Comparing Site Vists

I figured I would take another look at how the site is doing compared to last year. As of last week the site is still getting 10% more site visits than the same week last year.

Over all the web site is still getting more than 17% more web visits year-to-date, down 1% from 3 weeks ago. The first week of June had the lowest increase with only 6% more visits to the site. The largest increase in visits occurred on the last week in Jan with a 26% increase.

Over all the site is showing a 15% increase over 2008 visits, and about a 14% increase over 2007.

The graph compares web visits for 2010 and 2009 ~ by week.

Page-views [not shown] are down 4%, year to date over last year.

FAQ-Incoming-Linked-Sites

I'm up-loading a graphic from the site to reduce both the size of the graphic that gets loaded and a separate DNS look-up that is required to load the graphic. The graphic is of Linked-sites, or sites linking to interfacebus.com from external sites. The graphic was stored out on Google Picasa, and not on the server, which required the extra DNS look-up.

So the page that held the graphic was deleted and a link pointing to this blog posting added to the pages that pointed to the old page. The point was to reduce that page from maybe 12k of html text to zero, delete the external look-up for the 100k graphic and reduce the over-all load time for the site.

Last month I did the same thing to 2 or 3 other FAQ pages to reduce the over-all loading time of the site. Remember Google rank page loading time as well as just key words.

Pages Crawled Per Day

Crawl Stats;

Looks like Google is finally slowing down checking my Engineering Site. After 2 months of heavy Googlebot activity the spider seems to be taking a break, which is good. There's really no reason Google should be reading 2,000 pages a-day. The numbers shown in the graph are High number of pages, Average number and low number of pages spidered each day.

HTML Suggestions;

I finally got down to the point that Google does not have an suggestions of fixing any meta-tags. It seems like there were a few duplicate pages that were unused that I never fixed ~ so they may have been de-listed.Here is what Google has to say;
"We didn't detect any content issues with your site. As we crawl your site, we check it to detect any potential issues with content on your pages, including duplicate, missing, or problematic title tags or meta descriptions. These issues won't prevent your site from appearing in Google search results, but paying attention to them can provide Google with more information and even help drive traffic to your site. For example, title and meta description text can appear in search results, and useful, descriptive text is more likely to be clicked on by users."

Sitemaps;

1,653 URLs submitted, 1,426 URLs Indexed.

Speed Performance Overview

Just a quick up-date to show the latest graph on site download time, or site performance. Google now indicates an average download time of 2.8 seconds, which is faster than 51% of internet sites. This is down from 3.7 seconds in the last posting. The previous posting was Web Site Performance, 4/11/10.

Website Health Check

Because of the drop in web visits this month, which is common for June, I figured I would compare the number of visits this year to last.
 interfacebus.com Stats:

So there's an 11.78% increase in sites visits between June of 2009 and 2010.
June 2009; 63,814 visits
June 2010; 71,334 visits

The year to date difference is even better, with an 18.15% increase in visits
1/1 - 6/10 2009 = 994,854 visits
1/1 - 6/10 2010 = 1,175,422 visits

interfacebus [Google Sites] Stats:
Year to date for my Google Sites pages have a 104.56% increase in visits
1/1 - 6/10 2009 = 3,223 visits
1/1 - 6/10 2010 = 6,593 visits

Serialphy.com Stats:
Year to date for my serialphy pages have a 33.93% increase in visits
1/1 - 6/10 2009 = 1,441 visits
1/1 - 6/10 2010 = 1,930 visits

Knol pages [Google Knol] Stats:
Year to date for my Knol Site pages have a 42.9% increase in visits
1/1 - 6/10 2009 = 3,481 visits
1/1 - 6/10 2010 = 4,974 visits

Blogs [This blog and the other] Stats:
Year to date for my Blog pages have a 104.7% increase in visits
1/1 - 6/10 2009 = 3,679 visits
1/1 - 6/10 2010 = 7,532 visits

I guess I have to conclude the health of the sites is good!

Why You Should Blog

At first I figured I was going to just insert a reminder post about some of the reasons a web master should also run a blog. But, as I was reviewing posts looking for the last relevant posting I found some disturbing data as it relates to the rest of this blog post. I'll address that data at the end of the post, with a comment plus a few related early blog posts.

The numbers below represent visits to the Engineering Site from this blog over the last thirty days.
The date represent when the blog posts were written with the number of visits from those posts.
The number of visits from blogger represent 84 visits from 15 different pages off blogger [although they are just shown by year of post].
The point of this post is to show that even after years of writing a blog posting, it still gets visits and provides referring hits to interfacebus.com. [Why Blog; Sep 12 2007]

Referral visits from Blogger [blogspot]
by page generation date:

42 referrals from 2010 posts
0 referrals from 2009 posts
3 referrals from 2008 posts
15 referrals from 2007 posts
10 referrals from 2006 posts
14 referrals from 2005 posts

The other blog site which only shows new pages included on the engineering site brings in even more traffic.
But in the last thirty days that blog sent over 82 referring visits from 26 different blog posts.
I did notice a spike in visits yesterday which has me wondering why.

I should note that this blog had 771 page views over the last thirty days, while the 'what's new blog' only received 354 page views.
That's page views in these blogs not sending traffic to the main site which is discussed about in this blog posting.
However you can read these blogs from a newsfeed, but these numbers don't reflect that.

The graphic shows incoming visits to the Engineering Site from these blogs over the last four years. Looks like there's been a steady decline in visits from mid last year. The middle of last year I stared a newsfeed, then near the end of last year I removed some permanent links from this blog that pointed to the web site. Now the newsfeed carries all these same posts, but if some one clicks over to the web site it doesn't appear to come from this blog.

Google Caffeine and Indexing

So Google came out with a new indexing system which updates the Google index much faster than in previous years. The new indexing system is called Caffeine [Google Caffeine], and allows for a change in its index each night.

It was common knowledge that once every thirty days Goggle would shuffle it's indexed pages, so if you were number 2 in the search engine position [index] one month that same page would change position the next month. Some times the position would get a higher position, and sometime the index change would leave the page in a lower position.

However at the same time many of the new pages I generated would show up in the index the next day because I would blog about the new page address in the New Engineering Pages Blog. Google operates these blogs [blogspot] so they read or spider them all the time. Any new page mentioned in blogger gets noticed by Google much faster before it reads any external web site [I assume]. Before I started blogging I would wait 2 to 4 weeks before a page would get indexed, but after I started blogging any new page would get picked up within several hours.

So for Search Engine Optimization [SEO], blog about any new page you generate because it gets picked up much faster that waiting for the spider [Googlebot] to find it, and much faster than having to generate a new site map for each new page addition.

So what is Caffeine? Well it's Google changing your search engine position ever day instead of every month. However; because your page position might change ever day as it was displaced by one of my new pages, I'm not really sure I see the difference. However Google does indicate a recent drop in the number of my page that are being displayed in the search engine listing, or really the number of pages that are being clicked on.

So this graph depicts the the number of times my page is shown in the search engine results [blue] and the number of times somebody clicks on one of my pages [yellow] in the search results. A drop in click-through rate [yellow] in the last few days should indicate that although my pages are still showing up in the search results, they now appear lower down in the results ~ maybe page 2 of the results instead of page 1.
But at the same time Google Analytics reports that there has been no drop in visits, and I trust the Analytics report more than I trust the report from Google webmaster tools [graphic above].

If I run the above report for the term Can Bus I see no real reduction in the click through-rate.
While the graph shows all search queries [all search terms] that relate to my website.

Indexed URLs by Google

I'm always watching how many of my pages are indexed in Google's search engine, and a few days ago it reached 1,411 pages indexed. Here's a few data points over the last few months [years];

6/9/10   = 1,411 indexed pages.
5/31/10 = 1,400 indexed pages.
5/22/10 = 1,394 indexed pages.
4/7/10   = 1,309 indexed pages.
3/27/10 = 1,322 indexed pages. [Site-map loaded]
12/19/09 = 1,481 indexed pages. [Site-map loaded]
12/13/08 = 1,318 indexed pages. [Site-map loaded]

The number of pages most go up just after I add a sitemap, and then drop down a few weeks later after Google finishes reading them. However as I 'fix' the pages they slowly start to re-appear as indexed, but at a slower pace. Fixing a page is adding more data so it no longer appears as the page it was copied from.
No page rank; Component Manufacturers. Just one example of a page not indexed.....

I'm not going to show the graph but Google indicates that on average Google will spider 862 pages per day ~ every day. What is up with that, the site only has 1,600 pages? Google used 537MB of server bandwidth as the spider Googlebot down-loaded from the website.

Site views and page ranking

So the numbers are in for last month. Site visits are down over the last month, but over-all page views have increased a bit. Now if the numbers follow the same yearly trend there should be an even bigger reduction in site visits next month. I think I'll wait until next month before I re-post the page visits graph, there's really no need to post the same graph every month.

After I noticed yesterday that a few more of my pages have been included into the Google Index, I figured I would try and find a few pages that are yet to be added. It didn't take to long to find a year old page with out a page rank. The links are Component page followed by the mention of the component in the 'whats new blog'.
How to Derate 2N4239 Transistor. [posted on 6/4/2009 When to Derate a Transistor]
How to Derate a 2N3743. [posted on 1/30/2009 Derate a BJT by Temperature]
How to Derate a 2N3485 Transistor. [posted on 2/12/2009 Recommendation on how to Derate a Transistor]
How to Derate a 2N4931 Transistor. [never posted in the blog].

I found these pages by looking at the report from Google Analytics, and clicking on a few pages that were getting almost no page views. None of these four pages have a Google page rank, and have only received a few page views this year. Of course re-posting their links here may not help these pages at all, but it may help them get re-spidered ~ as they were just up-dated.

Link Checking

Just a quick note to indicate that the links on the engineering site have just been checked and verified. There were a few bad links, but not that many. I still have some sites to check within the next few days, but I need to give the web sites time to come back on-line.

The attached graphic shows the data generated by Xenu. It appears that at a minimum there were 99.29% good links on the site, and that number should increase as I check the rest on the remaining links in the report.

The last time I showed this report was in November 20 2009 as Xenu Report, Statistics for Managers. That report shows 99.25% good links, with a total number of URL links as 6950 URLs

Of course there is one day remaining in the month, but the site numbers look kind of low. Today should not be any different because the site never does well on holidays. This will be the lowest performing month of the year, still better than any month of any previous year. But I should be posting those numbers tomorrow, as my server counter does not up-date until about 5am.

Of course if you have off-page links on your site you should always check them to insure they point to a valid page address. I use a free link checking program called Xenu.

As a side note; the site visits so far this year have past all of the visits for the entire 2005 year.

Mezzanine Buses

There are a great many Mezzanine buses, some more out-dated than others. Mezzanine cards are small form factor boards that reside as a daughter card on a VME or cPCI mother-board to name a few interfaces. Mezzanine cards to not directly interface to the main system backplane as does VME for example. Most of the Mezzanine interface on this site are represented by two separate pages; one page carries the particular board manufacturers and another page covers the description of the electrical interface and mechanical form factor.

The oldest of the current Mezzanine board formats on the market is the IP Card or Industry Pack I/O Modules. Basically a non-intelligent board format used in a number of systems, a bit dated at this point in time, but there are still Companies Producing IP Modules. Page views are almost down to zero for this board format, which seems odd as it was an I/O based module.

The M-Module which is just as old still receives a few page views, but there doesn't appear to be much support in Producing M-Module Cards.

The mezzanine board format that replaced the IP board was the PMC format, or PCI Mezzanine Card. The PMC board added a controller but still allowed for any required I/O, previously handled by the IP card. So there are still many Companies Producing PMC Boards. A variant of the PMC interface is the PMC-X format. Both of these board standards used the PCI bus as the electrical interface, so really they are somewhat out dated [PCI Card Manufacturers].

The follow-on to the PMC standard was the PPMC interface, or Processor PMC format. There are a few companies that Produce PPMC Boards which were true processor based cards, on a PCI bus. Another related was board format is the PTMC Interface, or Telcom PMC standard [still using the PCI electrical interface].

These are all open standards so the board specifications could be used on any carrier card, but some interfaces were designed specifically for the VME bus, others for the cPCI Interface, and still others for the AdvancedTCA Interface. The AMC Mezzanine card was designed to interface to the new ATCA standard, being relatively new there is a small but growing number of Companies Producing AMC Boards.

Additional mezzanine boards include FMC,  and XMC standards.

Search Key Word vs. Search Key Phrase

For a number of years now I've noticed that many of my pages do not do well in a Google search when a one word phrase is used. I've never been able to figure out why my pages do better with two or more key words rather than one key word, but I guess I can't show up first for every key word.

So one example is the search term VME, which would be a one-keyword search. The VME page shows up in the ninth position on the first page. When the phrase is changed to VME Bus, the page shows up in the forth position [same for VMEbus]. When I add another key phrase, VME Bus Pinout, the page shows up as the first listing on the first page. First listing for the key-words VME Backplane, and VME Chassis also. Hmm, looks like VME Connector places first too, in both a text and image search.

I'm using the VME Bus as an example, but I could have used any computer bus because the same thing happens. I just don't under stand how Google determines what page should show up first when only one key-word is used.

Yes I know the rules, those other pages must be using the key terms more often at the beginning of their article. I' don't know, my pages do show up but I would rather have them show up higher for a few one word searches. I guess I need to somehow optimize the VME interface page up that it shows better in a Google search, maybe rewrite some of the text.

Blog Comments change

I made a change to how comments are added to a posting within this blog. Previously there was a 'Leave a Comment' line at the end of the posting. I think that text is gone now and replaced by a '# comments' which needs to be clicked on. So that text needs to be clicked, regardless of the number of current comments, it may say 0 comments. Now this only holds true if your on the main page or some archive page. If your within the posting than you will see the comments, if any, and a comment box.

So that's what I changed, I added the Comment Box to each posting. Now the comment box is within the blog post instead of on a pop-up page. However the comments are still moderated, as they always have been. Some people don't realize the comments need my approval before they get posted, because I see identical redundant comments come in.

Most of the comments in this blog are from me, because I use the comment field to up-date the blog postings. Also, like any other blogger I get comments that say very little but have a link embedded within the comment ~ I don't post those comments. Those comments are just spammers looking for a free link. Anyway it may be a few days before the links gets moderated and posted, or not.

Digital TV Reception

I finally got around to playing with Digital Television [DTV] this week. At first I didn't have a DTV antenna so I just hooked up an old FM dipole antenna I had. The FM antenna worked pretty well, as I was getting around 11 stations [Digital TV Channels]. However there was some interference, or bad reception because a few of the stations would come and go. Any why I went ahead and purchased a Digital Antenna.

I purchased a Omni-Directional Flat Digital Antenna from Best Buy. The digital antenna is a FDTV2 made by Trek. I only just purchased the antenna yesterday so I can't speak to how low it will last, but it comes with a one year warranty.

The antenna was about $55 and was in around the mid price range. There may have been a few both above and below this cost, but never having purchased a DTV antenna I just went with something in the middle price range. No I didn't read any product reviews, it's an antenna, but it is an active antenna. So I take it the only thing that can go wrong would be the internal amplifier, or maybe the attached cable. Anyway I would recommend this antenna because it seems to pick all the stations its supposed to; however I have not tried any other digital antenna.

I was wondering how long the cable from the antenna to the TV could be, but the attached 5 foot coaxial cable answered that question for me [standard BNC connector]. The antenna cable is permanently affixed to the antenna. Although a BNC Barrel [Female-Female] connector could be added to extend the maximum cable length, I wouldn't recommend doing it with out testing first. The antenna itself may be wall mounted, table mounted [comes with a base], or mounted flat on any surface

The attached table is a graphic of the available digital stations in my area from the FCC, showing signal strength, Network Callsign, Virtual TV Channel, and the Frequency band. For many of those stations listed, there are three different channels. That tells me that the stations are transmitting digital TV, but not high definition TV. [a few HD Acronyms].

The attached map is the relative location of the DTV transmitters and the DTV receiver. Although only three antennas are shown to the north many of the antennas are over lapped, with only one antenna being located south of the receiver.
Click either picture to see a larger view.

Related blog postings; HDTV Spectrum,
The TV used with the antenna is a Westinghouse TX-47F430S, 1080p HDTV [with internal digital converter].

Visitor Demographics, Web Site

Here is what Google Analytics has for the Engineering Site visitor demographics [year to date]. The graphic indicates visitors from the United States [US] only. Data is available for many other countries as well, but not all countries display this chart. There are also a number of additional tables of data that go along with graph, but there not as interesting.

Now I did get this chart from the Google Analytics, but it appears that the data was generated by Doubleclick [which Google purchased a year ago]. So I might assume that the data is not as good as if it really came from Google, because they track my visitors. However there are other sites that produce more or less the same data; but they use different methods of producing their data.

For example, Alexa produces some of the same graphs, but you would have to be using an Alexa tool-bar to be counted in their data. They indicate a small amount of Female visitors, while Doubleclick indicates none. Alexa also indicates many more visitors in the age range of 18-24, again doubleclick indicates zero visitors. I just looked at the Alexa data, but I do have a posting from 7/1/09 which shows what their charts look like [Alexa Demographics].

Quantcast also has some of the same data, but I'm not really sure how they get their information [see note below]. Here is another previous posting from 10/2/07 showing one of their charts [Quantcast data]. Their data is also different. I think it's so different I'll post their data to the right.  The time ranges between the two sets of the charts are different, but I'm just working with what I have. Either way, the two graphs show how or what different companies collect and how they display the data.

The point is who do you believe, and why doesn't Google use their own data. I guess Google is using the data from doubleclick because they already had the program. Then why doesn't Google send over the right data, they know every thing about my web site. Then again how would Google know if a visitor was a man or woman ~ there most have been some kind of sigh-up with doubleclick? 

Note: One method to generate information for Quantcast data is to add a small amount of code to your web page(s). For several months I had that code on my home page and a few other high traffic pages [out of 1,300 pages]. However I removed the code to increase page loading time. The real point is before I had added the code Quantcast had my monthly traffic at 9800 visitors, while after I added the code to those few pages the traffic went up to 39000 visitors. Currently it appears to be running around 22000, but that's not the point. How accurate is the data these sites provide.

Of course I don't care what my site's visitor demographics is, unless some company emailed and said they would only advertise if I had a particular kind of visitor ~ which has never happened.

I need to promote a few web pages that are getting no page views [less than 10 this year], sorry;
2N3867 Transistor derating curve............... 2N2944 Transistor derating curve.
2N6762 FET derating curve..................1N4454 Diode derating curve.
PCM Waveforms, IRIG.

Graphic: McDonnell Douglas KC-10 Extender in flight

Page De-listed and not found

I was working on updating a page the other day and I did a Site Search for it, so I could see how it displayed in the Web Browser. Well guess what, the page did not come up in the Google Search bar I have on the site. Two other pages that pointed to it showed up, but not the page it-self.
The page in question shows a Derating Curve for a 2N3762 Transistor, UA Package. However the two pages that showed up in the listing were for a Temperature Derating Curve for 2N3765 Transistor and a list of PNP Transistor Derating Curves. So the assumption is that the page relating to the 2N3762 PNP Transistor was delisted. Now I noticed that there is another related page; the main 2N3762 Transistor Design page and the 2N3762 Transistor in a U4 package which also should have showed up in the search results.

Now if I put a Google Site Search on my Engineering web site, I would expect Google to search my site ~ that's the point of the search bar. What Google does on their own site is up to Google, but my pages should come up on my site.

I was also working on another page [IRIG PCM Waveforms] and I checked Google Analytics stats for that page and only a few page views showed up. So I again did a Google site search and that page does not show as listed either.

These pages exist on the site, but many people may not navigate to the correct page to see the data, they may just use the search engine. I assume they click away when the correct page does not show up in the search.

Attached picture; A-10 Thunderbolt II In-flight, US Air force [USAF].

Page Optimization SEO

Just to follow up with the last posting regarding click-through rates from Google searches ~ I did look at a few of the Google search queries and worked on those related pages. Now I did not update the pages just to optimize them for the search engine. I used the list or my page in question and reviewed the web page to see if any additional work or data could be added to the page.

So in fact I did optimize several pages for the search engine, but only because I added more text to the page. I really just addressed page issues, if the page didn't require any work I left it alone. So I guess it was a good tool to use to find pages that need work. Well there are always many pages that need work, or maybe more detail, but there are always trade offs between what pages get up-dated.

Every Query in the list has a little star which I can high-lite so I'll know which keyword or query I've already worked. Of course there are a few thousand search terms in the list so it might take a little bit of time to work the entire list ~ like I'm going to do that.

Like any page up-date I make, I won't see any response or change in the page visits until the page gets spidered by GoogleBot [which could be a few weeks]. So that fact makes it hard to track page enhancements vs increased page views; as I made 14 page updates today, 30 pages yesterday and 40 pages the day before. What would I do keep a rolling list of page changes for 15 days until Google spidered them. Among other pages I updated include the 7 Segment Driver page and the 15 pin D-sub Connector page to name a few....

Oh, I added a comment to the previous post; I hadn't noticed but many of those numbers in the graph are the same [flat lines]. So the graph must be generalizing data, it's just not possible to have the same click-through rate each day.


Graphic: Iraqi T72 Tank Firing in the Desert [Soviet Designed Main Battle Tank].

Clickthrough Rate

Over the last month or so I've been updating pages to display the new Google Search bar, primarily because it uses 1kB less html code. So far just over a thousand pages have received the up-grade with the new code. But what about all the other page metrics I have to worry about, maybe it's time to work on a few of those; as in:

Page Bounce Rate

Pages with low Pageviews

Pages with a high Exit Rate
And on and on; there's just so many different ways to measure how a page is doing on the internet.

One metric I don't look at that often is search engine Click-Through Rate, or the amount of times a page shows up in a Google search but is not clicked on. Now the search term used or the query could be any term as long as it causes my engineering website to show up in the results returned in a Google search.

Of the 6,091 queries, Google provides no data at all for the lower 4,500 searches. I assume because the page impressions are to low [a percentage] to track; however I really don't know. There are another 1000 queries that display an impression but no Clickthrough data, maybe because the clickthrough rate is below 1%. Because the first term with a clickthrough shows up with a 1% clickthrough rate; which would be the worst search term with any data. The 'worst' Google search term in the report is "DVI" which I assume relates to the page I have on the DVI interface. That DVI page shows up on the first page as #8 out of 32,500,000 results. The DVI search term causes my page to be displayed in the Google search results 6,600 times with only 58 people clicking on the link [over the last month].

So there are thousands more search terms with different impressions and click-through rates. For example the term 'Derating' had only 58 search impressions but 12 click-throughs. So the question is, should I work on pages that have no or low search impression or web pages with a low clickthrough rate? What-ever; working any page based on this data would be Search Engine Optimization [SEO].

Related Blog posts 

[Custom Search Bar 4/2/10]
[Web Site Speed Performance 4/3/10]
[Web Site Speed Enhancements 4/15/10]
[Web Site Performance 4/22/10]

Graphic: Top Search Queries as Page Link Impressions vs. Clickthrough Rate.

Web Browser Usage

I figured I would post which Web Browser people use to visit the web site [the Engineering site, not this blog].
Every now and again I see some news article about how one of the web browsers is doing. Normally the article just posts data for the year, as in how one browser increased by a few percent for the year. In most cases the data is centered around how Firefox gained a percentage as Internet Explorer dropped a percent. Maybe sometimes that Google Chrome increased a percentage as Internet Explorer dropped some amount.

The point is that this chart shows four years of web browser usage for this web site, and not just one year of data.

Chart Data
Explorer; Microsoft Internet Explorer [MS IE]
Firefox; Mozilla Firefox
Chrome; Google Chrome
Opera; Opera Software
Safari; Apple Safari

The two things to note about the graph is that it only covers 2010 up to 5/3/2010, and Google Chrome only started in 2009.

Incoming Visitor Patterns by Month

At first when I saw my stats drop a bit last month I started to think my pages got re-shuffled in Google. Just because they do that ever month. I was also thinking about a few previous post relating to what Google thinks of my page loading times.

But now that I see the data for a full month, the trend appears normal. For the last five years the hits per month have been about the same for Feb and Apr [for the same year]. No they are not identical, but within a few thousand visits, the two months track each other.

Now it could be that more Engineering tasks start in March. Or maybe the jobs start in Jan but March is the month people start looking for parts and start designing the gear ~ I'm not really sure how I would ever know.

But if these trend lines continue I assume I'll see lower page visits for the next four months; although still higher than in previous years.

Taking note of a few pages added in Jan of this year
Power Line Filtering; Visits double each month [but it started at 13 visits]
Panel Mount LEDs; Visits dropped each month
Capacitor Networks; Visits going up a few hits a month
Resistor Networks; Visits are holding steady.
New page never do well the first few months any way.

Click on the graphic to the right to see Number of Visits/Month over the last five years.