Home > Paid Traffic Sources > Native

Is there Potential in this Campaign, or do i Move On? (14)

05-03-2020 07:41 PM #1 tacoloco (Member)
Is there Potential in this Campaign, or do i Move On?



Hey guys, can anyone give me some feedback on my next steps? I feel like its still early days on this campaign, but I don’t want to chase a loser too long. Is it worthwhile hacking away at this campaign?

I’ve only been testing 1 lander because its in French and I could only rip 1. I didn’t want to spend the money up front having a lander translated into French.

I’ve tested 16 creatives, but outbrain only really sent 1 of these traffic & ignored the others.

I’ve been playing around with CPC over the weekend and I have a better idea for the bottom bid that will get me traffic and how quickly I can deplete my budget.

The only optimisation I’ve done is pausing a few blatantly under performing publishers. I played with CPC a bit, but apart from that I let it run.

I haven’t tested any angles yet, just what I ripped from the spy tools.

This product is quite popular in English speaking geos, so I could get some more landers to split test, but translating it would be a PITA (doable, but enough of a hurdle to stop me at this stage).

I didn’t split test offers. I only had the one offer in French and couldn’t find any others in the same language, so I just ran with the one offer. The only thing I can think of is split testing it against the English checkout version?

What is my next move from here?
Is there any potential in this campaign, or do I cut and run?

Any feedback is welcome - THANKS!!

05-03-2020 08:34 PM #2 tacoloco (Member)
Is there Potential in this Campaign, or do i Move On?




Hey guys, can anyone give me some feedback on my next steps? I feel like its still early days on this campaign, but I don’t want to chase a loser too long. Is it worthwhile hacking away at this campaign?

I’ve only been testing 1 lander because its in French and I could only rip 1. I didn’t want to spend the money up front having a lander translated into French.

I’ve tested 16 creatives, but outbrain only really sent 1 of these traffic & ignored the others.

I’ve been playing around with CPC over the weekend and I have a better idea for the bottom bid that will get me traffic and how quickly I can deplete my budget.

The only optimisation I’ve done is pausing a few blatantly under performing publishers. I played with CPC a bit, but apart from that I let it run.

I haven’t tested any angles yet, just what I ripped from the spy tools.

This product is quite popular in English speaking geos, so I could get some more landers to split test, but translating it would be a PITA (doable, but enough of a hurdle to stop me at this stage).

I didn’t split test offers. I only had the one offer in French and couldn’t find any others in the same language, so I just ran with the one offer. The only thing I can think of is split testing it against the English checkout version?

What is my next move from here?
Is there any potential in this campaign, or do I cut and run?

Any feedback is welcome - THANKS!!

05-04-2020 12:08 AM #3 jack_l (Veteran Member)

I don't think that looks too bad...

I'd hate for you to spend more and not get it to work... but it clearly converts...

As a result I wouldn't worry too much about the translation... I'd test the images on the lp for now and leave the copy how it is since it seems to work.. even more important though I'd keep testing as many ads as possible...

And of course block bad performing widgets.. the offer payout is somewhat high so you'll have to lean on lp ctr for optimizing, but maybe block anything under 5% after 50 clicks with no conversions? Or also maybe 0 lp clicks after 25 overall clicks?

You could also clone it and set the clone at 20$ a day and block the image getting all the clicks on the main campaign so Outbrain is forced to try the other ones...

But yeah, it would be nice if it was either doing awful or great so it would be a clear choice, but that's not too bad at first on Outbrain given that you haven't blocked too many widgets yet (I don't really block any on Taboola but on Outbrain I think there are definitely bad ones that need blocked).

You can also raise the bid on your converting widgets, etc so they get more traffic.

And then as far as other things I suppose you could also target France + Quebec + Switzerland to try to get a bit more volume, but I don't know if that would make a big difference... or you could test the same offer in English on a 'scattershot' campaign targeting 10-20 geo's...

But yeah, really depends on you and your budget and goals and what you think about the offer, but I don't think there would be anything wrong with either decision - either cutting or continuing

05-04-2020 02:10 AM #4 tacoloco (Member)

Thanks for the advice Jack.
I think i'll keep plugging away it it for another $500 or so with the optimisations you suggested and see if i can improve what i've done to date

05-04-2020 08:10 AM #5 matuloo (Legendary Moderator)

CPC sources are like this, they pick the image with the highest CTR and push that one, so you need to fire up more campaigns, each with different ads, that way you will kinda force the algo into giving them at least a bit of traffic so you can see whether they stand a chance at converting or not.

I would vote for creating more LPs though, the reason is that you only have 1 and that means you really have nothing to compare the current one with. It might be good, but maybe it's not and you won't find that out without split testing another one. I'm not saying you should create a ton of them, but definitely add at least 1 or 2 more.

You're quite deep in the red, but it's quite common to see dramatically different performance with a different funnel, whatever traffic source we're talking about, so it's still worth a try in my opinion.

05-05-2020 03:18 PM #6 lemonads (Member)
Potential in this campaign

Hello Tacoloco,

The only pieces of information that is missing in your report is the origin of the sales. Are they all coming from a few publishers and or a few types of publishers? If so the fastest way to hack your campaign would be to restrict it to this type of publishers. Another question is that you say that you know the CPC that will drive you traffic but don’t name it.

Except that a quick examination of your stats reveals that :




So to quickly sum up my thoughts, there is potential in your campaign, especially if as you said this product is popular in english speaking countries, but yet there are still areas of shadows to really predict if it will be successful or not. So as far as I am concerned I would continue this campaign only if I feel that I can really scale it and that my learning budget will be in relation with my potential gains. The first step that I would take would be to optimize my creatives as it’s the most rewarding action with the minimum effort.

Last but not least you have 99% of failing by testing english offers on french market. Being french myself I know how much french people are bad at foreign languages

Hope this helps !

Jovana - lemonads

05-06-2020 08:59 AM #7 thedudeabides (Moderator)

Based on these stats I'd say probably not without some big changes. The ad CTR is too low for an ecom offer, so you're going to have much less room to lower CPCs and the LP CTR is quite low also, which should be closer to 20% than 10%

Maybe if both those things were improved you could get to break-even, but even then I think it will be tough to turn it profitable with a $35 CPA for an ecom offer in a Tier2 geo.

FYI if you want to get translations done on the cheap just use Fiverr.com will set you back maybe $15

I think your next steps would be getting some better ads, getting a translation done, and giving it another test. If the stats aren't looking much better after next $100 spend, move onto testing another offer. Just consider it the cost of education for learning what #s you need to hit to make something work.

05-06-2020 01:16 PM #8 matuloo (Legendary Moderator)

BTW: tacoloco, I just found another thread that you started with the same question and text, so I merged them into this one. Please don't start more than one thread with the same question, it will just make the replies more cluttered. Thanks

05-07-2020 05:33 AM #9 tacoloco (Member)

Quote Originally Posted by lemonads View Post
Hello Tacoloco,

The only pieces of information that is missing in your report is the origin of the sales. Are they all coming from a few publishers and or a few types of publishers? If so the fastest way to hack your campaign would be to restrict it to this type of publishers. Another question is that you say that you know the CPC that will drive you traffic but don’t name it.

Except that a quick examination of your stats reveals that :


  • A quick bayesian model of your sales says that your potential click lander to sales ratio is most likely to range from 1.5 to 3.5 percent. In any case it’s not enough for you to be profitable. As you can’t test new offer for the moment it’s not a good exit for you.
  • Regarding the other indicators :

    • Regarding your creatives, it’s strange that outbrain isn’t trying the other creatives, they should have a really low CTR compared to your main creative. In any case your CTR is quite low you can expect at least 2 times more, all will depend on how close to your product they are. Having a wider CTR wouldn’t necessarily mean that you will be more profitable, if your clicks are too far away from your product your other conversion rates will drop.
    • Regarding the click on lander ratio, it’s also a bit low, but there it will depend on the quality of your translation, on the type of devices you targeted etc, and of course the landing page itself. I would say that your profit relies on your capacity to optimize the creative CTR and the lander CTR.


So to quickly sum up my thoughts, there is potential in your campaign, especially if as you said this product is popular in english speaking countries, but yet there are still areas of shadows to really predict if it will be successful or not. So as far as I am concerned I would continue this campaign only if I feel that I can really scale it and that my learning budget will be in relation with my potential gains. The first step that I would take would be to optimize my creatives as it’s the most rewarding action with the minimum effort.

Last but not least you have 99% of failing by testing english offers on french market. Being french myself I know how much french people are bad at foreign languages

Hope this helps !

Jovana - lemonads
Thanks for the feedback Jovana.

The publishers that i converted on were:
Orange Homepage
Orange Tendance (Orange)
Outlook
Unknown

I played around with CPC between $0.60, down to $0.10
$0.30 seems to give a decent traffic, $0.20 is OK and anything under $0.20 is basically non-existant.
the sweet spot is around $0.25

I'm not familiar with the bayesian model. I assume there is a thread somewhere on STM where i can learn about it?

I've taken matuloos advice and set up a bunch of different campaigns in Outbrain, grouped by image, testing different headlines. This should hopefully give me some more conclusive results around creatives.

The lander i'm using was ripped, so i'm not sure of the quality of the translation.
Do you think it's worthwhile having my own translation done or at least checked for quality?
I've also created 3 new different variations of my existing lander with new main images and headlines i'm going to split test.

I'm going to throw another $100 or so at it with the new creatives and landers and see if i can improve my baseline, then i'll hopefully be in a better position to make some decisions.



05-07-2020 05:35 AM #10 tacoloco (Member)

Quote Originally Posted by thedudeabides View Post
Based on these stats I'd say probably not without some big changes. The ad CTR is too low for an ecom offer, so you're going to have much less room to lower CPCs and the LP CTR is quite low also, which should be closer to 20% than 10%

Maybe if both those things were improved you could get to break-even, but even then I think it will be tough to turn it profitable with a $35 CPA for an ecom offer in a Tier2 geo.

FYI if you want to get translations done on the cheap just use Fiverr.com will set you back maybe $15

I think your next steps would be getting some better ads, getting a translation done, and giving it another test. If the stats aren't looking much better after next $100 spend, move onto testing another offer. Just consider it the cost of education for learning what #s you need to hit to make something work.

Thanks, i'm going to throw some more budget at it with new ads and landers and see if i can improve against my current results.

05-07-2020 10:48 AM #11 matuloo (Legendary Moderator)

Quote Originally Posted by tacoloco View Post
I'm not familiar with the [/COLOR]bayesian model. I assume there is a thread somewhere on STM where i can learn about it?
Here is a simple bayesian calculator for you: https://marketing.dynamicyield.com/bayesian-calculator/

And more threads related to this method, with links to more calculators:

https://stmforum.com/forum/showthrea...ou-pick-offers
https://stmforum.com/forum/showthrea...est-Calculator

Caurmens tutorial on statistical significance : https://stmforum.com/forum/showthrea...nificance-quot

And to give you some food for thought, here's a great thread from "platinum" on the method, desccribing whether it's always the best one:
https://stmforum.com/forum/showthrea...-NOT-to-Use-It

05-07-2020 07:06 PM #12 tacoloco (Member)

Quote Originally Posted by matuloo View Post
Here is a simple bayesian calculator for you: https://marketing.dynamicyield.com/bayesian-calculator/

And more threads related to this method, with links to more calculators:

https://stmforum.com/forum/showthrea...ou-pick-offers
https://stmforum.com/forum/showthrea...est-Calculator

Caurmens tutorial on statistical significance : https://stmforum.com/forum/showthrea...nificance-quot

And to give you some food for thought, here's a great thread from "platinum" on the method, desccribing whether it's always the best one:
https://stmforum.com/forum/showthrea...-NOT-to-Use-It

Thanks Matuloo, i'll check out those threads.

Something i don't understand, and i'm sure this will come down to personal opinion/preference/style, is what my main goal is at this stage.
I'm spending money, testing an offer - am i looking for a minimum ROI, or just an offer that shows promise and can be improved upon?

at what stage do i give up on a campaign vs trying to optimise it and improve my ROI.

I was reading a thread about newbie mistakes and one of them was not trying enough angles & landers. This obviously doesn't apply to a bad offer, but if the offer is converting, is it worth testing further, instead of jumping to the next offer?

I've got some basic performance information on this offer using assets that i've ripped from the spy tools. I know the offer converts.
Should I try and increase performance with my own assets, or am i flogging a dead horse?

I realise there is end to optimisation and at some stage we need to be able to identify if an offer is worth pursuing.
I suppose a lot of it comes down to experience and confidence in your ability to turn a profit?

05-07-2020 09:30 PM #13 platinum (Veteran Member)

Quote Originally Posted by tacoloco View Post

Something i don't understand, and i'm sure this will come down to personal opinion/preference/style, is what my main goal is at this stage.
I'm spending money, testing an offer - am i looking for a minimum ROI, or just an offer that shows promise and can be improved upon?

at what stage do i give up on a campaign vs trying to optimise it and improve my ROI.

I was reading a thread about newbie mistakes and one of them was not trying enough angles & landers. This obviously doesn't apply to a bad offer, but if the offer is converting, is it worth testing further, instead of jumping to the next offer?

I've got some basic performance information on this offer using assets that i've ripped from the spy tools. I know the offer converts.
Should I try and increase performance with my own assets, or am i flogging a dead horse?

I realise there is end to optimisation and at some stage we need to be able to identify if an offer is worth pursuing.
I suppose a lot of it comes down to experience and confidence in your ability to turn a profit?
What I've seen happening quite frequently is that many affiliates tend to promote their offers with only what they see on spy tools. Thing which might become a costly mistake.
In general you might want to start testing an offer with what you see on spy tools just to get a rough idea on the offer it it's worth keeping active or not. At this stage you can quickly see the rough performance of the offer. Metrics like conversion rate and earnings per click are a good indicator as a start.

Once you have identified a potentially good offer after the initial test, you can start investing more time on optimizing your funnel along with other campaign optimizations. Regarding native, you can optimize your funnel on two levels:
- Ad Image and Headline
- Landing Page

The goal with introducing new ad images and headlines is to find better ads in terms of ad ctr level. This will first satisfy the ad network allowing them to deliver higher ctr ads on their publisher placements. At the same time having above average ctr ads, you can get your clicks for cheaper.

At the same time, you might want to optimize your landing page level, by testing new variations. Basically, in terms of variations you will see that playing around with images, angles and the copy of the landing page, you can significantly boost your landing page click-through rate as well as the offer conversion rate.

Most affiliates fail to properly optimize their landing pages, but the bottom line is that if you have a poor performing landing page, your publisher and ad level optimization will be highly influenced by that.

05-08-2020 12:10 AM #14 tacoloco (Member)

Quote Originally Posted by platinum View Post
What I've seen happening quite frequently is that many affiliates tend to promote their offers with only what they see on spy tools. Thing which might become a costly mistake.
In general you might want to start testing an offer with what you see on spy tools just to get a rough idea on the offer it it's worth keeping active or not. At this stage you can quickly see the rough performance of the offer. Metrics like conversion rate and earnings per click are a good indicator as a start.

Once you have identified a potentially good offer after the initial test, you can start investing more time on optimizing your funnel along with other campaign optimizations. Regarding native, you can optimize your funnel on two levels:
- Ad Image and Headline
- Landing Page

The goal with introducing new ad images and headlines is to find better ads in terms of ad ctr level. This will first satisfy the ad network allowing them to deliver higher ctr ads on their publisher placements. At the same time having above average ctr ads, you can get your clicks for cheaper.

At the same time, you might want to optimize your landing page level, by testing new variations. Basically, in terms of variations you will see that playing around with images, angles and the copy of the landing page, you can significantly boost your landing page click-through rate as well as the offer conversion rate.

Most affiliates fail to properly optimize their landing pages, but the bottom line is that if you have a poor performing landing page, your publisher and ad level optimization will be highly influenced by that.

thanks for the insight

Home > Paid Traffic Sources > Native