Many small B2B companies are wary of running PayPerClick campaigns because they do not want to waste their resources on a complex and difficult to understand process. They are afraid that the system is setup to make Google money and they cannot afford the time to learn how to change the default settings.

It is a difficult system to become expert in but its worth the effort because a well run PayPerClick campaign delivers these:

• Valuable enquiries
• Lots of business insight
• Best converting niche keywords

You get valuable enquiries when you start thinking about how your potential customers refer to their problems online; when your ads reflect how your potential customers speak; and when your landing pages address their problems as specifically as possible.

You get lots of business insight because your tracking systems help you identify which keywords, ads and landing pages give the best conversions to sales.

You can use your newly discovered niche keywords to create new content for your existing site. The pages will then be better optimised for natural searches and your organic traffic will become more relevant.

In summary, your PayPerClick campaigns will begin to work for your business if you:

• Understand your potential buyer
• Target the keywords they are using
• Track and measure everything

I specialise in getting quality enquiries for B2B clients from PayPerClick campaigns. Call me, Chris Bose, on 01488 674 203 for advice on how to get started.

In the beginning of this work, I did not delve into the quality of the similarity sets too deeply, hoping that the script would sort out any inadequacies.

Well quality in to get quality out. Obviously check the URL out to see if its relevant to your theme. Then check the log to see that the script is downloading the pages in the site, because sometimes it isn’t. This could be because of server issues or robots.txt files.

You can increase the quality of the runs by downloading entire sites with Web Dumper and serving the pages from your own server. This way you don’t piss off server owners by using their bandwidth every time you do a run, and you can tweak your own server to make sure the pages are being delivered in the time required. See the quality of your results go up!

Because the files are being accessed “locally” I can now run as many simultaneous runs as my bandwidth will cope with for a particular similarity set.

WordPress 3.2 requires PHP version 5.2.4 or later. My hosting server recently upgraded to 5.3.6. This generated a lot of error messages from WebCalendar.

To turn them off, edit your php.ini file found in /etc/php5/apache and turn display_errors and log-error to Off.

Restart the apache server by ssh’ing to your server and running the following as root:

/etc/init.d/apache2 restart

Don’t forget to exit when you are done

A new WordPress install worked fine on an 1and 1 account, but when I activated Thesis I got a Server 500 error on all the admin pages even though I could see the website.

So google search gave a page on the WordPress forums which said to add a php.ini file to the wp-admin file with the line:

memory_limit = 32M ; Maximum amount of memory a script may consume (8MB)

Did this and got back all my admin pages immediately.

To produce useful results, a focused web crawler needs a good starting URL and a relevant similarity set. I began by thinking that advanced manufacturing was a good niche. But the crawler is telling me that it is no niche at all. The results have been very variable and poor quality and my interpretation of this is that the similarity set is really composed of unrelated URL’s.

The only niching was in my mind, the script does not see the connections my mind thinks it makes.

Advanced manufacturing niches do exist of course and I shall be exploring print electronics, micromoulding, 3D manufacturing, 3d scanning, and optoelectronics.

Our focused web crawler produces large raw data text files up to 150 Gb. These have to be split to work in a Text  editor like BBEdit whose top limit is 300Mb.

You can use a utility like Split and Concat but this can only be used manually and has no batch facility.

I wanted to schedule split to run overnight, and found the ‘at’ command.

However, the ‘at’ command is used to schedule the commands and the’atrun’ utility is used to execute the jobs.

By default, the ‘atrun’ utility is not enabled so any jobs scheduled just sit there and never run. To enable ‘atrun’ execute the following:
sudo launchctl load -w /System/Library/LaunchDaemons/com.apple.atrun.plist

Now you can use the ‘at’ command.

Thanks to DeveloperCoach for this

To disable the atrun command from running, you can again manipulate launchd settings via launchctl.  You would do this if you were not actively using at and wanted to prevent extraneous disk access and to ensure the computer sleeps properly.

sudo launchctl unload -w /System/Library/LaunchDaemons/com.apple.atrun.plist

Splitting a 27Gb into 300Mb chunks took 16 minutes on my desktop, running in the background.

Starting URL’s are first determined by identifying useful directories, PR sites and sometimes listening to your spam email. Thats how I found www.splut.com.

But there will come a time when these avenues are exhausted and you have run all your target markets and similarity sets and you are stuck.

Then I had an idea out of the blue because I had been thinking hard about it. I believe that the solution may be in a Google search. Search for some relevant key phrases after turning off Google instant and selecting 100 results per page.

Clean up the page by removing the irrelevant links Google puts there, install MAMP, put the page in the htdocs file and add the local path to your script.

You have now constructed your own focused directory in a sense which you can populate with as many links as you think necessary by adding other search results pages to your own.

Then test and re-test.

I have found that best results come from a page that is the result of an advanced Google search with things like -pdf and -jobs in the search string. I am sure you can think of others.

Your focused web crawler now has access to an infinite number of starting URL’s. This combined with the ability to run multiple python scripts in different windows on powerful desktops leads to monthly terabyte bandwidth usage yielding 20 to 30 new target companies per month.

A focused web crawler needs a similarity set and a starting URL to run. The similarity set is the set of URL’s that resemble the target market sufficiently to guide the crawl. The starting URL is where you begin.

Initial similarity sets come from your client and prospect database.

A useful way to determine new target markets in high technology areas is to LISTEN to directories and take your direction from their categories.

Our focused crawler sometimes only finds results within the starting URL directory. This is most common with Applegate, Zibb and EngineeringTalk.

I call these “closed” because they only produce results from within the domain.

The results are accurate and precise, so this directory run could be used to find relevant targets for a similarity set quickly and so aid the wider search later on.

Ushahidi is free and open source software for information collection, visualization and interactive mapping.

Swiftriver is an open source platform that helps users add context to realtime data.

Sweeper is an application within the above framework. I am not sure of its capabilities but what caught my eye was this: “Automate the addition of context to data: Location, priority, influence, reputation, tags etc.”

So I tried to install it to have a play. I found 2 install links: the first on the Ushahidi wiki and the second at the Swiftriver Google Group.

Both were too tedious and long winded for me.

I installed Sweeper on Mac OS 10.5.8 in 5 minutes:

Download and install MAMP.

Download Sweeper and put the folder in the htdocs folder.

Start the MAMP server and open your preferred browser at localhost:8888/sweeper and begin to follow the install instructions.

The browser tells you that PEAR logging is not installed.

Open a Terminal and follow these instructions:

curl http://pear.php.net/go-pear.phar > go-pear.phar

sudo php -q go-pear.phar

Then change the install location to /usr/local. Thanks to Bombshelter for this.

The install will then tell you your configured directory: /usr/local/share/pear

Then run cd /usr/local/bin

sudo pear install log

sudo nano /Applications/MAMP/conf/php5.3/php.ini

In nano add the following to the php.ini file:

include_path = “.:/usr/local/share/pear” underneath the line include_path = “.:/Applications/MAMP/bin/php5.3/lib/php”

Restart the MAMP server. You might have to restart the whole computer sometimes.

Go back to your Sweeper install page and finish the install.

Simples!

PS: of course, it took me a few hours to get the procedure above just right. But for you it will only take 5 minutes.