[loxly]

NyFalcon and Loxly team up to explain A/B (and maybe even C) testing for affiliates. Once you listen, post follow up questions right here and we'll see what we can come up with.

We are keeping these short, 15 minutes max so you don't have to take hours to listen this time

Enjoy!

[ladidah]

Great job! I like the informal laissez-faire style.

I heard or read the same thing about the Adsense color but I disagree somewhat since I have tested my sites with zillions of colors and the white/black/blue does not work well for me. I think for general business sites yeah but not niche specific with site that don't have those general colors. (think feminine frilly)

Looking forward to hearing about the RSS adsense through Feedburner.... teasers :P

[John Powell]

I really enjoyed it and even went to hear the first one. That way I get to see the evolution. Hearing voices is a good thing especially that NY accent. This print mode misses that.

Check the link to "Download the MP3" as it seems broken.

[sfcom]

It sounds like you both had fun doing the podcast. Although I do no AdSense at all, I understood the concepts that Vinny was talking about and can see how they would improve CTR. Will you be covering other topics besides AdSense in the future or are you just taking it week by week? Either way, I will probably listen in.

-sfcom

[loxly]

Thanks! Fixed the links, sorry about that.

Yes we will be covering more than just adsense. Post topics and even specific questions you are interested in hearing about and that's what we'll cover We have some ideas of course, but if we know what's on people's minds, that's where we'll go first. He happens to know a lot about adsense, so some of the examples may start there

And yes, a/b testing is good because not the same colors work for every site So testing makes sense.

[Geno Prussakov]

Good podcast, Debbie and Vinny!

Every successful affiliate marketer pays a tremendous attention to web analytics. Every super-affiliate I have ever conversed with performs an ongoing, never-ceasing testing of their campaigns. After all, there is no way to know for sure unless you test.

I am just surprised to be finding out how many affiliates are not into testing. Popularizing the idea is exactly what is needed. And I applaud Deb and Vinny for this.

G.

[Vinny O'Hare]

Thanks Guys

Geno you would be amazed at the amount of affiliates that asked me what a/b testing was after we did the first podcast. That is why we touched on it in the second episode.

You Am's may want to add what it is in your next affiliate newsletter?

[Daniel M. Clark]

Okay, first of all, I haven't listened yet. I'm about to. I just downloaded it, and I'm going to put it on my iPod and listen while I get ready for bed. So, I'll leave feedback on the actual content tomorrow. I wanted to write this now, though, so I wouldn't forget.

Please, please, please, set your file attributes and meta data. The file name - makingmoney2.mp3 - is fine. Clearly, since the last one was makingmoney1, you're just numbering, and that's cool. What needs attention is the meta data, the ID3 tags, for those of us that download and put it into iTunes or whatever software we use.

iTunes lists the file thusly:

Name: test
Genre: blues
Artist: <none>
Year: <none>

At the very least, those four fields in the mp3's ID3 tags should be set.

[loxly]

The files get sent to iTunes via Podango and I haven't found where to change those settings Daniel. I will look into it tomorrow. Thank you for the heads up. I don't use iTunes at all, so am clueless there.

[loxly]

Personally I think that a lot of affiliates don't realize they are doing A/B testing And that is mentioned in the podcast. Of course affiliates are always looking at conversions and improving their pages, but they may not be paying attention to some factors they should be and they may not be testing Page A vs Page B to see what is converting better.

[chetf]

Pretty good bringing it up.

Definitely formalizing your a/b testing helps. I don't trust my feeling at looking at the numbers, I always "run them". Being strict about this helps you find things that are not logical or obvious. It also helps to do big sweeping tests, and then take the winner there and tweak test it to optimize the page within the constraints of the larger test.

While looking at how smart companies do it is a good idea, don't be scared to try things out of the box and compare those results, I have been surprised more than a few times by the results - especially when they were the results I didn't want to win...

I would guess the biggest problem with all of this is, most new affiliates do not have the traffic numbers to base their testing on and end up getting faulty results because their test pool is too small.

Chet

[Daniel M. Clark]

I'll IM you later about the MP3 tags issue, Deb, but basically, whatever program you use to listen to MP3s should be able to edit the tags.

The podcast sounded very good, just like the first one. It sounds like you're in the same room instead of talking over Skype. One suggestion - Vinny came through louder than you. Either dial him down your dial you up just a bit. Other than that, it sounded really, really good.

You both sound like you're getting used to the format. There were fewer gaps of silence (though there were a few near the top) and you're both good about the "um's" and "ah's".

Content-wise, I don't agree with the premise that pages are best when white/black/blue. Different sites have different user bases and different expectations from those users. More important than the colors is the contrast, I think. I don't think white is necessary, but light is. I don't think black is necessary, but dark is. Links, I think, don't need to be blue, but they need to be a unique color that stands well out from the color of the non-link text. On one of my sites, I used red instead of blue because my logo was red and I wanted to match it. The site looked good, and I don't believe it suffered from simply having red links instead of blue.

I really liked the point about only having one AdSense block on a page. I hate going to a site that puts the max of four on the page... it's always overkill. Plus, it doesn't earn as much for the site owner.

That's all I got right now. If I think of anything later, I'll post again

[loxly]

Originally Posted by HecticDMC

Content-wise, I don't agree with the premise that pages are best when white/black/blue. Different sites have different user bases and different expectations from those users. More important than the colors is the contrast, I think. I don't think white is necessary, but light is. I don't think black is necessary, but dark is. Links, I think, don't need to be blue, but they need to be a unique color that stands well out from the color of the non-link text. On one of my sites, I used red instead of blue because my logo was red and I wanted to match it. The site looked good, and I don't believe it suffered from simply having red links instead of blue.

That's why testing is good There are general rules that have been proven effective however there are always sites that break the rules. Test your sites, change the link colors and see what happens. You may be surprised, you may not.

At least we got people to start thinking And that is good for everyone.

[Emilio-ARC]

Lox,

Listened to the podcast and found it to be informative and light hearted. I love the laid back tone of the training and how gracious Vinny was to share many of the techniques he uses with the rest of the community.

Great job!

[e-Gazer]

Great podcast guys - I listened to it the other day but didn't have a chance to comment right then/there. I am sure there are many who will find it useful!

It is amazing how many are not yet familiar with the concept of A/B testing - but hey, everyone has to start somewhere.

Thanks to resources like these, we have lots of great opportunities for continuing education in this biz.

[Vinny O'Hare]

Loxly I know how to edit the stuff Hectic was talking about. I skipped over that part, didn't think it mattered.

Hectic I could agree that the page background doesn't have to be white, it could be a shade on the gray side. The link color I will have to disagree with you as I would never make the clickable link any other color then blue. When I see a site with links other then blue I click off the site because it doesn't look like a credible site. Could be a user preference thing on my behalf, but if I am doing it odds are others are also.

[chetf]

Originally Posted by nyfalcon

Loxly I know how to edit the stuff Hectic was talking about. I skipped over that part, didn't think it mattered.

Hectic I could agree that the page background doesn't have to be white, it could be a shade on the gray side. The link color I will have to disagree with you as I would never make the clickable link any other color then blue. When I see a site with links other then blue I click off the site because it doesn't look like a credible site. Could be a user preference thing on my behalf, but if I am doing it odds are others are also.

I think you just invalidated your entire podcast.

[loxly]

Originally Posted by chetf

I think you just invalidated your entire podcast.

Not really He talks about testing, and his testing results and the results of others. There is a lot of data to back up his statements about text, link and background colors. That is only one element of testing. There are lots of other page elements that produce, or don't produce, results.

[Lanadili]

I listened to the podcast yesterday and really liked it. I'm looking forward to the next one on RSS feeds

[Vinny O'Hare]

Chet

Thanks for your opinion. Before I say something I will run it by my validation team first

I am also a numbers freak, having a vast number of pages and sites it is easy for me to get a good grasp on what works and more importantly what doesn't work.

I agree most affiliates don't have that type of data to pull from, you would be surprised at the amount of email from affiliates asking me what a/b testing even was.