Home > >

What should you optimize first and how : Offers, Banners or Landing Pages? (29)

03-24-2017 02:40 PM #1 matuloo (Legendary Moderator)
What should you optimize first and how : Offers, Banners or Landing Pages?

We had several threads lately here on STM, where aspiring affiliates asked for some kind of plan on how to structure the optimization process, or to put it in a different way : what to optimize first – banners, LPs or the offers.

First of all, let me emphasize that EVERY successful affiliate has specific optimization methods and they don't necessarily have to be identical with my own, but there are some rules that are pretty much universal, so let me focus on those in this article.

THE OFFER IS THE KING!

I've been hearing it for years … Traffic is King! … then it changed to Content is King! … but since we are a community of media buying affiliates, I believe that Offer is the true KING You can have the best traffic in the world, but you can pretty much show it up your ass in case the offers at your disposal are crappy.

You can be a banner Ninja, you might design the most killer LPs in the world, but still, it's all not worth shit if you don't have a good offer that you can send the traffic to. Keep this in mind, when coming up with new campaigns : “A superb offer can convert even poor traffic, but a bad offer won't convert, no matter what traffic you send to it!”

Weeding out the poor offers and finding the promising ones MUST be the first step in your optimization process, everything else comes after this!

HOW TO FIND THE GOOD OFFERS?

This article is about optimization process, not picking offers, but anyways, let me write a few lines about it too.

So how do we actually find the promising offers …. surprise, surprise … by testing But seriously, there are several ways of getting your hands on a good offer.

1. Ask your Affiliate Manager – this is the most obvious way, but still not a bullet proof one. You never know how well the AM slept the night before, how many beers they had at the show they just returned from … so the recommendations might as well reflect that. Some AM's are also not the most skilled people in the universe...

But these people still see what's making money for the network, so definitely consider the advice you are getting from them. Keep one thing in mind though – the EPCs they see are AVERAGES. Yours can be higher but also way lower!

Having good relationship with your AMs and working with the good ones can be a MASSIVE advantage!

2. Spying on your competition – this is a very good option, if not the best one. Get a spytool, or use some VPN to spy manually … check what others are promoting and find the offers at your affiliate networks.

AM is not a CHARITY, so if you see an offer being pushed by several large affiliates for multiple days or weeks, it's obviously working for them. They are certainly not doing that to blow their money out of the window.

3. Pick on your own – browse the offer DB of your favorite networks or direct advertisers and pick something for a test. When doing this, ALWAYS double check with the AM whether the offers of your choice are active at all! Many networks are not exactly pedant when it comes to removing dead/paused/capped offers.

OK, I PICKED SOME OFFERS, NOW WHAT?

So, by using any of the 3 methods (I recommend a combination) you now have a bunch of offers to test and we need to figure out which ones of them are bad... this must be the first step of our optimization process.

In order to make sure we keep the same conditions for all offers, we have to start with a fixed funnel that we won't touch until the offer testing is over. It's very wise to use a spy-tool to get the initial funnel, in order to make sure you are using something proven.

As an experienced affiliate, you will have your own benchmark/testing funnels from past campaigns, but when starting out, it's simply the easiest to rip some banners and LPs and use those. Do not choose too many banners or LPs at this stage. Don't start with 1 banner/LP either, you need some variety.

Let this run for at least 2x offer payout x # of offers. So if you are testing 5, $5 payout offers, you need to spend at least 2x5x5=$50. Then take a look at the offers and see if any start to stand out or fall behind. I think it's better to see some offers fall behind and cut those, once enough traffic has been sent to them. In case the trend isn't clear, I keep on running. 2x offer payout is the initial spend you have to make, before even considering doing something with the campaign, it won't be enough to make a final decision, but it's enough to see if you are onto something or the whole funnel is a total failure and you need to change something : see the NOTE below.

This test should leave you with 2 or 3 offers that performed the best in this test. Sometimes, it will be just 1 offer, but I always prefer to move on with at least 2.

NOTE: in rare cases, you will end up with 0 conversions after the test runs out. In such case, there must be some flaw in the funnel – you either picked very bad banners/LPs or the traffic sucks in one way or another. Assuming that you were testing multiple offers, it's highly in-probable that they all suck. In this case, you need to pick different creatives and/or run some BOT tests on your traffic source.

OFFERS TESTED, WHAT'S NEXT.

We have 2 more factors to test now : Banners and Landing Pages. Technically, you could start with optimizing either of them, but I personally prefer to start with Landing Pages. Wondering why? Landing Pages are less prone to burn out, you don't need to change them that often and many of them are transferable from one source to the next … LPs are simply not so sensitive and once you have a good one, it can work for a longer time.

When testing LPs, I build several variations, covering several angles and let them run head to head for a significant amount of traffic. While testing LPs, I don't change banners and offers, they have to stay the same in order to get representative data.

Just like with offers, I let them run until some LPs start to fall behind and cut those.

Once we have the best LP's narrowed down, we can start with testing banners. This is the part where I usually spend the most time, since I test dozens of them. All kinds of variations, headlines, colors … you can go really wild here and look for elements that seem to improve the performance.

It can be the image on the banner, the headline … some graphic element … whenever you think you detected something like that, make more banners with the element that you believe affected the performance.

WHERE TO MOVE FROM HERE?

This process should result into a solid funnel, that you can use to build profitable campaigns with. The war is not over yet, as you need to optimize placements and other targeting options, but you should have a solid benchmark funnel now, that you can count on.

I'm a big fan of building a few benchmark funnels like this, it makes starting new campaigns way easier. This way, you're not starting out blindfolded, you already know what to expect. Want to test a new offer? Take the funnel, replace just the offer and let it run for a while, based on the performance you will know whether the offers has some promise or not.

So let me sum it up :

1. Start with optimizing the offers, everything else has lower priority.
2. Optimize Landing Pages as the second factor. You could also sort out banners first, but I recommend to handle LPs first.
3. Next come the banners, take your time and test dozens of them if necessary.
4. Now you should have a solid funnel that you can use as a benchmark.
5. Move on to optimizing placements and whatever applicable targeting options.

Thanks for reading

03-24-2017 03:05 PM #2 eike_west (Member)

Quote Originally Posted by matuloo View Post
Let this run for at least 2x offer payout x # of offers. So if you are testing 5, $5 payout offers, you need to spend at least 2x5x5=$50.
This would be your rule of thumb for the whole budget (50$ initial testing budget) or per combination of Banner x LP x Offer?

03-24-2017 03:30 PM #3 thien425 (Member)

Thanks for the awesome post Matuloo! I have a question about this though. I've read about testing staggered bids usually from the start. If you do this at all, which part of this process would you test staggered bids?

03-24-2017 03:40 PM #4 matuloo (Legendary Moderator)

Quote Originally Posted by eike_west View Post
This would be your rule of thumb for the whole budget (50$ initial testing budget) or per combination of Banner x LP x Offer?
Initial budget for the whole campaign, to see whether you are onto something or it's a total miss. I've used that rule-of-thumb just to give you an idea, if the trends are not clear after this, which they most likely won't be yet, keep on testing. I also made a note about this in the original post, so it's more clear.

03-24-2017 03:41 PM #5 matuloo (Legendary Moderator)

Quote Originally Posted by thien425 View Post
Thanks for the awesome post Matuloo! I have a question about this though. I've read about testing staggered bids usually from the start. If you do this at all, which part of this process would you test staggered bids?
When working with banner traffic, testing several bid levels comes later on, not from the start. You need to make sure you have at least somewhat working funnel first, especially the offer, then you can test the other variables.

03-24-2017 04:09 PM #6 Mr Payne (Member)

Solid post matuloo!

Another way to find good offers is to mastermind and/or network with other affiliates. When you build relationships you share ideas and insights on what is currently working well for others.



Andrew

03-24-2017 04:21 PM #7 thien425 (Member)

Quote Originally Posted by matuloo View Post
When working with banner traffic, testing several bid levels comes later on, not from the start. You need to make sure you have at least somewhat working funnel first, especially the offer, then you can test the other variables.
Oh ok, how about pop traffic? I'm guessing it's the same since the offer and funnel is also more important?

03-24-2017 04:43 PM #8 matuloo (Legendary Moderator)

Quote Originally Posted by thien425 View Post
Oh ok, how about pop traffic? I'm guessing it's the same since the offer and funnel is also more important?
With pops, you have one less optimization factor - the banners. Banners can help you pre-filter the traffic in a way. Using staggered bids gives you some control over the traffic quality, so you can use this technique instead of optimizing banners. But yes, offer is still the most important part of the puzzle with POPs too.

03-24-2017 04:44 PM #9 matuloo (Legendary Moderator)

Quote Originally Posted by Mr Payne View Post
Solid post matuloo!

Another way to find good offers is to mastermind and/or network with other affiliates. When you build relationships you share ideas and insights on what is currently working well for others.



Andrew
Definitely, networking can help a ton Good point.

03-24-2017 06:30 PM #10 cmdeal (Veteran Member)

Quote Originally Posted by matuloo View Post
We had several threads lately here on STM, where aspiring affiliates asked for some kind of plan on how to structure the optimization process, or to put it in a different way : what to optimize first – banners, LPs or the offers.

First of all, let me emphasize that EVERY successful affiliate has specific optimization methods and they don't necessarily have to be identical with my own, but there are some rules that are pretty much universal, so let me focus on those in this article.

THE OFFER IS THE KING!

I've been hearing it for years … Traffic is King! … then it changed to Content is King! … but since we are a community of media buying affiliates, I believe that Offer is the true KING You can have the best traffic in the world, but you can pretty much show it up your ass in case the offers at your disposal are crappy.

You can be a banner Ninja, you might design the most killer LPs in the world, but still, it's all not worth shit if you don't have a good offer that you can send the traffic to. Keep this in mind, when coming up with new campaigns : “A superb offer can convert even poor traffic, but a bad offer won't convert, no matter what traffic you send to it!”

Weeding out the poor offers and finding the promising ones MUST be the first step in your optimization process, everything else comes after this!

HOW TO FIND THE GOOD OFFERS?

This article is about optimization process, not picking offers, but anyways, let me write a few lines about it too.

So how do we actually find the promising offers …. surprise, surprise … by testing But seriously, there are several ways of getting your hands on a good offer.

1. Ask your Affiliate Manager – this is the most obvious way, but still not a bullet proof one. You never know how well the AM slept the night before, how many beers they had at the show they just returned from … so the recommendations might as well reflect that. Some AM's are also not the most skilled people in the universe...

But these people still see what's making money for the network, so definitely consider the advice you are getting from them. Keep one thing in mind though – the EPCs they see are AVERAGES. Yours can be higher but also way lower!

Having good relationship with your AMs and working with the good ones can be a MASSIVE advantage!

2. Spying on your competition – this is a very good option, if not the best one. Get a spytool, or use some VPN to spy manually … check what others are promoting and find the offers at your affiliate networks.

AM is not a CHARITY, so if you see an offer being pushed by several large affiliates for multiple days or weeks, it's obviously working for them. They are certainly not doing that to blow their money out of the window.

3. Pick on your own – browse the offer DB of your favorite networks or direct advertisers and pick something for a test. When doing this, ALWAYS double check with the AM whether the offers of your choice are active at all! Many networks are not exactly pedant when it comes to removing dead/paused/capped offers.

OK, I PICKED SOME OFFERS, NOW WHAT?

So, by using any of the 3 methods (I recommend a combination) you now have a bunch of offers to test and we need to figure out which ones of them are bad... this must be the first step of our optimization process.

In order to make sure we keep the same conditions for all offers, we have to start with a fixed funnel that we won't touch until the offer testing is over. It's very wise to use a spy-tool to get the initial funnel, in order to make sure you are using something proven.

As an experienced affiliate, you will have your own benchmark/testing funnels from past campaigns, but when starting out, it's simply the easiest to rip some banners and LPs and use those. Do not choose too many banners or LPs at this stage. Don't start with 1 banner/LP either, you need some variety.

Let this run for at least 2x offer payout x # of offers. So if you are testing 5, $5 payout offers, you need to spend at least 2x5x5=$50. Then take a look at the offers and see if any start to stand out or fall behind. I think it's better to see some offers fall behind and cut those, once enough traffic has been sent to them. In case the trend isn't clear, I keep on running. 2x offer payout is the initial spend you have to make, before even considering doing something with the campaign, it won't be enough to make a final decision, but it's enough to see if you are onto something or the whole funnel is a total failure and you need to change something : see the NOTE below.

This test should leave you with 2 or 3 offers that performed the best in this test. Sometimes, it will be just 1 offer, but I always prefer to move on with at least 2.

NOTE: in rare cases, you will end up with 0 conversions after the test runs out. In such case, there must be some flaw in the funnel – you either picked very bad banners/LPs or the traffic sucks in one way or another. Assuming that you were testing multiple offers, it's highly in-probable that they all suck. In this case, you need to pick different creatives and/or run some BOT tests on your traffic source.

OFFERS TESTED, WHAT'S NEXT.

We have 2 more factors to test now : Banners and Landing Pages. Technically, you could start with optimizing either of them, but I personally prefer to start with Landing Pages. Wondering why? Landing Pages are less prone to burn out, you don't need to change them that often and many of them are transferable from one source to the next … LPs are simply not so sensitive and once you have a good one, it can work for a longer time.

When testing LPs, I build several variations, covering several angles and let them run head to head for a significant amount of traffic. While testing LPs, I don't change banners and offers, they have to stay the same in order to get representative data.

Just like with offers, I let them run until some LPs start to fall behind and cut those.

Once we have the best LP's narrowed down, we can start with testing banners. This is the part where I usually spend the most time, since I test dozens of them. All kinds of variations, headlines, colors … you can go really wild here and look for elements that seem to improve the performance.

It can be the image on the banner, the headline … some graphic element … whenever you think you detected something like that, make more banners with the element that you believe affected the performance.

WHERE TO MOVE FROM HERE?

This process should result into a solid funnel, that you can use to build profitable campaigns with. The war is not over yet, as you need to optimize placements and other targeting options, but you should have a solid benchmark funnel now, that you can count on.

I'm a big fan of building a few benchmark funnels like this, it makes starting new campaigns way easier. This way, you're not starting out blindfolded, you already know what to expect. Want to test a new offer? Take the funnel, replace just the offer and let it run for a while, based on the performance you will know whether the offers has some promise or not.

So let me sum it up :

1. Start with optimizing the offers, everything else has lower priority.
2. Optimize Landing Pages as the second factor. You could also sort out banners first, but I recommend to handle LPs first.
3. Next come the banners, take your time and test dozens of them if necessary.
4. Now you should have a solid funnel that you can use as a benchmark.
5. Move on to optimizing placements and whatever applicable targeting options.

Thanks for reading
This is spot on!

03-24-2017 11:27 PM #11 falconpunch (Member)

Hey, thanks a lot for this awesome thread matuloo!

I have one question, when you look for 3 or 4 different offers to test, do you take all the same? Like 4 different win iPhone 7 sweepstakes? Because if the offers arent for the same then you can not share the LPs between them.

My confusión here is, whats the diference of trying different offers and test múltiple campaigns? Are we talking about the same then?

Im very confused about what kind of offers you should split-test and if they share something (LPs or banners)

Thanks again!

03-25-2017 12:55 AM #12 matuloo (Legendary Moderator)

Quote Originally Posted by falconpunch View Post
Hey, thanks a lot for this awesome thread matuloo!

I have one question, when you look for 3 or 4 different offers to test, do you take all the same? Like 4 different win iPhone 7 sweepstakes? Because if the offers arent for the same then you can not share the LPs between them.

My confusión here is, whats the diference of trying different offers and test múltiple campaigns? Are we talking about the same then?

Im very confused about what kind of offers you should split-test and if they share something (LPs or banners)

Thanks again!
Yup, in case you're targeting in some way, you should stick to one offer type or at least a similar one. The more specific angle you are using, the more specific you have to be with the offer too. On the other hand, when buying POP traffic for example and sending it directly to an offer, you can use multiple offer types, since there is no targeting involved.

03-25-2017 05:31 AM #13 falconpunch (Member)

Quote Originally Posted by matuloo View Post
Yup, in case you're targeting in some way, you should stick to one offer type or at least a similar one. The more specific angle you are using, the more specific you have to be with the offer too. On the other hand, when buying POP traffic for example and sending it directly to an offer, you can use multiple offer types, since there is no targeting involved.
Hey Matuloo thanks for the reply!

What did you mean with On the other hand, when buying POP traffic for example and sending it directly to an offer, you can use multiple offer types, since there is no targeting involved.?

Do you recommend NOT using landing pages for the initial phase?

03-25-2017 09:58 AM #14 matuloo (Legendary Moderator)

Quote Originally Posted by falconpunch View Post
Hey Matuloo thanks for the reply!

What did you mean with On the other hand, when buying POP traffic for example and sending it directly to an offer, you can use multiple offer types, since there is no targeting involved.?

Do you recommend NOT using landing pages for the initial phase?
That was an example to demonstrate that in some cases, it's ok to mix offer types. I don't recommend to run pretty much anything without LPs, except for some PIN submits.

You can also mix offer types in case of generic LPs, so some simple warnings, entrance gates, maybe game/videos screenshots. But as long as you dive deeper into some angle, you should stay consistent from the LP to the offer.

04-19-2017 11:26 PM #15 thien425 (Member)

For the initial offer test, how many offers is a good number to split test? I have about 20 similar offers for the same geo just from one network but I'm not sure if I should run traffic to them all right away or just a few at a time and rotate them out when I cut offers.

04-24-2017 11:42 AM #16 matuloo (Legendary Moderator)

Quote Originally Posted by thien425 View Post
For the initial offer test, how many offers is a good number to split test? I have about 20 similar offers for the same geo just from one network but I'm not sure if I should run traffic to them all right away or just a few at a time and rotate them out when I cut offers.
I don't thing testing more than let's say 5 at once is a good idea ... the more offers you test, the more traffic you need to run to reach significant results. You could theoretically run all 20 at once, but in order to assure even traffic distribution for all, you would have to run for a long time and buy a lot of traffic.

I do tests like this sometimes - on a smaller target group that I limit by targeting options, then I let it run for a few days so all offers get some traffic throughout the day and from all placements.

If you are after fast results, stick to 5 or less offers in one test.

04-26-2017 06:15 PM #17 thien425 (Member)

Thanks for the reply matuloo. I have another question if you don't mind. I am running the initial offer testing now until i reach at least 2x offer payout x # of offers. Right now I am only testing wifi/os and its not looking good so far. I plan on testing carrier/os next, if I don't get any conversions from the wifi/os campaign, shouId I even bother testing the same offers for carrier/os or just test new offers? I hope this question makes sense.

04-27-2017 10:29 AM #18 matuloo (Legendary Moderator)

Quote Originally Posted by thien425 View Post
Thanks for the reply matuloo. I have another question if you don't mind. I am running the initial offer testing now until i reach at least 2x offer payout x # of offers. Right now I am only testing wifi/os and its not looking good so far. I plan on testing carrier/os next, if I don't get any conversions from the wifi/os campaign, shouId I even bother testing the same offers for carrier/os or just test new offers? I hope this question makes sense.
Depends on the offers - if it's carrier billing offers, you shouldn't be testing wifi at all for example and you should go for carrier traffic directly.

There are situations where carrier traffic could work better than wifi, even if it's not a carrier billing offer ... for example offers that cater to people on the move, something related to traveling for example. Some placements can also have higher than normal % of carrier traffic, because they again cater more to people on the move : reviews of restaurants, some location services, maps ... and these placements might have better traffic for your offers. So yes, it could make sense to test on carrier traffic, even thou you didn't have good results with wifi ... depending on the offer.

04-27-2017 11:49 AM #19 thien425 (Member)

Quote Originally Posted by matuloo View Post
Depends on the offers - if it's carrier billing offers, you shouldn't be testing wifi at all for example and you should go for carrier traffic directly.

There are situations where carrier traffic could work better than wifi, even if it's not a carrier billing offer ... for example offers that cater to people on the move, something related to traveling for example. Some placements can also have higher than normal % of carrier traffic, because they again cater more to people on the move : reviews of restaurants, some location services, maps ... and these placements might have better traffic for your offers. So yes, it could make sense to test on carrier traffic, even thou you didn't have good results with wifi ... depending on the offer.
Oh I see, these offers are carrier billing offers I think but the descriptions says wifi traffic allowed, which is why I tested wifi. So just to clarify, if the offers are pin submits that accept wifi traffic, I shouldnt even bother testing wifi and just carrier?

04-27-2017 02:16 PM #20 matuloo (Legendary Moderator)

Quote Originally Posted by thien425 View Post
Oh I see, these offers are carrier billing offers I think but the descriptions says wifi traffic allowed, which is why I tested wifi. So just to clarify, if the offers are pin submits that accept wifi traffic, I shouldnt even bother testing wifi and just carrier?
Wifi is allowed and they do monetize it with a more complicated flow, so the traffic is not totally lost, but it will convert WAY worse than carrier traffic. But, you have extremely low chance to profit on it, I personally do not bother with wifi traffic at all when promoting carrier billing offers.

Have you seen my article about these offers? https://stmforum.com/forum/showthrea...o-promote-them

01-01-2019 07:09 PM #21 dtalexone (AMC Alumnus)

Thank you for this resource, @matuloo!

In any phase of the campaign launch and optimization, you are testing the same types of elements against each other (ads vs ads, lps vs lps) and not necessarily a combination of elements, right?

Let's envision this funnel:

Click image for larger version.  Name: funnel.png  Views: 23  Size: 8.9 KB  ID: 20274

There are 6 combinations of ads, lps, and offer.

Ad 1 + LP 1 + Offer
Ad 1 + LP 2 + Offer
Ad 2 + LP 1 + Offer
Ad 2 + LP 2 + Offer
Ad 3 + LP 1 + Offer
Ad 3 + LP 2 + Offer

Whether using a rule of thumb of 4x payout or waiting for statistical significance, you are not testing these combinations, correct?

And when you are testing LPs, are you waiting for LPs to reach a certain ad spend with each phase to make a decision or are you waiting for statistical significance regardless of ad spend?

Thanks in advance for your help!

01-01-2019 09:47 PM #22 matuloo (Legendary Moderator)

Quote Originally Posted by dtalexone View Post
Thank you for this resource, @matuloo!

In any phase of the campaign launch and optimization, you are testing the same types of elements against each other (ads vs ads, lps vs lps) and not necessarily a combination of elements, right?

Let's envision this funnel:

Click image for larger version.  Name: funnel.png  Views: 23  Size: 8.9 KB  ID: 20274

There are 6 combinations of ads, lps, and offer.

Ad 1 + LP 1 + Offer
Ad 1 + LP 2 + Offer
Ad 2 + LP 1 + Offer
Ad 2 + LP 2 + Offer
Ad 3 + LP 1 + Offer
Ad 3 + LP 2 + Offer

Whether using a rule of thumb of 4x payout or waiting for statistical significance, you are not testing these combinations, correct?

And when you are testing LPs, are you waiting for LPs to reach a certain ad spend with each phase to make a decision or are you waiting for statistical significance regardless of ad spend?

Thanks in advance for your help!
At some point, I was trying to focus on the specific combinations ... but quite often, when I isolated these winning combos, they stopped to work when run exclusively. There are several reasons for it ... bit of luck for some combos, bots that make good combos look bad ... etc.

So even though I still try to push the best combinations when I spot clear winners, usually I'm trying to find the best of each as you said .. so best banners, best LPs and best offers.

You could say I'm mixing the two approaches a bit ... if I see a banner winning, I will check whether it performs on a comparable level with all the LPs in the test ... if it is, I focus just on the banners, but if I see certain LP performing way better, I would try to isolate the combination and test it isolated.

Hope I didn't confuse you to much

Matej.

01-02-2019 01:18 AM #23 dtalexone (AMC Alumnus)

Quote Originally Posted by matuloo View Post
At some point, I was trying to focus on the specific combinations ... but quite often, when I isolated these winning combos, they stopped to work when run exclusively. There are several reasons for it ... bit of luck for some combos, bots that make good combos look bad ... etc.

So even though I still try to push the best combinations when I spot clear winners, usually I'm trying to find the best of each as you said .. so best banners, best LPs and best offers.

You could say I'm mixing the two approaches a bit ... if I see a banner winning, I will check whether it performs on a comparable level with all the LPs in the test ... if it is, I focus just on the banners, but if I see certain LP performing way better, I would try to isolate the combination and test it isolated.

Hope I didn't confuse you to much

Matej.
I understand what you are saying and it makes sense.

It would seem also that it can get quite expensive testing different combinations. In that event, are you comparing combinations against each other until statistical significance or using a rough rule of thumb?

Thank you!

01-06-2019 09:21 PM #24 matuloo (Legendary Moderator)

Quote Originally Posted by dtalexone View Post
I understand what you are saying and it makes sense.

It would seem also that it can get quite expensive testing different combinations. In that event, are you comparing combinations against each other until statistical significance or using a rough rule of thumb?

Thank you!
Any extensive testing will get expensive at some point, but that's normal

When running specific combos, they already have to be showing SOLID performance from the initial tests, so I'm not really comparing them to others, I just try to push them hard to see if they can hold the performance on high volume and focus on the ROI.

01-06-2019 10:30 PM #25 dtalexone (AMC Alumnus)

Your help is appreciated, @matuloo!

01-07-2019 09:59 PM #26 matuloo (Legendary Moderator)

Quote Originally Posted by dtalexone View Post
Your help is appreciated, @matuloo!
You're welcome dtalexone, anything you need ... ask away

04-22-2019 11:23 PM #27 sorreforum (Member)

Amazing..thank you for the article!

07-02-2020 03:22 PM #28 the visionary (Member)

Great post!

One more question, if you mind. When testing offers should you go only broad with your targeting? Or you can target as you like, just keep in mind that all the ad sets should be similar for all the offers to gather the relevant data?

07-03-2020 10:06 AM #29 matuloo (Legendary Moderator)

Quote Originally Posted by the visionary View Post
Great post!

One more question, if you mind. When testing offers should you go only broad with your targeting? Or you can target as you like, just keep in mind that all the ad sets should be similar for all the offers to gather the relevant data?
Depends on the traffic source too. In case there is some interest based targeting, it makes sense to utilize this and target users that might be interested into what you are offering.

In case of sources that don't have such targeting, it's better to go broad and just use the targeting options that assure you're buying the traffic the offers accept. For example, you don't want to send desktop users to offers designed for mobile users. Or android users to offers that require iOS users.

Other than that, let the creatives do the filtering so yes as you said, make sure the adsets are relevant to the offer, and stay within one vertical.

Home > >