Alex Schultz: APIs

Keyword Spamming YouTube

Google is being exploited by spammers on Google video and YouTube. When they start paying for content it will only get worse! Jeremy wrote a little rebuttle of a MyBlogLog spammer, cornwall seo talks about it too and browsing google video directly after that I started seeing the same there.

I am a big fan of NCIS... I can't help it I like the humour coupled with the overwhelming sense of coolness and fairly decent plots. Very quickly since google video integrated YouTube listings into it's search results (great quick exploitation of synergy there well done G men + women) I realised that all my fave shows were now being offered up as bootleg copies:

This in itself isn't the end of the world since it costs $1.99 to get the CBS version and I can keep it on my machine easily and (thankfully) don't have to get up every 10min to move onto the next clip but it led me to test out a theory on another nasty spam technique I used to use in my early days of SEO (which timed with Napster). Keyword stuffing...

On this page you can see a search for NCIS turns up Gilmore Girls and South Park (page 2 of results). You can also see that the descriptions are keyword stuffed with "ER", "CSI" and various other popular TV shows.

The joy of using back links to help determine the keywords a web page is valuable for in Natural Search means you can devalue the keywords on the page (to some extent) or at least get a hint to pick the most relevant ones. Combine this with the fact that (excepting cloaking) the page has to be user readable it's not too hard to see why keyword stuffing doesn't work for webpages any more. I however rarely read the descriptions of the YouTube and G Video's I watch and so spamming the description isn't going to hurt me. Combine that with the fact all Google has to go on is title, description, #times viewed and a little bit about the uploader I think we are coming to an interesting challenge for Google to deal with, similar to Jeremy's mybloglog problem and eBay's keyword stuffing problem in search results.

Good luck Google... get your brains on it (and get your users tagging that content ;) )!

YouTube vs Revver the fight is on

I am very excited about this recent post on bbc news stating YouTube is intending to start sharing it's revenue with it's users!

Since revver is now making a good 10-20% of the revenue I get on my paper airplanes website. I started off using YouTube for the videos on my site and now have c. 300k views on my YouTube videos but just a 3 short months later shifted to revver since they paid me for my videos and I have now earned just shy of $1000 from them and been paid $700 of that through paypal (quality service doubt youtube will do that :( ).

The questions I have outstanding are:

1. Will YouTube allow you to earn money from embedded videos or just on YouTube?
2. How will the ads work, are they inline or are they the ones surrounding the videos?
3. How can YouTube pay kids?
4. Did the you tube founder just tread on his tongue by saying this in the quiet period pre earnings :s ?

Overall this is a great shift and one I have been hoping for for a while. It opens up a whole new territory on the web for people like me who create how to sites and have skills to share (as well as kids who set themselves on fire and steal copyrighted material ;) ) and with a big power in advertising behind it (Google) I am totally stoked.

eBay Related Searches through Get Search Results

Ok I am simply blown away by the latest offering from the eBay API as an awesome way to make a bunch of money. eBay is now revealing (without the data + formula behind it of course) the results of their related searches algorithm. When you search on eBay for "tmx" on eBay at the moment it will give you 937 items but just above that it will give you the following related searches: ps3,   wii,   tmx elmo,   tickle me elmo,   playstation 3

This is really useful information you can use to determine (as a paid search advertiser for example) hey I don't want to buy keyword "tmx" it's too pricey but keyword "tickle me elmo" is closely related... I will buy that instead and send the clicks to the same page as TMX. This data is quite simply awesome and exposed world wide for free. Make use of it and you'll do great, esp. as an ebay affiliate and you can access it via the REST API, I am playing with this and a little bit of MAA right now :)

Skype4Com - Create Chat with VBS

I have spent a lot of the last week hacking about with the Skype api and vbs. I have wscript on my computer and wanted to open a skype chat window and send a message to the user based on the content of an API call response. Here is the code I used with the Skype4Com dll to do the skype call adapted from some JS API call code already available online. As always this is a true hack and so may not be 100% stable or great code it is offered as is for you to play with.

 

' // Create a Skype4COM object:
Set oSkype = WScript.CreateObject("Skype4COM.Skype")

' // Create a UserCollection object and add the multichat users to this collection:
Set oMembers = WScript.CreateObject("Skype4COM.UserCollection")
oMembers.Add(oSkype.User("alexapitest"))

' // Create a multichat:
Set oChat = oSkype.CreateChatMultiple(oMembers)

' // Open a chat window:
oChat.OpenWindow()

' // Send a chat message:
Set oMsg = oChat.SendMessage("Greetings!")

Banker's Rounding - WTF

So today I was hacking around with the Skype API to put together a little bot which will contact me once an hour while I am at work with some interesting data from the web. The idea is it will eventually be a bot that uses the eBay API to tell me when a certain item I am looking for is listed as a BIN under a certain price. I've been looking at various ways to do this but (since I only have one computer here and don't fancy going to the hassle of partitioning the hard drive) in the end I chose to use window's scheduled tasks and VBS (although perl with active perl gave it a run for it's money).

Due to this I am interested when random facts about VBS and indeed visual basic turn up but this one truly made me fall over backwards. WHO by default would set up a programming language to use banker's rounding WHY would I want 4.985 to round to 4.98 rather than 4.99 when I go to 3 significant figures!

Ah well thanks to the comments section of daily wtf for clearing that one up :)

SEO table stakes, my top 3 - Porthos.com

So I spent a little while chatting to the awesome folks (Hal and Ryan) over at porthos.com over the weekend. They have been making an effort to improve the SEO on their website and after a chat we had back in december their results have improved considerably. This chat really set me thinking alongside a few articles by Rand over at SEOmoz.org which talked about the basics of SEO and how having been involved in SEO for almost a decade many people who know about SEO consider the basics to be universally known and understood. That of course isn't true! So I thought I would post my 3 SEO basics here and share the knowledge:

1. Think of the keyword first: there are three components. This holds true for every small and large web retailer I have worked with from trulia.com to arenaflowers.co.uk
• what is the value of traffic from a keyword for you? (variable a)
• how much competition is there on that keyword? (variable b)
• how many searches are there for that keyword?  (variable c)
• (a*c)/b = relative focus you should be putting into ranking for that keyword
2. Content is king: You hear this a lot in the market these days esp. with social media but I don't neccessarily mean huge volumes of user generated content (although this has worked for me in cocktails). I mean make good quality content for search engines and users.  Why and how?
• search engines are looking for good quality content and striving for that end so my philosophy is don't keep trying to be one step ahead with spamming them
• users need to like your site and convert when they land on your site
• good quality content = good coding = usability. Use <hX></hX> tags, use Alt text, use meta tags, structure your site right for partially sited accessibility AND therefore search engine friendly code.
3. Get your linking strategy right: I see this as having three components
• lots of high quality, relevant inbound links with your target keyword (point 1) in them from other sites
• a flat internal linking structure within your site so that search engines can reach all your pages within about 2 clicks of any page
• get your keywords in your url structure. We all know it helps with search engine ranking and user click throughs on your search results, so why have horrible urls?

It was esp. interesting to me going through these points with Hal and Ryan because in return they treated me to some fun wine tasting and taught me a little bit more about wine. 2 simple points came out: see how a young (2004) cabernet tastes sweeter than a more mature wine (2001) and the bitterness of tanins from a mature wine made with unripe grapes is too strong to go with steak or a mild meal but tastes awesome with a strong cheese. Now I have heard them I get it and will use them a little bit in future at the summits and dinners I hold in my work. Perhaps whenever someone talks to you about SEO you will remember that just maybe the stuff we take for granted as universally held fact just isn't!

Awesome spoofs of the I'm a PC and I'm a mac adverts

So I hate how pretentious apple is, I love a lot of what they do and think they are great innovators. I wish that windows already ran as many cool little apps as Mac have and I would adore something like dashboard being integrated by default into all windows systems already... not after vista however despite this Macs are still stuck up. Which makes this esp. funny, enjoy:

http://tv.truenuff.com/mac/

Oh yes and I am very geekily enjoying linux again (for the first time since my Physics degree 2yrs ago) so I love the "I'm gonna block access to your gaming servers, dorks.... I AM ROOT!!!" joke in this video of the set: Gaming AWESOME!!!

Running a PHP script with a Cron job in 1and1

Following on from running my first cron job with 1and1 hosting what I really wanted to do was to be able to run a php script, perl script or some other form of script on a timed basis. This article is a description of setting up a cron job on 1and1 hosting to run a simple "hello world" style example. I have the 1&1 linux business hosting package so can only confirm this example works with that.

First I set up my helloworld.php script, this was a simple script to create a file called helloworld.txt and write the words "hello world" to that file and save it.

<?php

// the filename
$filename = 'helloworld.txt';

// the content
$content = 'hello world';

// create the file
$f=fopen($filename, "wb");
fputs($f, $content);
fclose($f);

// now done

?>

I tested this file by running it on my server and it worked fine. I would recommend you do the same.

The next thing I had to do was log on to my 1and1 server using SSH, the program I use for this is PuTTY which I have been using since university and works well. You enter the username and password for FTP and SSH access to your site in the PuTTY command prompt and then are logged into the site. The next two steps you have to go through are:

1. to find the location of your current directory: enter "pwd" to find this... mine was: "/kunden/homepages/XX/dXXXXXXXX/htdocs"
2. to find the location of php on your machine to do this I entered the command "php -info" in the command prompt and looked for the location in the output which was: "/usr/local/bin/php" however due to the nature of 1and1 hosting the full path was actually "/kunden/usr/local/bin/php"

Now I had to create my crontab file. In the file below the 14 refers to the 14minutes past the hour, you can find out about the 4 *'s from the crontab page on wikipedia. So I entered the command "crontab -e" (which means edit crontab) and came to the VI editor which 1and1 use on their debian linux installation. In that line I entered the command: "14 * * * * /kunden/usr/local/bin/php /kunden/homepages/XX/dXXXXXXXX/htdocs/helloworld.php > /dev/null" all on one line. This command means at 14minutes past every hour on every day using the program found at /kunden/usr/local/bin/php execute the file found at /kunden/homepages/XX/dXXXXXXXX/htdocs/helloworld.php and send the output to the location /dev/null (which means get rid of the output). Using this command I found that on 1and1 hosting my helloworld.txt file was updated between 14 and 18minutes past every hour.

I know this description would have helped me had it been available when I was looking for information on running a php file through a cron job on 1and1. I hope it helps you.

My First Cron Job with 1and1 hosting

A cron job is a way in Linux of getting a command to run on a schedule of your choosing. That command can be a script or a basic Linux command string such as copying a file. For me being able to execute cron jobs on my server was a key gap in my arsenal of web development skills that has led to me building all kinds of strange hacks to run timed jobs.

This weekend I set myself 3 tasks: 1. start my green card application, 2. find somewhere to start taking GMAT lessons and 3. learn how to do a cron job. In order to start this I needed a cheap host which provides linux hosting with cron jobs and the obvious choice seemed to be 1and1 hosting. I have spent much of today reading how terrible 1and1 hosting is and beating myself around the head with trying to get a cron job working on 1and1 shared hosting. Here is my first very simple cron job with 1and1 shared hosting:

Firstly you must SSH into your 1and1 hosting, I like to use PuTTY to do this.

You then have to log in to the SSH once connected using your FTP username and password from 1and1. My first cronjob was to create a copy of a file inside the directory structure, I uploaded the file helloworld.php to my directory using FTP and then copied the file helloworld.php to the file hello.php using the command "cp helloworld.php hello.php" in my SSH connection. To check this had worked I then ran the command "ls -l hello.php" and found that the file had just been copied by looking at the date and time of last update.

Having confirmed I could run these commands myself I wanted to create a cron job to execute the command. To edit my cron job I needed to create a crontab. The command to open and edit a crontab is "crontab -e" and I entered this in my SSH connection and saw the following screen:

This is the vi editor which is the editor 1and1 uses on debian linux. I personally was trained on emacs and so it was a bit tricky getting used to this editor. The first thing to do is hold down the shift key and hit "I" on your keyboard which accesses the INSERT mode on vi editor. Then you need to enter your Cron Job code. Mine looked like this:

45 * * * * cp helloworld.php hello2.php

The "45" stands for 45 minutes past the hour and the following *'s mean first hour (0-23) then day of month (0-31) then month of year (0-12) and finally day of week (0-7) (but I decided not to specify any of those).

The final thing to do is press "esc" and then hold down the shift key and press "Z" twice which saves your job and exits the vi editor. You should then get a message which says "crontab: installing new crontab" and if nothing goes wrong that's all you will see. The cron job will then run at 45minutes past the next hour assuming everything works smoothly.

If you want to see how close to that time you are you can always type "date" into the command prompt and hit return.

Fingers crossed I will be able to do something cleverer with my crontab soon involving scripts and when I can I will share that here because I am suffering extreme frustration finding information on and working out how to do anything with scripts and crontabs on 1and1.

2007 Resolutions - cos predictions get me in trouble

Last year I put together some predictions when I first started my blog which I have removed. They turned out to be a bit of a liability at work, so this year I am not going to make that mistake again.

I do think two (non-controversial) things will be true in 2007:

1. Widgets/Gadgets will be a big revenue generator
2. Mobile use of the internet will soar but not be as huge as Battelle thinks

So I thought instead of detailed predictions I would share some of my resolutions for 07:

1. Move to San Francisco in April
2. Do my GMAT and apply for MBA
3. Start the Green Card process
4. Run 10k twice a week for the first 8wks of the year (bit of a cheat I do this already)
5. Do weights training twice a week for the first 8wks of the year (haven't done weights since I gave up rowing in 2005... this will hurt)
6. Drink water at work

Nothing too challenging on that list and in fact I have made progress on every front already. For me I am hoping 2007 will be more memorable than 2006. This may sound weird having moved to the US, bought my first car, been promoted, travelled 15 times on work, done a triathlon and run summits in San Jose and Tallinn but somehow 06 was lacking something. It was a year I didn't get emotionally involved in and I feel I lived my life as a bit of a dispassionate observer... I don't intend to do that in 2007.