Home > Paid Traffic Sources > Native

---

Do native algorithms really want us to make a lot of money? (18)

---

Strange question and I wanted to clarify that I am not a conspiracy theorist!

But I'm bring you some data:

- Campaign without any conversion tracking on the platform, without pixels, blindly in a nutshell: 200% ROI (all widgets open, even the most spammy)

- Same campaign with conversion tracking and widget optimization: 50% ROI

Clarification:
COD Ecomm product
Price: 90-100 €
Payout: € 20 per lead

First blind campaign:
(200% ROI) running in November, December

Second campaign (50% ROI) running in spring

In my opinion, the time of year may have influenced, around Christmas the ecomm products sell well.

But isn't that a bit absurd?

To clarify: I already knew that the campaign was working well, before launching it blindly...

I would like to know some opinions, I do not want to state that native networks work against those who convert "too much", but sometimes when the CPA Is fantastic, comes a period in which everything slows down or almost freezes.

Are these algorithms LAZY or a little CRAFTY?

Sometimes simply running copies of the exact same ADS helps convert better ... Same teasers and images... Clone campaigns, same thing...

I need opinions

---

I don´t run Natice traffic but I´ve seen such things on Push as well that I can´t really explain or that don´t make sense.

For example I run a campaign on CPA Goal and it´s doing very good.

Then I duplicate the campaign and run it with a 50% higher bid but I barely receive any traffic in the high bid campaign.

Had this on Push as well.

I had campaigns that were doing good and then suddenly the traffic stops almost completely.

When I kept them running it happenes that a few weeks later suddenly traffic volume increases alot and the campaigns are back.

Or I start a campaign with broad targeting that performs much better than the same funnel with tighter targeting that matches the funnel much better.

---

Yeah is absurd... The comparison with betting agencies came to me, that is, if you earn too much, you become a problem for them because they lose money ..

But in this case the advertising platforms do not lose money, which is why they should support because they also benefit from it.

Youtube, on the other hand, thinks differently: if your video works, they give you a push to the moon...

---

Well yes, youtube is different since they are a content platform and it's in their best interest to push quality and engaging content. Even when it comes to Ads, whatever their users respond better to is better for them to have on the platform as it generally makes the users happy and stay on the platform.

Ad networks generally care just about the revenue per mile, so how much they can earn from each ad impression. Their ultimate goal is to maximize their revenue, not yours.

It's quite possible that they have set certain threshold that labels the campaign as "good enough" and once that threshold is reached, the algo would feed you some lower quality traffic too. I mean, people usually don't complain about 50% ROI, so maybe they aim for something like that and than focus on just selling you more traffic while still maintaining the still healthy ROI of 50%... more money for them, still decent results for you It would make sense for sure, but we're just speculating here of course.

But if this is the case, you can try to game them a bit... lower the conversion value that you report to them! That would lower the ROI and the algo would have to try harder... this should be a simple test and the results should be visible quite fast, how about trying that?

The other option is that the algo simply isn't advanced enough and it fails at delivering the right traffic. Maybe your skills are actually better than their AI and trying to rely on it just hurts your overall performance.

---

Well, there are many features that affect to ad-campaigns. As you mentioned, the time of the year is one of them. If you could identify what was the most effective channel/site in your previous ad-campaign, it would be easier to optimise another campaigns.

---

I missed this. I have an opinion:

It depends on the Native traffic source.

This is my experience, for whatever the fuck that's worth...

For revcontent, it doesn't matter what the fuck you do, you can pass back 0 conversions or 1000000 conversions, traffic is the same and the algo doesn't optimize for conversions at all, just clicks and money. I pass back conversions there anyway as it's more convenient to adjust bids and creatives. Otherwise, I probably wouldn't bother.

For Taboola and OB, with the smartbid stuff, it's very similar to FB. If you don't pass back any conversions, they will start throttling your traffic and you won't get anything. Basically have to passback conversions and play with Fixed bid a lot to tame certain sites. Really annoying balancing act.

For MGID, they absolutely fuck you over if you pass conversions back. I've had many campaigns do amazingly well without passing back conversions, and then once I started to do so (at reps pleading request so they could 'optimize', they all mysteriously took a gigantic dump to keep my camps at barely breaking even or less. I don't pass back conversions to them anymore.

---

I have heard it said a number of times that MGID has their own internal buying team and you will end up competing against them. I suspect they wanted that data to optimize their own campaigns.

---

Allow me to add my two cents here...

Being in a position who works everyday with both the buying and selling side of native traffic I can confirm the following.

All native ad networks are quite aware of the fact that if the campaigns don't really ad up for their advertisers, they will very likely stop using their platform. In other works, no results - no more ad spend.

At the same time, all native ad networks need to make things work for their traffic suppliers, aka content websites. As @matuloo mentioned above, sites get paid on a RPM basis, meaning that they have to make it worth that for every 1000 ad impressions, the publisher site should get $X-$XX in revenue, otherwise they will start looking for alternative monetization options.

Bottom line, as "the man in the middle", native ad networks need to make things work on both ends, otherwise they can get out of business quite easily.

To put things under balance between both sides, most native ad networks optimize towards a specific CTR / vCTR threshold, which in most cases if not reached yields a higher CPM. This way they can at least make both sides happy while getting their share. Once CTR and CPMs are settled, the traffic source can start looking for additional optimizations like conversions or other engagement metrics like page views and stuff.

MGID is not the only traffic source that might have an internal media buying team. Other native ad networks do have them too, and that's not necessarily a bad thing. It's worth noting that these media buying teams in most cases provide managed services for big brands or clients that don't have the right media buying experience and expertise for native.

Also, when various native ad networks internal media buying teams buy within their own platform, they buy the same way all of us buys. So it's pure competition on a real-time bidding model. You can imagine how much traffic are these ad networks moving daily. If they were to screw things up this easily, would have switched to closed ad networks or just play it safe with reselling remnant traffic to someone else. Not to mention also the budgets they would need to block to own the whole inventory.

Getting back to conversion data vs. no conversion data, this is honestly a tricky part. I've seen many native campaigns with brand awareness objective, yielding way better results than conversion campaigns, like I've seen also the opposite. There's no guaranteed results in neither one. The difference when running conversion objective campaigns, is that the traffic source algorithm will try to serve your ads to where people that convert to it are most likely to do so. This using network-wide data or campaigns specific data - this last one is obviously to be treated as a black-box since no one is willing to shed light on it.

The other problem I often see with many native campaigns, is that many people want to always play it safe while keeping their budgets to the bare minimum while promoting high ticket products.

I know and I fully agree about the fact that it has to work for me in order to scale things up. However, when considering that someone else is promoting a similar product, with a higher bid (but not necessarily) and a waaayyy higher budget; me as a traffic source will prioritize the higher spender requests over the lower ones. It's simple math after all, nothing personal.

I believe the best way to deal with these is be straight forward with your reps right from the beginning like: "If things work, I can scale to $XXXX /day if not, don't bother me to continue spending with you".

When doing that, keep in mind that you have to be ready to show what you promised in order for them to continue collaborating moving forward.

We all try focusing on what's bringing most business to our pockets, and odds are that the others we're working with are doing the same too.

---

Thanks for your reply @platinum, very valuable!

One question if I may since this has been buggin me for years

Do you know this for a fact? I mean the part that they simply use the platform when mediabuying themselves?

When handling the campaigns for big brands, it totally makes sense that they would do it this way. But what about the internal teams that actually run campaigns themselves? Somehow I believe they fuck with the system in one way or another as they are also getting the commissions from the networks so they can cause some initial loss to the network by buying the traffic at a lower rate than the actual competition.

I cant help myself but I don't like the fact that traffic networks have their internal media buying teams too... not that I could do anything about it except for bitching a bit though

---

And not only high ticket. This is a hard for a lot of media buyers, and agencies, to understand as well. Many are only aware of how FB traffic works. So it blows their mind that you can literally spend $10k on a Native campaign before breaking even. Most expect instant resuls with a budget of $100 a day. There's just so many damn sites on Tier 1 Native, that you can spend $3k on random sites in one day (don't do this) before seeing one $40 conversion.

Smart Bid, Target CPA, Target ROAS, etc. attempt to help with this, but they're nowhere near perfect and still need lots of human intervention, and still require lots of spend. Unfortunately, I have no t found a way around this on Native. It's all a numbers game. And they're big fucking numbers.

---

Yeah, as far as I know most internal media buying teams buy ads from the same platform UI the average advertiser does.

Apart from handling campaigns for big brands, they also handle campaigns as an agency for smaller brands with considerable budgets.

This is a pretty normal business practice I honestly see nothing wrong about. After all, in most cases they are promoting completely different funnels, which shouldn't scare affiliates at all.

What I would worry mostly about is offer networks running their own media buying activity. Allow me to lay down two simple reasons.

First, as opposed to the ad networks that don't necessarily know all the details about a specific offer; offer networks know a hell lot more about what's going on with that offer.

Second, aside from having a way bigger margins than their average or top affiliates do, they also have huge traffic insights. After all a lot of them require passing traffic source data, as they would need it for optimization and compliance proposes.

Third, if not enough volume is coming from their network of affiliates they have to compensate it themselves in order to keep access to that offer and keep their clients happy. I guess we've seen enough offers with extremely aggressive angles being run by the offer networks throughout the years. And the funny part is that they wouldn't allow anyone run such angles due to compliance concerns.



This is the sad truth indeed. The average native campaign on a Tier 1 and some Tier 2 countries will get traffic from around 1500-3500 publisher sites on average, and this when running on a conservative bid.

It's not hard to understand what to expect from a $20-$50 daily budget campaign.


Building an algorithm that works for everyone's needs is not an easy task.

It's virtually impossible to find what works best for Offer XYZ after $100 ad spend. There are hundreds if not thousands of offers that are competing at the same time under the same network, and all of them have different landing pages, different conversion flows, payouts, everything.


I think its best to spend our energies and focus on how to make things work on our interest, rather than stressing ourselves about what traffic and offer networks are doing on their end.

---

As long as they just act as agencies that manage campaigns for specific clients, that's fine. These are likely deals where they charge some fee based on the Ad spend and they probably have to meet some KPI in order for the client to stay with them.

What I meant was traffic networks having internal media buying teams that actively look for offers, attend conferences with the goal to make deals with offer owners or affiliate networks and then promoting their stuff. This was pretty common in adult and even though I have no proof for this, I have lost a few campaigns because someone from the traffic source stole my funnel and "booked" some placements that were working really well for me.

But I guess the likes of Outbrain or Taboola are too big to pull shit like this... with the smaller ones, I wouldn't be surprised

You are right with the networks providing offers, they can pull some nasty shit too. I've experienced my share of weird calls from them such as random kicks from an offer when it was obviously working super good just a week ago etc...

Same with the traffic sources.... they reject more aggressive creatives while there are many just like the rejected ones running on the platform... and if you show them, silence follows

---

It's like domain parking providers that prohibit arbitrage, but do it themselves on all platforms.

In Italy on native I have to compete with all those stupid ads on SUVs that are given away for free...

Contacted one of them via email, "thank you for reporting ..."

But SUVs are still given free to all Italians

---

crazily a lot of golden nuggets here in this thread. I used to run native too and barely breakeven with 10k spent on mgid. i listened to all of the rep's advice and there's just no way for me to get past that breakeven point, it's ridiculous every single optimization i made doesnt work, lol. I can see deeply now how these networks can screw advertisers over, and definitely native is not something for small affs like me

---

This is actually quite common with many verticals or setups. Take dating (my favorite vertical) for example... as long as you know what you are doing, almost every test will end up close to the breakeven point or in small profit, but taking it to let's say 30%+ is much harder.

You need to make THAT ONE tweak to make it happen... it can be a better offer, small change in the copy, slightly better Ads. At first, it seems almost impossible and then it happens

The reason for this, I believe, is that pretty much everyone is able to rip a campaign or construct a decent funnel so most people are reaching a similar level of returns, which actually sets the bids in the end. To put it simple, the current bids in a specific target are usually set based on how much the average ROI is, the market kinda takes care of this. Those who can beat this % are in profit.

So it doesn't necessarily have to mean that the networks are somehow screwing the advertisers over... not that it couldnt be the case of course, but I don't think it's universal.

---

We've ran tons of native over the years and my observation is if you start cutting placements, they will kill your great placements too.

In short, they want you to BUY ALL.

About reps, 99% are clueless, they went through training looking at slides and hearing theories like at university from a professor, but never ever ran anything themselves to understand what the 'real world is like'. Hence why their suggestions don't work.

Every single native platform is different on how it behaves, if you spend a lot of money on testing you will start feeling it. Ie on taboola, excluding placements meant they cut down your premium spots bringing the profits, decreasing bids was a better way to go... or redirecting those placements and selling their shitty ass traffic and taking a tiny loss even better. There's so many tricks, but it takes time to master a source. A friend did 8 figures spend (not main stage speaker at AWA/AWE, no course, no agency to sling but a true underdog).. and he really knows insane tricks based on so much experience he built over the years.

---

I never thought of that. I feel dumb now. Off to set up rules in tracker now:

If site =/= MSN then redirect to adult smartlink.

I’m sure Taboola won’t mind

---

I've seen the same with many networks. The more you try to isolate the best spots and target just those, the less you are loved by the algo. It makes sense from the networks point of view but not so much for an affiliate

---

Home > Paid Traffic Sources > Native