Sunday, July 24, 2011

Should I combine pages to consolidate information

In the last few day I've been removing a few pages from the site. Well I remove pages all the time, but in this case I've been combining pages. The thought is, take a page that gets almost no incoming visitors and combine it with a related page, that may also not receive many hits producing a page that may get more hits because it twice as large. Perhaps not twice as large, because any duplicate information would have to be removed.

Removing a page means one less pages that have to be maintained. However it also means one less page that may bring in a visitor. The opposite view would be, one less page that competes with an existing page.

Pages on the site are not introduced based on the amount of traffic they may generate. So when a new page is uploaded there is no way to tell how many visitors the page might bring in. Normally it could takes months before a page is determined to be generating any type of visits, as it could take a month just to be found by the bots.

So any way back in 2009 two pages were generated covering 2N930 High Temperature Operation; one for the TO-18 package and one for the surface mount version. Well neither page generates any kind of traffic. The surface mount version had 103 pageviews and the metal can version had 342 page-views, both really low. It could also be that many of those 103 visitors were due to a click-through from the 'main' 2N930 page, meaning that the surface-mount page was getting even less page-views.

Now the combined page is 20% larger than before. The additional information is below the original data so should not decrease it's value, but increase the value of the page. That is, moving data lower down a page may decrease its value to a search engine. Leaving the text as it was maintains its prior importance, while adding more text increases the importance of the page. Another benefit is that reduces the number of pages similar to each other.

The pages have to be combined. The site map has to be updated. The deleted page has to be redirected and any graphics have to be removed from the server. Otherwise the bots would continue to download the graphics, using bandwidth for no reason.

Another page that got combined was the How to Derate a 2N2604 Transistor page. Having even less page views than the 2N930 page...... The next page to be combined will be the When to Derate a 2N2484 Transistor some time today. I've already redirected two of the sub-pages.

Also just combined the Derating a 2N4931 Transistor and 2N3743 Transistor page as well, which had the lowest page views of another of the other sets.

Other than blogging about when and why to combine pages, this post may also serve to get the updated versions spider-ed, so Google will see the update sooner.

The only down side to removing pages is that internal site pages see fewer internal links. The pages that were removed linked to other pages on the site, now they don't. Of course I might miss a link or forget to delete a graphic off the server.

Friday, July 15, 2011

Google Labs Site Performance

So I haven't posted in awhile. I've just been off working on the web site, which brings in a lot more money than this blog. I have started a few posts, but never ended up finishing them, so they got deleted.

Anyway I wanted to re-post some more data on Google's Site Performance, which shows page loading speed. This is their estimate of how long it takes a page to render in a browser. Than all the pages viewed that day are averaged into a single number.

There's nothing new with this post from any of the previous posts on this topic, which can be found by searching in the blog bar above for 'site performance', but I did change a setting on the server the other day.

I had the server set up to log visitor data for both AWSTATS and Webalizer. I really use AWSTATS each day, if I look at it, but there were times when I wanted to see data from Webalizer (which formats the same data a bit different). I also use Google Analytics to capture visitors data, so I had three different programs collecting the same data. The Google product is the best, but I have history data with AWSTATS so I really want to keep that running for comparing previous years.

On the 7th of this month I turned, the server-side,  Webalizer off, and waited to see if that improved the site performance of the web-site. A week later I still can't say if removing that program helped at all, but the page speed seemed to get faster {at first}.

One week of data is not much, so it's still to early to make any kind of judgment, but I am watching. The graph below shows the data for the last six months. The first big drop (increase in speed) in July is the day I turned of Webalizer. Note that the speed decreased a few days later and than decreased again, but changes in speed should be considered normal. It's the drop that got my attention.


There's nothing special about the speed, the site has hovered around 3 seconds for over a year. At this point all the pic files have been optimized, excess white space has been removed form the site and so on. Not much more I could do. Maybe send out a mess email to the world to upgrade their PC, so the site loads faster.....

I should say, that I'm always adding new data, either as new pages or additions to an existing page. So I would consider it 'normal' for the page speed to get slower regardless of what I do, because all the pages are slowly growing larger. More picture files are also always being added. So I may have reduced the [code] size of a page by removing all the white-space, but added more text off setting any speed improvement.

I'll add a comment in a few days to update the data, if it changes one way or the other. But it's something to think about, if your running a bunch of server side stuff that may not be required.

Saturday, March 19, 2011

What is Googles Farmer Update

The Farmer update, or Panda update was one of a few hundred algorithm changes Google makes to its search each year. The Farmer update got its name because the change was aimed at removing content farms from the search results.
A content farm is a web site set up just to trap key-word searches, but really has little or no content. Maybe hundreds or thousands of articles designed to bring in search traffic. However the articles are usually poorly written and are just filled with key words to attract traffic.

The Farmer algorithm update went live in the US on February 24th, with an update to non-US searches some time later but has yet to be announced. So if you have a site in the US and saw a decline in search traffic from Google starting on or after the 24 of last month you could be effected by the Farmer update.

One of the few comments Google has made regarding the update is that a weak page on your site could effect the entire site [which is some what new]. In the past it seemed that a weak page with little or no content would just never be indexed by Google, but never seemed to draw down other pages in the web site. Any way the forums are full of people giving advice, but the poster are just guesses.

So any way I wanted to detail a few changes that I made to the web site, not that I know I've been effected.
1. Deleted a dozen orphan pages from the server. The orphan pages had been out there for years and were mis-spelled page addresses or abandon pages that were no longer used. At one point they had news groups or what ever linking to them, so I saw no reason to show a 404 page not found. But after a few years, I'm sure that no one is reading those old news group posts and clicking in.
2. Redirected another dozen pages to something else. These page were being used, but had more weight years ago. As time has past they slowly lost content due to links going bad, or no interest from me in trying to fix them; however 10 years ago they were valid pages.
3. I removed about a dozen pages relating to advice on using Adsense, as their page views had been falling over the last few years and they did not have any relation to an engineering topic.
4. I removed a few ads on some of the pages, only because the content had decreased over time, because links on the pages were no longer working. Removing an ad was part of a normal page review and had nothing to do with the algo update. How ever some people in the forums are saying to many ads could effect a pages ranking, and pages with ads above the fold ~ I think that's garbage.
5. I fixed a number of bad links on the site, but I do that all the time. Out-going links are always going bad
6. I removed dozens of blog posting from early years on this blog. I've also done this from time to time in the past. Some times I'll write a blog post to keep the blog going, but in the big picture the post is of no interest to any one but the few people following the blog. But if Google is going to rank the entire blog based on these thin posts, then it's better just to remove them. This blog had 534 posts in Dec, but now has 512.
7. I've been optimizing HTML code on the site to reduce loading time, but I've been doing that for a year now.
8. Removing pic files from Google Picasa and storing them on the server, which should help with page loading and I've also been doing that for a year now.
9. I removed the Google tool bar because it seemed it was tracking me more than the site visitors. The tool bar measures page loading time and was showing wide variations when the load times should have been more even. I figured the tool bar was tracking 'my' page loads bringing up slower pages looking to optimize. In other words I would open a 'slow' page to fix, but showing Google a slow page. So Google was seeing more slow pages because of my tool bar usage than it normally would.
10. Updating pages with the new faster Google Analytics tracking code. I was doing that any way, and again to decrease page loading time.
11. Updating or enhancing page content ~ I do that any way.
12. Deleted or removed dozens of out going links, that were valid but seemed to be outdated ~ maybe they had a copy right of 2004 on their page. Normally the page had to many links any way, so no harm done.
13. I added a few more no-follow tags to some out going links, again their pages seems outdated.

I've been making these changes either before or after the 24th of last month and I've seen no difference in page views. However any decrease in page views that has occurred has really only occurred to a few high page-view pages, other pages are up or down as normal.

Comparing Feb to March the site is down 10%
Comparing Jan to March the site is down 0%
However March is normally higher than either of the previous months.

At the same time it could be that my site was not hit by the Algo change, rather some other sites are just doing better in the SERPS now ~ there's no way to tell. In any event these changes are normal Search Engine Optimization or SEO type changes.

Tuesday, February 08, 2011

Increasing Blog Traffic

First you would need to write about what people are either searching for or wanting to read about.

Although I started writing this blog in 2005, I've only just started tracking visitors. Google added a 'stats' section in July of 2010. Before that, I would only track the number of referrals being sent to my Engineering Data Base. So as of 7/2010 this blog has had 15,334 pageviews, according to the data in the blog Dashboard.

Over the last 24 hours the blog received hits for posting made in 2005, 06, 07 and 2010, which should indicate that older pages still bring in traffic. That's not to say stop writing, because one of the highest page views relates to a review I wrote just 3 months ago. However the counter did just start working 6 months before that.

So sometimes I'll blog about a new product I just purchased, or review a software package I just tried. It seems that a blog posting reviewing a product brings in a lot of visits, I guess people want to find out if they should buy it as well. Many of my posting center around some electrical engineering topic, standard, or interface bus which all bring in traffic [but to a lesser degree].

Adding a graphic might also help with increasing your blog traffic, as the blog may now show up in a pic search, in addition to a web or blog only search.
The attached photo of an graphic equalizer has nothing to do with this posting topic, but it serves to introduce the links I'm adding.

Some posts are more self serving, when I post about a new page addition to the web site; however some of those posts might also serve to introduce a new product as well.

For example a week ago I added a few new pages to the engineering site, which have yet to be spidered. So what better way to get them found, than by adding them to a blog posting [and they relate to the audio pic].
A Passive Audio Base Control Circuit.
A Passive Treble Control Schematic.
A Midrange Audio Control Example.
In fact these pages have been indexed because they come up in an on-site search, but have as yet received any page views. Of course that could just be that no one is interested in a simple audio control schematic.

This blog sends about 100 visitors over to the engineering site as referring traffic. That's compared to the 2,000 plus page views the blog receives. However in the past I was promoting the web site a bit more at the bottom of this blog than I am now. Back in 2008 the blog would refer 300 web visitors per month.

Friday, January 28, 2011

Unique Vists to the Web Site

This graphic shows the unique visits to interfacebus.com in three different views, in the same graph. The largest of the three shows monthly visits, month after month.

The smaller insert also shows unique web visits, but with the years overlapped one another. The smaller of the inserts also shows visits but as a year to year comparison.

The variations from month to month indicate some fluctuation in the incoming traffic. However the site receives several thousand visits per day, so many of those dips and peaks represent just one day of traffic, and some may just be the site counter going off line.

Some of the months perform better or worse than other months only due to the number of days in the month or holidays occurring during the month. A work day brings in several thousand visits, while a week-end may only bring in 50% of that. A holiday only drops the visits by 75%, mainly due to 45% of the visits coming from other countries.

This posting allows the blog to serve the graphic instead of my server, but the web site does point to this page.

Click the graphic to see the larger view.

Although it's not that relevant any longer, the large dip or drop in visits during 2005 was due to changing the web counter mid year.
.

Tuesday, January 18, 2011

Year End Site Performance

Last year ended up with the highest number of visits to the web site so far. Google would say that the total number of visits is not the best metric to watch, but I would contend that any visit is a good visit. However Google likes Time on Site or Bounce Rate, but a bounce [one page view] does not mean the visitor does not come back later.

The graph shows the number of visits to the site. The number of visits indicates a visit and/or a returning visit from the same person, as opposed to unique visits, when a returning visitor is not counted. The graph compares month-to-month visits and year-to-year visits. Except for December, this year out performed every month in all the previous years.

Around mid-year I finished implementing the new Google search bar. They say it's better {?}, but the code is smaller so I changed out the code. In 2008 I changed out all the Google banners, for that new code too, which was also smaller.

Any time I update a page I look for white space I can remove to reduce page loading time. White space is a 'space' in the html code that severs no function. Normally I can reduce a page by 200 bytes for a 10k html file and 2k for a 50k page by deleting [not visible] white space. The down side is that I also add new content when I up-date a page, so in many cases removing the white space evens out with the new text.

So it's been seven months now of optimizing the web site to increase Google's Site Performance data. Now, 7 months later the site's performance is 2.8 seconds [average load time] which happens to be what the loading time was back in June. However the average 'average' is closer to 2.8 than before, which hovered into the 3 second range.

Two weeks ago I started trading out the Google Analytic tracking code, which they also say is faster. So I hope to see a speed improvement in a few more weeks, as I get more pages running the new code. Currently there are just over 300 pages that are running the new Analytic code, but the Site Performance data is weighted so heavy trafficked pages count more. I up date pages based on need, not traffic.

Any way I'm always adding new content, regardless of any SEO things done to the site.