Dayparting in split testing


You know that traffic doesn’t perform the same  every day of the week, which is why you should make your tests last several days.

But had you thought that one version may be better at certain times while another version better the rest of the time?

Look at this split test:

Makes you want to break down your tests by day and time and after getting the results, have a script serve the winning versions based on that. :P

Save time when watching videos


I’ve been going through tons of training videos the past months, but more so now that I’m helping Mike review new products.

It really is a very time consuming activity.

When going through books, it’s easy to skip parts and know what you skipped, or browse very fast and know what you went past fairly accurately. Not so with video.

Also, many presenters talk slowly, or talk a lot of fluff to make the video longer, consuming more of my time.

Talking with my friend Mark Jones last Friday, he mentioned a service from Enounce that allowed to play a video up to 2x its speed. I thought it was a great idea, but didn’t like having to pay for that.

Actually, I had read about this concept before, in a sci-fi book (The Invaders Plan) and liked it a lot, but that was like 13 years ago and I didn’t make the connection with my current needs.

Fortunately, I use a couple video players that are quite able and with a little searching, found how to adjust the playback speed in both. :)

One is KM Player, the other VLC Media Player.

KM Player is Windows only and I like it better. The way I found to adjust the speed is going to the Control Box (Alt+G), in the Audio tab (it’ll keep video synched).

VLC is available for Mac as well, and the speed control is available in the latest versions, so if you have an older one, update it. You just click the “1.00x” text next to the time, or use the arrows on each side of the timeline. Here’s a screenshot:

Video rebuffering will cost you 4 out of 5 people


Found this article that tells about a TubeMogul study where they found that 81.19% of people will leave if the video rebuffers even once.

Of course, a video doesn’t rebuffer for everyone. The study says that out of the services they tested, 6.81% of the streams rebuffered. 81.19% of that would be a bit over 5.55% of the total: the percentage of people that left the page because of a rebuffer.

Of course the conditions of that test may -most probably- be different from your own. I know I’ve been to the sales pages of different offers that had their video rebuffer, sometimes repeatedly, so I assume it’s likely that some marketers don’t worry about this point as much as they should.

I know that a sales page has more things for one to do in it that just watch a video, compared to a page that is there just for the video, but I’m sure that the rebuffering affects sales page visitors as well, especially those relying on video to sell.

It’d be interesting to monitor the rebuffering in sales pages, but the best bet will be to never have it do that.

Thought it’d be important to show you guys this, because it’s common sense that video should start fast and not be interrupted, but I’m sure most of us didn’t have an idea of how crucial it is.

To avoid the rebuffering, I’d say there are a few important things to make sure of:

  1. Your server hardware is fast. This is particularly important if it has to serve several visitors at the same time.
  2. Your server connection is fast. Also the closer the server is to the visitor, the better. A solution like Amazon’s CloudFront makes a lot of sense, especially for an international audience.
  3. Your video is very optimized. If the compression doesn’t reduce the file’s size enough, lower the quality or/and reduce the resolution.

Have everyone you can dummy run the sales process with people from different places to make sure they have no problem whatsoever, I’d suggest having friends from different locations load your page and tell you if they had rebuffering happen during video playback. Make sure you get some dialup and mobile users in the test too!

ClickHeat tracking


I forgot to mention ClickHeat, a heat map script that I found some time ago over at SourceForge.

I haven’t tried it, but it’s supposed to do heat maps (don’t know about confetti, though).

I particularly like that it’s free and self-hosted.

Got a copy to figure it out and it was easier than I expected!

  1. Get the newest file from SourceForge.
  2. Extract the archive and upload it on your server. There’s a clickheat directory in the archive, you can just upload it to the server root, so that you’ll find it at I haven’t tried using a diferent name for the directory, but it may work just the same, which would be desirable if you don’t want others to guess your directory for it.
  3. Go to that URL with your browser and follow the instructions. It’s mostly clicking the button to continue, only the the configurations page will need you to add the admin’s username and password.
  4. After that you’ll be shown the login page, where you use the username and password you just configured for it.
  5. You’ll be in the reports page, empty because it’s a new install, and there’ll be a link to get the JavaScript code for the template. Click it.
  6. Copy the code and paste it in your template (just like when you add tracking code from Google Analytics or similar).
  7. In the page where you got the JS, you will also see some settings. One of them talks about grouping. Use the page titles to track, it will cause ClickHeat to generate a drop down for each one of your pages. It says “not recommended” because it has a higher performance hit, but if needed, there’s a hack that deals with it.

That should do it. Let me know if you had any trouble. :)

CrazyEgg analytics


I just found another cool tool that gives data about the user’s activity on the page, but not really the same, so it complements what you had to make better decisions when creating those page variations in your testing batteries.

The page is CrazyEgg and they have heat maps and confetti maps (you’ll get why with the screenshots). These offer a visual summary of the clicks on the page.

Page heat map according to Google


Here’s a tips article from Google pointing out the best places for ads in a page.

Of course it’s particularly relevant for those of you that run ads on your pages, and you -probably- already know these, but I’m posting it for another reason.

It gives a good idea of where the visitor focuses on a page, or where not, which you can use to decide where you place your page elements when designing those split testing versions/variations of your pages.

Using trust seals on your page

A trust seal is a seal awarded by proprietary companies to businesses to display, in an attempt to boost customer confidence.

Does it make a big difference what seal one uses? I find there are several services to choose from.

This is the list I have compiled so far:

Which have I missed? Which would you recommend?

They have different prices, so if the result is pretty much the same in terms of trust in the client’s eyes, it’d be good to have the option of choosing a cheaper one when starting out with a limited budget.

Split testing the seals would be a good idea to see what impact it has on conversions. Maybe just using a seal created by you, as long as it looks pretty solid and official, will improve the conversion.

I found another trust site, although not exactly like the ones mentioned above. It’s not for every business, but seems great for those that can benefit from it.