SEO Junk

Webmaster resources and topics from the best threads

Let’s talk about traffic washing and dilution

This is a long post, but you will like it. So, get a cup of coffee and some popcorn. For some, this will change the way you see the well known traffic brokering business model. You will learn how traffic brokers wash traffic and learn how to wash their already washed traffic against them and make money out of nowhere.

Many complain their ratios have gone down the drain the last 9 months. One reason for these bad ratios are traffic brokers doing traffic washing. This dilutes the real traffic value. With more traffic being washed from the usual traffic circuits, the traffic you get is getting less value each day.

I decided to do scientific research on how much does brokered traffic get diluted and washed. In the end, the buyer (webmaster) is scammed by buying traffic which has much lower value.

1. The traffic broker
During the last few years, it seems I am the only traffic trading script supporter who has never sold his script payload to traffic brokers. The main reason is traffic brokers always tend to wash a good percentage of their traffic. The other reason is I know how to convert my traffic into money.

During these last few months, several brokers have approached me about selling my traffic. This included a couple of non-brokering companies, so it seemed something fishy was going on. Then, people suddenly started reporting their sales ratios getting 10 times worse while my sales ratios remained exactly the same. That raised alarms to me.

2. The setup
This test took 4 months and almost 15 days more more to analyze the data. The total amount of money spent for the test is near $7000 + 2 dedicated servers during 4 months.

Since I wanted to compare apples to apples, it meant creating the same conditions for the traffic. There were two identical sites, running the same scripts, having the same layout, and the same server and script configurations. The servers had exactly the same hardware and system configurations.

One of the servers would be used with brokered traffic, exclusively. The other would be used with my traffic as control traffic. This let me ensure both systems ranthe same traffic numbers during the same hours, so chances for clicks and sales should be exactly the same.

The control traffic was the payout traffic coming completely untargeted from multiple sites that running CjOverkill. The other traffic would be brokered traffic. It would also be untargeted, for multiple niches (nichemix).

3. The setup
Ok, two servers are ready. One was set as a control server and the other was set as a study server.
The first thing to do was test traffic brokering till I found the right traffic to match the same productivity as my own traffic.

For most traffic brokers, it was the “niche mix” traffic. Fortunately, it was very cheap traffic. Some traffic brokers let you buy traffic from other trade scripts. However, in practically all the cases, the results were worse than my control traffic.
Once traffic performance was matched, the whole system was ready for real life tests.

4. A TGP/MGP test
Neither TGP traded traffic with other sites. Both sites had the same galleries listed and exactly the same layout. The traffic total was 170k, distributed 1k per hour, during one full week period

The control site, running my traffic completely untargeted, had 120% page views/uniques and made 4 sales. The study site running the traffic broker traffic did about 125% page views/uniques and made no sales at all.

5. A paysite test
The paysite test of choice was an AFF landing page looking very much like a TGP. This ensured surfers would not close it immediately. This test used 90k brokering traffic vs 75k of my control traffic.
The results were more than bizarre:

Tracking ID | Clicks | Uniques | Total Signups | Uni/Signup | Women Signups # | Orders | $ Orders | Adjusted order Users /Returns / Credits
control | 100718 | 76848 | 63 | 1219.8 | 5 | 0 | $0.00 0
broker | 88570 | 87783 | 35 | 2508.1 | 1 | 0 | $0.00 0

As you can see, my traffic had 63 free signups while the broker traffic had only 35, while my traffic had 10k uniques less.

Since it’s a dating site, I decided to give these free signups a month to see how much they would convert into paid signups.

The traffic broker traffic made no sales (zero, nothing, null, nada). My control traffic generated a total of $648.89 after a month since the test finish. At this point, the red lights were flashing all over the place. So, I decided to further investigate the broker traffic.

6. Traffic GeoIP Analysis
The GeoIP country stats showed most traffic brokers sent a huge percentage of traffic from non USA/EU countries. Remarkably, one traffic broker had more Russin Federation traffic, while others had lots of South American traffic.

The countrol traffic showed a very balanced country mix. That raised alarms and made me further investigate the traffic brokers country mix. Some of the traffic brokers were shaving the USA/EU countries during some hours when their traffic was practically 40% Russian or South American. During the rest of the hours it was still a high percentage, but not that high.

In any case, the USA/EU counries traffic was much lower than it was for my control traffic. Between 15% and 30% less, sometimes even more. So, it was clear the traffic brokers were redirecting most USA/EU traffic and selling the crap countries traffic they didn’t want.

7. Traffic hitbot analysis
Now knowing the so-called “reputable” traffic brokers were shaving the USA/EU traffic, I just wanted to test if they also spread hitbots. So far, practically all the traffic brokers were spreading bots. Some did more, others less. Between 15% and 65% was all bot traffic. Note that 65% is alot of bot traffic.

Some bots were so rudimentary, they were more than easy to detect on runtime. Others required more difficult tests to expose, but still doable and possible.

Some of the different types of bots I detected:
1 ) Bots clicking several times per second. Even the fastest clicking human cannot make 5 clicks in one second.
2 ) Bots loading and clicking on hidden images. Some of these were hidden under a HTML layout, so even knowing where the 1×1 pixel image is, you cannot click on it because it’s under another layout.
3 ) Bots with a known screen saver autosurfing application signature. You know all these screensavers that surf while you are away?
4 ) Bots running IE API pretending to be FireFox browser.
5 ) Bots accepting conary cookies and then passing them to the site again. A conary cookie is a cookie a real browser would not accept, while a bot would be unable to distinguish whether it’s real or not.
6 ) Bots once marked with a UniqueID which changes with each load or click would load or click the site again with the old ID instead of accepting the new one.

It seems traffic brokers either don’t have the needed tools to filter their traffic, or they just sell the garbage while keeping the good traffic for themselves. No wonder your traffic is converting worse if you buy traffic from brokers or trade with sites buying from brokers. About 50% to 80% of that traffic is either hitbot or has no credit card at all (making a sum between bots and crap countries).

What can you do to benefit from this situation?
Wash the traffic again. This worked well for me so far and it was really funny doing it.

1 ) Create traffic seller/buyer accounts with these same brokers.
2 ) Buy traffic from Broker A and sell it to Broker B.
3 ) Filter the USA/EU traffic in the process and send it it where traffic converts.

It’s a bit more complex and is beyond the scope of this article. The end result is you recover the money spent on non-USA/EU traffic you bought from broker A by selling that same traffic back to traffic broker B,. In the end, you are paying $2 for 1k of USA/EU traffic with about a 20% hitbot factor. Depending on the broker A and broker B prices/payouts, the traffic may be free for you.

The results in this post are from the worst traffic broker of the test. Yet none of the others was even close to the control traffic.

If you enjoyed this post, make sure you subscribe to my RSS feed!

Subscribe to SEO Junk by Email

If you enjoyed this post, please consider to leave a comment or subscribe to the feed and get future articles delivered to your feed reader.

Powered by Wordpress | WP Premium theme by WP Remix
Copyright 2007. SEO Junk. All rights reserved

• Home
• About