Native Ads placement evaluation (mini-study)

[theosap]

I'll keep it short

Native Ads for EU Country
Desktop / Mobile / Tablet
Evaluated 14 Ad networks, between fraud and low volume we excluded 6, we were left with 8
Since we used a DSP, the available capacity in the form of bid requests was limited for some of the Ad networks. Yet, significant for our goal.

Those 8 Ad Networks made available yesterday 10.7 Mil Bid Requests.

We applied different filters on metrics we collected, such as user behavior, traffic sources, traffic type distribution, domain age, etc. to evaluate each placement that had over 500 Ad Bid Requests. Besides, we manually further reviewed the top 500 or so placements, and we removed any arbitrage sites, novel/book reading, nudity, ones with a bad match in regards to targeted region, etc.

That left us with 6.5 Mil Ad Bids from the original 10.7 Mil we started. Therefore, we removed from the get-go 39% of the bids inventory.

At this point, 97% of that traffic comes from just 170 domains. We originally started with 4,700 domains. We now have 170 domains to either evaluate by running some campaigns or further filter to see if there is now a match of audience.

I won't hide you that this work took more than a few hours already. Yet, it produced a list that I believe will need a data refresh once every 3-6 months to see if any other new quality sites made it to our desired pool.

 

[clearmachine]

This is an awesome post. I'm happy to read it this morning!

I have a couple of questions.

How exactly do you whitelist the widgets? I noticed most of my accounts won't allow that.... at least anymore. What networks are you able to do it?

 

[theosap]

You're welcome. They come from Voluum DSP which allows you to upload a whitelist either in the form of domain or widget. Now, I can't tell you 100% that it will run across all networks. I personally would never run any campaign to any ad network, ever, without being able to control the origin source.

Also, I find ridiculous the claims of some of the hyped media buyers saying you need to throw 5-10-15k/day to optimize a campaign in USA in some niches. That is absolute retarded. I don't doubt they do operate like that though. It's waste of time and money and most likely they do give up early on potentially profitable campaigns due to limited signals.

Bottom line, initial experimentations should take place on a laser-focused, ideally matched audience. If you can't succeed on that, you won't be able to succeed in a bigger scale.

Yesterday, I captured a glance of a video from a media buyer running an apparently popular so-called university. Now, I'm not here to question if he is a "profitable media buyer". I have no reasons to believe he is not. But the whole "for every successful campaign you see us demonstrate, you haven't seen the other 150-200 we failed"

Do you want to be teached by someone with a half percent success rate? hahh

 

[clearmachine]

Thanks for sharing theosap. Great information!

 

[klpardo]

I hope to one day be able to understand half of this post

 

[joeybabbs]

OP are you using an API to do any automation on finding your placements or any other way of automating some of that work?

 

[theosap]

We did some .csv exports of 200 rows. While the original list of 4700 domains may sound a lot, it's not. If we had to do the same for US traffic sources, we would probably need to connect to some API(s).

That will probably add at least a few hundred dollars on top of what you're already paying for Ahrefs, Semrush, or Spyfu. We happen to use all 3 of them, but for different tasks. You don't have to, for SE metrics you need 1 source, and any of those three will do it. If I recall well, Spyfu API comes with the starter package at no additional cost. For those starting or want to limit their spending, do .csv exports. Exclude very small traffic sources.

I'm anticipating a significantly higher % of fraudulent traffic when we'll analyze the US or other Tier-1 countries. I took the time to review a large percentage of the domains manually, get a better understanding of available placements per ad network, etc. It may sound excessive, but if you have an active interest in a particular country, you should consider doing the same.

What else... Instead of Google Sheets, we tend to use Airtable for similar tasks. It's visually more appealing and has more database characteristics.