its more difficult for me.


On top of that I will probably switch from self hosted Thrive Tracker to tryout Voluum.
Recently I have way to many problems with thrive starting from delayed reports, incorrect data, wrong conversion stats, postback problems - so i will make the change probably asap.

I have few pieces missing so ill ask:

1) lets say I test completely new offer with new landers that never tried before. based on what do you decide to kill the campaign and move on to the next offer?


Thanks,
Max

---

This is exactly why 1)we should lock down a good lander and offer (after lots of testing) first so that we'd make enough of the traffic profitable, and 2)target larger volumes so we can cut more stuff and still have enough left over to profit from.


This is where having multiple camps with staggered bids will help. (I mean for sources where you can't specify the bid for each individual campaign.


In that case, I would focus on testing many landers with the aim of locking down a good lander first.

In order to identify the best lander of the bunch, you'll need to have conversions.

This means you'll need at least one half-decent offer during this initial testing - to help you lock down a decent lander.

Here are 2 approaches I like to use - try one or both to see which one suits you better:

1)Ask AMs for offer recommendations (for a specific geo and offer type). Pick 2-3 that have similar payouts - the more similar the better. Rip all the most popular landers from Adplexity (ones that have received the most traffic, that are in the target language, for that offer type, ones that look really different - don't pick landers that look similar). Have at least 5 landers, 10 would be better. Throw them all into a campaign. Cut offers and landers down to winners (using stats calculators). Then test another batch of offers.

2)Compile a big list of all offers that exist for your chosen geo and offer type. Again rip landers as above. Run all offers and landers together until one of the offers has made 2 conversions, then pause all offers except that one. Run that offer and all landers, and cut landers until the last (winner). Then use that winning lander to test more offers as above.

Further explanation:

-In the 1st approach, the offers need to have very similar payouts, because when we compare lander stats using the recommended stats calculator, this calculator will assume all conversions to be worth the same. So if the payouts are too different then the results will be inaccurate.

-In both approaches, the central idea is to test all the popular landers in use to find a winner first, and then use this winner to mass-test offers. Once you find a decent lander, you can use it for at least a few months, because landers don't "burn out" half as quickly as banners do. Offers will also come and go, but you could keep using that same lander to test new offers for quite a while.

-Basically, to test landers and cut down to a winner, you need at least one good offer that will give you the required conversions without breaking the bank. In the 1st approach I suggest to use AM-recommended offers because they're more likely to give conversions (usually the reason why they're recommended by the AM is because they've converted for other affiliates). In the 2nd approach, I suggest to pick the offer that reaches 2 conversions first, because it's an indication that it's converting at a decent rate.

-This whole approach in a nutshell: Use one or more converting offers to nail down a winning lander, then use that to mass-test offers. Once you have a good lander+offer that will hopefully make enough of the traffic profitable, cut the unprofitable parts to result in profits.


Where are the others? You only asked one.



Amy

---

Here one more bomb lol,

I just came across smartlinks - is it worth tying out?

I understand that its super fast rotation of offers so its basically 1 thing less to worry about in the setup.

question is do i craft the funnel i mean which landing pages ?

---

Smart links are certainly a viable option to test, I've had a lot of success with networks that offer them.

When it comes to landing pages and smartlinks, you have 2 options basically : Either use very generic LPs that do not try to sell a specific offer type - for example some warning page, age check, general screenshots ... OR use a smartlink that is more configurable (for example the matrix from Kimia or the Hybrid link by Bitterstrawberry) These allow to specify certain offers to rotate, or they are selected based on their performance. In either case, you would know what kinda offers would be rotated by the smartlink algo, so you can use LPs that would suit them.

Another option would be to use a direct link, so without a LP. This will NOT work in most cases, but it's still possible to make it work on sources with good targeting.

---

Hi!

Another thing to add - you can use the Smartlink as a "backup" option for the traffic that didn't convert with your offer. Meaning, that if for some reason the user doesn't appear to see your offer, you can set a rule for him to be redirected to a Smartlink. Then the algorithm will determine his segment and show an offer anyway. This way you make sure that you monetize all traffic possible.

---

You can craft your campaigns as you wish for each targeting using HybridLink® by BitterStrawberry.com
Generally, Smart Link rules are based on all active traffic of the specific network, therefore it's possible your specific traffic sources won't have the same performance as the overall of the network.

Cheers!

---

Home > Mobile > Follow-along Campaigns