Sources of traffic for arbitrage?

[peakaboo]

Hi guys! I've been lurking here and there for a while without creating an account and that's the reason I lost traces to some good articles that I read before.

Does anyone remember a thread that summarized all good sources of traffic to buy for arbitrage? Or can you suggest me again a finite list? Thanks

 

[iconnek]

perhaps @se900se can help

 

[Badboy Enterprise]

and you think that they are going to share to you the money pot?

 

[Deleted member 813988]

The sources can be very different, a lot depends on your budget and final goal.

This is the list of 100 ad networks, you can buy traffic from them, as well as you can open publisher account with them and monetize your traffic: https://www.blackhatworld.com/seo/top-100-ad-networks-adsense-alternatives.1045408/

You can go for social media ads, depending on your budget and content, or run social media ads and/or promotions.

 

[on1]

The sources can be very different, a lot depends on your budget and final goal.

This is the list of 100 ad networks, you can buy traffic from them, as well as you can open publisher account with them and monetize your traffic: https://www.blackhatworld.com/seo/top-100-ad-networks-adsense-alternatives.1045408/

You can go for social media ads, depending on your budget and content, or run social media ads and/or promotions.

Do you like any specific networks for insurance?

Besides Google/Bing/Social/Taboola

 

[Deleted member 813988]

Do you like any specific networks for insurance?

Besides Google/Bing/Social/Taboola

I would run such a campaign among multiple ad networks and ad exchanges, e.g. I wouldn't target particular ad network but run it on 50-60 ad networks at the same time. This would require higher budgets but it could bring better results. A lot depends not on just specific ad network, but on the audience that you build and show your ads to.

 

[on1]

I would run such a campaign among multiple ad networks and ad exchanges, e.g. I wouldn't target particular ad network but run it on 50-60 ad networks at the same time. This would require higher budgets but it could bring better results. A lot depends not on just specific ad network, but on the audience that you build and show your ads to.

What's your retargeting process?

 

[AdvertisingGuy]

The sources can be very different, a lot depends on your budget and final goal.

This is the list of 100 ad networks, you can buy traffic from them, as well as you can open publisher account with them and monetize your traffic: *url removed*

You can go for social media ads, depending on your budget and content, or run social media ads and/or promotions.

Most of the companies on that list are supply-side platforms (SSPs). They are essentially ad inventory wholesalers that sell banner, video and native ads to demand-side platforms (DSPs). Very few of the companies listed have a UI where you can book campaigns, or at least not without significant upfront financial costs (AppNexus) or filling out an insertion order. Most of those companies are inappropriate for ad arbitrage, but might work for CPA offers.

@peakaboo if you want to buy traffic for ad arbitrage, you have native networks (Outbrain, Taboola, ZergNet, RevContent, others), social media sites (Facebook, Twitter, snapchat, others), and incentivized / toolbar traffic and a few other services, some of which can be found on this forum. But you must be very strict on the quality of traffic you're getting--if you get bot traffic, even accidentally, you could be blacklisted from the exchanges and have your revenue withheld.

 

[Deleted member 813988]

Most of the companies on that list are supply-side platforms (SSPs). They are essentially ad inventory wholesalers that sell banner, video and native ads to demand-side platforms (DSPs). Very few of the companies listed have a UI where you can book campaigns, or at least not without significant upfront financial costs (AppNexus) or filling out an insertion order. Most of those companies are inappropriate for ad arbitrage, but might work for CPA offers.

@peakaboo if you want to buy traffic for ad arbitrage, you have native networks (Outbrain, Taboola, ZergNet, RevContent, others), social media sites (Facebook, Twitter, snapchat, others), and incentivized / toolbar traffic and a few other services, some of which can be found on this forum. But you must be very strict on the quality of traffic you're getting--if you get bot traffic, even accidentally, you could be blacklisted from the exchanges and have your revenue withheld.

That's right, companies that allow you to run your ads among multiple ad networks in most cases require big budgets + monthly payment + they also take percentage of your revenue depending on how you run a campaign and what sources you use, therefore I mentioned that the traffic source depends a lot on what your budget and goal is. @on1 mentioned he's looking to get traffic for insurance - therefore I mentioned this option as I run multiple campaigns between 2k and 30k per day on one of such platforms.

For everything else, you can simply use any network from the list I linked above or social media and you're good to go

 

[peakaboo]

Most of the companies on that list are supply-side platforms (SSPs). They are essentially ad inventory wholesalers that sell banner, video and native ads to demand-side platforms (DSPs). Very few of the companies listed have a UI where you can book campaigns, or at least not without significant upfront financial costs (AppNexus) or filling out an insertion order. Most of those companies are inappropriate for ad arbitrage, but might work for CPA offers.

@peakaboo if you want to buy traffic for ad arbitrage, you have native networks (Outbrain, Taboola, ZergNet, RevContent, others), social media sites (Facebook, Twitter, snapchat, others), and incentivized / toolbar traffic and a few other services, some of which can be found on this forum. But you must be very strict on the quality of traffic you're getting--if you get bot traffic, even accidentally, you could be blacklisted from the exchanges and have your revenue withheld.

Thanks a lot for the list. It's exactly what I am look for. Any suggestion on how to manage the quality of traffic I bought from these providers (native networks + social media sites) ?

That's right, companies that allow you to run your ads among multiple ad networks in most cases require big budgets + monthly payment + they also take percentage of your revenue depending on how you run a campaign and what sources you use, therefore I mentioned that the traffic source depends a lot on what your budget and goal is. @on1 mentioned he's looking to get traffic for insurance - therefore I mentioned this option as I run multiple campaigns between 2k and 30k per day on one of such platforms.

For everything else, you can simply use any network from the list I linked above or social media and you're good to go

I'd start small with limited budget so couldn't start with many providers/campaigns at the same time. Do you have any suggestion when starting with a small budget (e.g. < 100$ a day) ?

 

[AdvertisingGuy]

That's right, companies that allow you to run your ads among multiple ad networks in most cases require big budgets + monthly payment + they also take percentage of your revenue depending on how you run a campaign and what sources you use, therefore I mentioned that the traffic source depends a lot on what your budget and goal is. @on1 mentioned he's looking to get traffic for insurance - therefore I mentioned this option as I run multiple campaigns between 2k and 30k per day on one of such platforms.

For everything else, you can simply use any network from the list I linked above or social media and you're good to go

This is 100% correct. Would you mind sharing what DSPs you use? I'm curious.

Thanks a lot for the list. It's exactly what I am look for. Any suggestion on how to manage the quality of traffic I bought from these providers (native networks + social media sites) ?



I'd start small with limited budget so couldn't start with many providers/campaigns at the same time. Do you have any suggestion when starting with a small budget (e.g. < 100$ a day) ?

"Quality" in this case can mean a few different things. If you mean "quality" as in "human / not bot," first you must be able to differentiate between a human and a bot over the Internet. How can you tell them apart? Ad verification companies use JavaScript to collect information of a browser and/or an ad viewer. I could write thousands of words on this but some major flags are:

• Automated Browser: in JavaScript this is navigator.webdriver = True. This is 100% a bot.
• Data Center IP address: if the visit is coming from Amazon, Microsoft, Google, DigitalOcean, etc, then it's 100% a bot.
• Referrer: Where is this visitor coming from? Since this would be a native network, referrer would be Outbrain or someone similar. When going to multiple pages of a site, a real visitor will have a document.referrer on each visit.
  
• Spoofed User Agent: Is it a Chrome browser pretending to be Firefox? This is done by looking at the Fonts the browser has enabled, Canvas fingerprinting, among other things.
• Valid Graphics Card: Is the user visiting with a valid graphics card or is it a headless browser? This is analyzed using the WebGL protocol. If it's headless, the graphics card will be something like "Google SwiftShader." If it's valid, it will be something like "AMD Radeon HD 6970M OpenGL Engine."
• Visitor Behavior: What does the user do once it hits the page? If it's a slideslow, the normal behavior is to go from page1 to page2 to page3, etc. If they go from page 1 to page 5 to page 30 then it's a bot. Is the user clicking around the page? If you have HotJar or CrazyEgg, does the visitor follow mouse movement / scroll patterns the same way as the rest of your visitors do? There are tons of ways to analyze this.

To manage the quality of your traffic, you can use a tracking pixel that would collect all of this information and have your click redirect to hit this pixel first. So it would be NativeNetwork->Your Pixel->Landing page. You'll also want to have a tracking pixel to determine what the user does once he hits your landing page. Some will bounce off, but if your content is engaging and your page isn't too cluttered with ads, they will continue. If you see an automated browser with a data center IP address clicking clicking on an outbrain link and visiting your site (or if they bounced before ever getting to the first page), you could try and get your money back for the click. But the only way to do that is to collect data on every one of your clicks.

In my experience, social networking sites are pretty clean.

 

[peakaboo]

"Quality" in this case can mean a few different things. If you mean "quality" as in "human / not bot," first you must be able to differentiate between a human and a bot over the Internet. How can you tell them apart? Ad verification companies use JavaScript to collect information of a browser and/or an ad viewer. I could write thousands of words on this but some major flags are:

• Automated Browser: in JavaScript this is navigator.webdriver = True. This is 100% a bot.
• Data Center IP address: if the visit is coming from Amazon, Microsoft, Google, DigitalOcean, etc, then it's 100% a bot.
• Referrer: Where is this visitor coming from? Since this would be a native network, referrer would be Outbrain or someone similar. When going to multiple pages of a site, a real visitor will have a document.referrer on each visit.
  
• Spoofed User Agent: Is it a Chrome browser pretending to be Firefox? This is done by looking at the Fonts the browser has enabled, Canvas fingerprinting, among other things.
• Valid Graphics Card: Is the user visiting with a valid graphics card or is it a headless browser? This is analyzed using the WebGL protocol. If it's headless, the graphics card will be something like "Google SwiftShader." If it's valid, it will be something like "AMD Radeon HD 6970M OpenGL Engine."
• Visitor Behavior: What does the user do once it hits the page? If it's a slideslow, the normal behavior is to go from page1 to page2 to page3, etc. If they go from page 1 to page 5 to page 30 then it's a bot. Is the user clicking around the page? If you have HotJar or CrazyEgg, does the visitor follow mouse movement / scroll patterns the same way as the rest of your visitors do? There are tons of ways to analyze this.

To manage the quality of your traffic, you can use a tracking pixel that would collect all of this information and have your click redirect to hit this pixel first. So it would be NativeNetwork->Your Pixel->Landing page. You'll also want to have a tracking pixel to determine what the user does once he hits your landing page. Some will bounce off, but if your content is engaging and your page isn't too cluttered with ads, they will continue. If you see an automated browser with a data center IP address clicking clicking on an outbrain link and visiting your site (or if they bounced before ever getting to the first page), you could try and get your money back for the click. But the only way to do that is to collect data on every one of your clicks.

In my experience, social networking sites are pretty clean.

Wow thanks for the insight. At the beginning I thought traffic would be safe if I buy it from popular native networks like outbrain / taboola / etc but it seems not be the case.
I started to doubt if traffic arbitrage is profitable if we need to buy it from these (expensive) networks + invest in the quality management

 

[Deleted member 813988]

Thanks a lot for the list. It's exactly what I am look for. Any suggestion on how to manage the quality of traffic I bought from these providers (native networks + social media sites) ?

Each network has some percentage of low-quality traffic - that’s normal. It doesn’t have to be bots, though, accidental clicks are also don’t bring any conversion.

It terms of ‘managing’ the quality - targeting correct audience/users, on correct sources at the right time will bring you quality that you need. Targeting through ad network can be more detailed and precise than FB targeting (depending on the ad network you work with).

Ad networks have filters to check for bots, as well as you can have filter on your site as well. Depending on what your size is and what you do - you might want to invest into that (could be quite expensive).

Wow thanks for the insight. At the beginning I thought traffic would be safe if I buy it from popular native networks like outbrain / taboola / etc but it seems not be the case.
I started to doubt if traffic arbitrage is profitable if we need to buy it from these (expensive) networks + invest in the quality management

Popular at networks popular for a reason, and they do everything (a lot) what they can to prevent low quality traffic or bot, however setting ads is something that you have to do yourself. Low quality traffic can be just bad ad settings.

If Traffic Arbitrage wasn’t profitable it wouldn’t be that big these days and wouldn’t have so many sites doing that!

 

[AdvertisingGuy]

Wow thanks for the insight. At the beginning I thought traffic would be safe if I buy it from popular native networks like outbrain / taboola / etc but it seems not be the case.
I started to doubt if traffic arbitrage is profitable if we need to buy it from these (expensive) networks + invest in the quality management

Outbrain / Taboola are expensive networks. CPCs on these platforms can start (on the low-end) for USA Desktop clicks at a $0.06-$0.12. And even some say that might be low. I think your standard native CPC across all native networks (meaning MSN, Yahoo Gemini, others) would be around $0.17. Clicks on Reddit from certain subreddits can be up to $0.40 CPC. Native networks are generally speaking cleaner from my experience, but there's no guarantee. "Trust but verify." You should be checking every click for bot signatures and comparing the clicks/visits to your G***** Analytics or whatever analytics provider you're using to make sure the visits actually get on the page and do something once they're there.

That's not to say that incentivized or toolbar traffic is bad / non-human per se. It will be cheaper ($0.03-$0.05) but you have to vet them carefully. I've tested dozens of traffic sources, most of them were crap and very easily identifiable as such. At some point I will write a post that goes more into depth on the bullet points I listed above.

 

[AdvertisingGuy]

Targeting through ad network can be more detailed and precise than FB targeting (depending on the ad network you work with).

This point deserves an additional comment.

The three behemoths of online advertising at this moment are Google, Facebook and Amazon (who is well behind the first two).

• Google knows what you want (via search queries / Google Analytics)
• Amazon knows what products you're interested in and what you want to buy (based on product search / purchase history)
• Facebook knows who you are (based on likes / analytics from Facebook like buttons all over the Internet).

Broadly, Facebook has some of the most effective audience based targeting ever invented. But @se900se point still rings true. Let's use @on1 's insurance example. There are certain websites made specifically for people that are looking for information about insurance or insurance products. Or people that are looking to become insurance brokers or something like that. A well-placed banner or native ad on one of these sites is often more effective than Facebook's audience targeting because when the user is on that website, they are specifically looking for information about insurance. Meaning they are especially receptive from an advertiser with authority, and said advertiser has the authority because they're running ads on that site!

Obviously the ad copy of the banner and landing page post-click need to look legitimate as well, but ad network buying can oftentimes be better than buying Facebook ads. And not everyone knows that.

Ad networks have filters to check for bots, as well as you can have filter on your site as well. Depending on what your size is and what you do - you might want to invest into that (could be quite expensive).

PerimeterX, Dstil Networks and Encapsula are some companies that will filter bot traffic. They will ID a visitor coming to a page and if the visitor is a bot, they will serve the bot a broken page. One could feasibly build this themselves using the most common bot behaviors listed in my earlier comment. I've done it and it works decently well to keep easily ID'ed bots off a site.

 

[Deleted member 813988]

This point deserves an additional comment.

The three behemoths of online advertising at this moment are Google, Facebook and Amazon (who is well behind the first two).

• Google knows what you want (via search queries / Google Analytics)
• Amazon knows what products you're interested in and what you want to buy (based on product search / purchase history)
• Facebook knows who you are (based on likes / analytics from Facebook like buttons all over the Internet).

Broadly, Facebook has some of the most effective audience based targeting ever invented. But @se900se point still rings true. Let's use @on1 's insurance example. There are certain websites made specifically for people that are looking for information about insurance or insurance products. Or people that are looking to become insurance brokers or something like that. A well-placed banner or native ad on one of these sites is often more effective than Facebook's audience targeting because when the user is on that website, they are specifically looking for information about insurance. Meaning they are especially receptive from an advertiser with authority, and said advertiser has the authority because they're running ads on that site!

Obviously the ad copy of the banner and landing page post-click need to look legitimate as well, but ad network buying can oftentimes be better than buying Facebook ads. And not everyone knows that.



PerimeterX, Dstil Networks and Encapsula are some companies that will filter bot traffic. They will ID a visitor coming to a page and if the visitor is a bot, they will serve the bot a broken page. One could feasibly build this themselves using the most common bot behaviors listed in my earlier comment. I've done it and it works decently well to keep easily ID'ed bots off a site.

That’s right! Another thing in advertising is that you can target right user but on the wrong site, e.g. when someone is on entertainment site he’s unlikely to think about insurance. Blocking sources as well as targeting specific ones is another big part of advertising process.

For the filters - many networks use IAS, DV, Moat and other filters. Filter can just prevent ads from being shown, however, most of the ad networks will just send you an email if they feel you get the bot traffic. In many cases (for small publishers) spending lots of money on filters can be avoided by simply checking your analytics for traffic sources as well as traffic behavior.

 

[AdvertisingGuy]

That’s right! Another thing in advertising is that you can target right user but on the wrong site, e.g. when someone is on entertainment site he’s unlikely to think about insurance. Blocking sources as well as targeting specific ones is another big part of advertising process.

Yes to this! 100 times!

For the filters - many networks use IAS, DV, Moat and other filters. Filter can just prevent ads from being shown, however, most of the ad networks will just send you an email if they feel you get the bot traffic. In many cases (for small publishers) spending lots of money on filters can be avoided by simply checking your analytics for traffic sources as well as traffic behavior.

This is someone who knows what they're talking about! For some additional background--Moat / IAS / DV (DoubleVerify) are all large companies (that were acquired by larger companies) and are very expensive to use--IAS is on the cheap side and it's around $2K-$3K / month. I've heard Moat quoted at $5K/mo. I've heard rumors that WhiteOps costs $25K/month, but I doubt it's true. I wonder if a few few publishers can band together / pool their resources and their own Moat account?

If you're clever, you can have your own ad verification pixel that works 90% as well as Moat for much cheaper and will get more data on what's bot or not. I'll show some data from my own pixel at some point.

 

[Deleted member 813988]

Yes to this! 100 times!



This is someone who knows what they're talking about! For some additional background--Moat / IAS / DV (DoubleVerify) are all large companies (that were acquired by larger companies) and are very expensive to use--IAS is on the cheap side and it's around $2K-$3K / month. I've heard Moat quoted at $5K/mo. I've heard rumors that WhiteOps costs $25K/month, but I doubt it's true. I wonder if a few few publishers can band together / pool their resources and their own Moat account?

If you're clever, you can have your own ad verification pixel that works 90% as well as Moat for much cheaper and will get more data on what's bot or not. I'll show some data from my own pixel at some point.

The pricing is crazy -, the cheapest was Forensiq ($1.5K p/m) Others vary from $3k to $15K - I don’t know how they are able to justify that, but many companies prefer to pay for ‘piece of mind’.

 

[AdvertisingGuy]

The pricing is crazy -, the cheapest was Forensiq ($1.5K p/m) Others vary from $3k to $15K - I don’t know how they are able to justify that, but many companies prefer to pay for ‘piece of mind’.

This might be a little "inside baseball" off-topic, but the ad verification industry is equal parts mafia and CYA (Cover your @$$).

Mafia:
Here's how it works. If you work at a big publisher, you'll get an email from a seller at Moat (all companies work the same way, but we'll use Moat) saying that "our clients were buying on your site and saw that you had a high percentage of fraud and ads that weren't in view. If you don't improve things, you'll get kicked off the buy / blacklisted / banned. Can we take some time to discuss?" They're basically saying "Nice website you have there... shame if something were to... happen to it."

And lo and behold, the only way to fix your problems is to work with Moat. Meanwhile if you have an ad verification provider already and that provider says that you're generally clean, you'll still get these e-mails, because everyone has their own definition of fraud (or everyone is lying in the first place).

CYA:
So here, a big media buying agency (let's say GroupM, but most agencies work the same way) says that your ads need to be 100% in-view for two seconds, otherwise we'll demand make goods, lower CPMs, cancel our buys, etc. And since we use Moat, we expect you to use Moat as well, otherwise our numbers won't match up. This of course leads to even more confusion because how ad verification pixels are implemented will vary from site to site and advertiser to advertiser and sometimes, two Moat pixels running on the same ad space will give widely different viewability numbers! But if you work with Moat, there is a "CYA" element to things, and if you're working with a vendor, and then vendor says you're clean, then you're generally clean. At least according to people that care, and that counts most of the ad networks and G***** (maybe).

Now, that's not to say that the companies themselves are totally fraudulent. If you're seeing 100% bot traffic from a traffic source according to Moat, most other services will give you the same data because they all generally work the same way. But the secret sauce isn't all that special. And the "piece of mind" does count for something, especially if you have no idea how ad fraud works in the first place.

Anyway, you're a really bright guy who is really knowledgeable about this stuff. Not that you needed me to tell you that.

 

[peakaboo]

Each network has some percentage of low-quality traffic - that’s normal. It doesn’t have to be bots, though, accidental clicks are also don’t bring any conversion.

It terms of ‘managing’ the quality - targeting correct audience/users, on correct sources at the right time will bring you quality that you need. Targeting through ad network can be more detailed and precise than FB targeting (depending on the ad network you work with).

Ad networks have filters to check for bots, as well as you can have filter on your site as well. Depending on what your size is and what you do - you might want to invest into that (could be quite expensive).

Popular at networks popular for a reason, and they do everything (a lot) what they can to prevent low quality traffic or bot, however setting ads is something that you have to do yourself. Low quality traffic can be just bad ad settings.

If Traffic Arbitrage wasn’t profitable it wouldn’t be that big these days and wouldn’t have so many sites doing that!

This point deserves an additional comment.

The three behemoths of online advertising at this moment are Google, Facebook and Amazon (who is well behind the first two).

• Google knows what you want (via search queries / Google Analytics)
• Amazon knows what products you're interested in and what you want to buy (based on product search / purchase history)
• Facebook knows who you are (based on likes / analytics from Facebook like buttons all over the Internet).

Broadly, Facebook has some of the most effective audience based targeting ever invented. But @se900se point still rings true.

I think both Google and Facebook knows very well who you are and what you want, given the Google/Facebook profile + their network of facebook plugins, GA plugins on most of the sites available.
Correct me if I'm wrong, imho, traffic arbitrage is simply:

• Buy good traffic that is highly targeted to your site's niche
• Utilize effectively the traffic: attractive content, smart navigation, and encourage the users to share the content
• Goal is to multiply the views and improve user retention on the site

Simply buying cheap from source X and sell directly to source Y wouldn't work.

@AdvertisingGuy can you share your tracking pixel stats?