Another Wordpress Blog

Alexa is widely used: Pay Per Post, Text Link Ads and many other service use this to estimate traffic. Today, I’ll explain the common rank gaming method — which I will call “The Alexa Project”, that is known to work. I’ll also describe why Alexa Projects generally work only for a short time and then describe what’s required to make them work forever .

The Alexa Gaming Method

A well known method to trick Alexa into believing you have high traffic relies on four things:

1. A group of friends who agree to work together. This group might call themselves the “Alexa Project”.
2. All friends installing both Alexa and “Linky” extensions to their browsers. (The Alexa toolbars are available FireFox and Internet Explorer. Linky is available for Firefox.)
3. At least one friend sets up a web page to act as a “hub”. This web page includes links to every blog in the Alexa Project.
4. Using their Alexa browser, all friends in the Alexa Project agree to visit the hub regularly and automatically open every link on the page, either manually, or automatically using their Linky Tool bar.

The behavior of an individual participate is illustrate to the right. Basically, the visit the “hub page”, and click open every link. In principle, they have visited their friends site, and so, in some sense, their friend deserves to have Alexa give credit for the visit.

So… it’s not really an unfair right?

Yes, the method is, in fact, unfair. After all, what everyone who joins the project knows is that somewhere between 1 in 10,000 to 1 in 100 people have Alexa toolbars installed. (I tend to think the value is roughly 1 in 500, and I’ll use that number from now on.)

So, if a blogger can get 10 friends a day to visit using Alexa toolbars, Alexa credits the blog with equivalent of 5,000 visits by “random” people. By banding together, a group of 10 people who visit each other blogs every day can seriously drive down their Alexa ranks. (With Alexa, #1 is the best rank. )

Does this really work?

Yes. This drives down your Alexa rank.

In fact, this method relies on Alexa measuring traffic exactly the way Alexa tells people it measures traffic. If someone visits using an Alexa browser, Alexa counts it. Otherwise, Alexa doesn’t count that traffic.

In case you are wondering why you can’t just reload your blog over and over and over. Well, Alexa only counts any individual IP once. You can give yourself one “Alexa hit” a day. After that, you need real visitors with their own Alexa bars installed.

What Goes Wrong?

Frailty, thy name is “Alexa Project Participant”. Over time, each participant begins to neglect their job. (And it is a job — unless you would have visited the blog itself to read that blog anyway. But you wouldn’t; otherwise, no one would need “The Alexa Project”. )

Anyway, maybe the unreliable participant visits on Sunday, but forgets to visit on Monday. Then, they forget again. A few more forget to visit. Eventually, everyone begins to see their Alexa ranks degrade.

At that point, the more reliable members get discouraged: They know they are giving their friends a boost, but their friends aren’t returning the favor!

Soon, everyone stops. And everyone’s Alexa rank starts to rise up again.

Sometimes people start to regroup and try to convince others to hold up their end of the bargain. That can work for a while, then the whole cycle begins again.

How could the Alexa juicing be made to last forever?

Why, by writing plugin!

What would the plugin need to do:

1. Permit users to enter their friends blogs into a database.
2. Create a button that lets users auto-visit their blogs from their own blogs. The p
3. Include a script that detects their friends visits. (This can be done by reading referrers and logging the ones that match their friends blogs.
4. Once underway, the blogger using “You Visit, I Visit” would click a button and only visit people who actually visited them! (And if the other bloggers has installed the plugin, well, they know they’ll get a return visit. )
5. And everything should be fairly invisible to outside observers, because, well… There are people who claim this is “just visiting their friends blogs” and “just getting credit for traffic”, but there are others who bet to differ.

As my regular readers know, I am currently running public tests to see if rumored Alexa gaming work. Because Pay Per Post has discussed the issues surrounding my experiment twice, I want to also discuss my current view of Pay Per Post.

The short story is that, having given it some thought, I realized that I had to sever any relationship with paid-to-post programs if I wish to run the tests to gaming strategies. By my own request, I ceased to be a postie on Monday. But, that’s actually the small story, as it happens to have to do with me, personally.

The larger story has to do with Pay Per Post reaction: When they became aware of the experiment, Karen, of Customer Love, announced that Posties would not be permitted to participate in even short term tests to see if Alexa can be gamed. Later, “froogle” announced that PPP will be auto-detecting attempts to game Alexa using these redirects.

Karen and Froogle at Pay Per post are correct to announce this policy and check for compliance. (Checking is easy. I set this up so that participants cannot conceal participation.)

One reason given for PPP’s action appears to be this: PPP currently uses Alexa for segmentation. Therefor PPP cannot be permitting Posties to be gaming Alexa to improve the price they charge advertisers.

As it happens, though I didn’t really think of this before launching the experiment, I recognize PPP’s policy as manifestly correct.

I suspect people will wonder why I didn’t think of this in the first place. Part of the reason is that, initially, I was focused only on testing the rumored gaming method. I’m afraid, like the famous cat, I will someday be killed by my own curiosity.

Also, as it happens, I hadn’t taken a PPP opp in over 6 weeks or so and didn’t intend to take any if I actually managed to game my Alexa. So the thought “Oh, I’ll be able to take better opps” didn’t really hit me. (If it had, I would either have done the test secretly or resigned from PPP immediately. )

I also believe Alexa is inherently deficient due to the self-selection bias introduced by the toolbar and I believe it is already heavily gamed. So, I sincerely don’t believe a few blogs plublicly testing method makes much of a difference in overall accuracy or fairness of Alexa ranks.

Of course, the other reason I might have not thought of it is simply, ordinary stupidity and thoughtlessness.

Nevertheless, the fact is, PPP is correct: regardless of Alexa’s faults, because PPP uses Alexa for segmentation, Posties cannot be conducting of Alexa gaming on their blogs! (At minimum, the tests must be designed in ways that do not affect a PPP blogs’ Alexa rank in even the slightest. But that is difficult to know in advance.)

But I want to conduct tests and do so publicly. So here’s what I did.

Recognizing Pay Per Posts position is correct, I did two things rather quickly:
1) I yanked the redirect images from blog footers. Because of the experiments design, this immediately ended the first experiment which persisted only a few hours and so can have very little effect on anyone’s rank.¹

2) I voluntarily resigned from Pay Per Post.² I had to ask what the process was, and did a bit of fiddling at the PPP interface, but unfortunately, the process isn’t obvious. I sent pm’s to four PPP staff asking the procedure and also filled out a ticket. Within two hours of filling out the ticket, I was out of the system. I thanked Karen of Customer Love. (For those wondering about any financial consequences to me: there are none. I hadn’t taken a PPP opp in at least 6 weeks. )

3) I redesigned the experiment in a way that will actually give clearer results without the need of many blogs.¹

So, now I am free to conduct the current experiments and, what’s more important to me, I am free to conduct future experiments without violating PPP policies or exposing PPP or posties to any unfavorable publicity.

I also think Posties taking opps from Pay Per Post should applaud PPP for their quick response on this.

---

1. In fact, for those wondering, preliminary results suggest the short term test had absolutely zero effect. Because I was in the process of resigning, I kept images on my own blogs. The preliminary data suggest that this method does not work. LordMatt’s sudden increase in Alexa when he asked a few people to include images at his blog was probably due to a real increase in blog visitors with Alexa toolbars. Still, I need to continue 3-4 days to get incontrovertible evidence which I think is useful because rumored methods are constantly popping up.

2. I also yanked my blogs from ReviewMe and Sponsored Reviews. It is suprisignly difficult to automatically delete a blog from many of these services!