Euphoria turning to disappointment

I've now tried Amazon's Simple Queue Service, and it doesn't work like I expected.  Once I queue stuff up, and try to retrieve it, it doesn't all come back.  It comes back in random chunks of data.  I never know how many of my queued items will be returned, or what order they'll come back in.  This despite running the proper queries to retrieve everything.  I need to run the "return all" query a minimum of 5 times before I get all 10 items in the queue back to me.  I posted to the (empty) developer forums for the simple queue service 6 hours ago, and have received no reply.  I'm not sure if the lights are on over there because the forum is totally empty...

So then I tried to make REST requests to the Alexa Web Information Services (AWIS) web service.  Despite using Amazon's own signing test page, I only receive a very non-descriptive "not_authorized_11456" error.  That sucks!  Browsing through their sparse forums, I've found a couple of people looking into the same issue.  I even saw one post of somebody from Amazon that said "we're looking into it" and then never responded back saying it was fixed, or that it was a bigger issue.

Same non-descriptive error as above for the REST request to the Alexa Web Search Platform (Beta).

So my question to Amazon is.  WHY SHOULD I USE YOUR WEB SERVICES IN MY APPS?  I can't get them to work.  The documentation is incomplete, the forums are empty, the support avenues are limited, the response is non-existent.  

The potential is great!  The ideas are awesome!  Oh the cool stuff I could build - if only it worked!  aurgghhhh!  That's the answer - that's why I'll use them.  I'll continue to be frustrated and wait for Amazon to get it together because the promise of these web services are too great to ignore.

In the mean time, don't ask me anything - I might bite your head off...

Goodbye big hosting costs!

Amazon just came out with S3 - Simple Storage Service.  This just gives me goose bumps all over...

 

I'm just high on Amazon as a whole right now... I plan to post more at length about all the cool things they're doing right now.

 

S3 is a hosted hard drive space - the same storage space that they use for all of their own sites.  It's accessible only through REST or SOAP (for downloads you can call a file through http).  It costs a paltry, miniscule, (ridiculous) $.15/gig/month and $.20/gig of transfer.  Amazing!  Just think of all the video, audio, and images you could put on your site.

 

The days of paying big costs to web hosts for disk space and bandwidth are over.  I almost feel sorry for the dedicated server salespeople of the world...  I do work for a site that figures they could save around $120,000/year of co-location and server costs if they used this.

 

Who are the potential customers?

1. Porn.  Talk about lots of bandwidth...


2. Video/audio.  I wonder how much You Tube pays for their disk space and transfer?


3. Image hosting - same sentiment as above, but for flickr.


4. Outsourced hard drive space - cool idea check out Box.net and X Drive


5. The fifth big customer?  Hopefully, me!  Just don't know what I'm going to put there yet...

Check it out here

State of the Spam Site

As long as there is money to be made from putting content on the Internet there will be blackhat spammers. Blackhat spammers of the past are the ones that tossed up that site of jibberish, full of ads. They're the scourge of Google, hated by Matt and Jeremy. Of course, I've never done this... haven't even tried... ok, there was that once, but

A couple of the more notorious black hats these days are Dave Naylor (Dave N) and SEO Blackhat. These guys make a living on understanding search engine algorithms and manipulating them to gain traffic across their sites. They do not value a domain, or web host as much as most webmasters because they're always trying new things, getting booted, and moving on to the next. When they hit - they hit big, make a few bucks for a few months, get banned, and move on.

Blackhat spammers are the reason that it takes most sites a long time to get decent rankings with their sites these days. Recent algorithm changes in Google, and to a lesser extent, Yahoo, have made the tenure of a site on the web an important factor. The longer that your site has been around, without being flagged as spam, the more likely that the new content you put up will rank well.

Because of this fact, a growing trend is to purchase existing sites, and site valuation is becoming more of a science than the art that it was in the past. Black hats are the first to catch onto this trend, and snapping up old sites because they're at the forefront of the algorithms. SEO blackhat put up a very interesting post about his current strategy of buying Google's trust.

If you could come up with hard and fast rules about how much to pay for a site based upon longevity, alexa rank, pr, and size, you could have a very nice little long term business. I firmly believe that domains are the new real estate, and the internet is a solid investment vehicle that's currently under-valued. Like all good investment strategies this will only last as long as it's relatively under-the-radar. Once it gets on the average Joe webmaster's screen, the bubble inflates and pops. Or at least, evens out.

Watch the black hats - much like the porn industry, they're always at the edge. Bleeding first, so the rest of us can learn what works, and what doesn't.

TG

How to measure if people scroll down a sales letter

I wanted to be able to see how many people were scrolling down my sales letters, and presumably reading the text, so I came up with a cool use of javascript.

 

I found a piece of javascript code that tracks the mouse movement.  When it gets above a certain value on the y-axis, I submit a page in a hidden frame, and insert a database record to log that this person had scrolled down the page.

 

Here's the code for the hidden frame, inserted just above the closing </body> tag.

 

<iframe name="LogFrame" src="" width="0" height= "0"></iframe>

 

Here's the javascript code that tracks the mouse:

 

// THIS PIECE OF CODE TRACKS THE MOUSE, AND FIRES CODE TO DETERMINE WHETHER THE USER HAS SCROLLED DOWN THE PAGE OR NOT
// Detect if the browser is IE or not.
// If it is not IE, we assume that the browser is NS.
var IE = document.all?true:false

// If NS -- that is, !IE -- then set up for mouse capture
if (!IE) document.captureEvents(Event.MOUSEMOVE)

//Set-up to use getMouseXY function onMouseMove
document.onmousemove = getMouseXY;

// Temporary variables to hold mouse x-y pos.s
var tempX = 0
var tempY = 0
var LogIt = 0

// Main function to retrieve mouse x-y pos.s

function getMouseXY(e) {
if (IE) { // grab the x-y pos.s if browser is IE
 tempX = event.clientX + document.body.scrollLeft
 tempY = event.clientY + document.body.scrollTop
} else {  // grab the x-y pos.s if browser is NS
 tempX = e.pageX
 tempY = e.pageY
} 

if (LogIt == 0 && tempY > 1500)
{  //this is where the frame is submitted with the page that will insert the record into the DB

 window.LogFrame.location.href = "ebook/LogScrollDown.aspx";
 LogIt = 1
}
 
return true
}

This method of scroll-tracking is far from perfect.  If the person is using a non-supported browser, or if cookies are turned off, it won't work.  If the person uses only their down arrows to scroll and never move the mouse, it won't work either.  But, it will work for around 95% of people, and that won't change from sales letter mod-to-mod, so it's a good test.

This technique has been very informative for me so far...

--

>> There are only 10 kinds of people in the world - those that know binary and those that don't.

Tracking sales

I just added a "sales letter version" key to one of my sales pages.  I wanted to track the conversion percentages for each modification I made to the sales letter, now matter how miniscule.  I was tracking conversion percentages by day, but now I'm also tracking by "version."

 

Here's the breakdown:

 

Sales Letter Version (the version number of the text that's currently active)

Version Date (date the current version went into effect)

Splash Count (number of people that saw the sales letter)

Scroll Count (number of people that scrolled down, and presumably read, the sales letter)

Scroll % (scroll count / splash count)

Checkout Page Count (number of people that saw the checkout page)

Checkout Page % (checkout count / splash count)

Purchase Count (number of people that purchased)

Purchase % (purchase count / splash count)

 

With this quick overview, I can run a new headline for a day or two and see if it's more or less effective than the previous headline.  If I make changes frequently to my sales copy (which I do, daily), this method isn't very scientific.  First of all, it doesn't gather a large sample size if you change it too frequently, secondly, it doesn't take into account things like seasonal or day-of-week shopping patterns (like Wednesday sales traditionally greater than Monday sales).  The better way to do this would have been to run an A-B test for a longer period of time, sending some customers to each version of the sales letter and measuring them against each other.  But... this will do for now!

--

>> There are only 10 kinds of people in the world - those that know binary and those that don't.