‘Cause I’m seeing right now a huge improvement on one website that was quite severely penalized by Panda 2.5.

The only change that I made on it was to improve the bounce rate, as I was counselling here (yup, with some simple design tricks anyone can improve this parameter). Before the Panda 2.5 the bounce rate was constantly around 85%, now it is constantly around 75%. Of course, by improving the bounce rate automatically I improved the page views stats, too.

Now, since this morning, I’m seeing on this website some 25-30% extra traffic, which is very encouraging. In fact, taking into account that Panda updates are coming approximately once per month, it is about time for Google to rollout the Panda 2.6 update.

In the next hours and days we’ll find out more about this.

The usual bullshit. Blah blah.

Unfortunately, the Panda update is much more difficult to deal with, because, as I said here, Google started to consider into its algorithm some usage factors that normally aren’t under the webmaster’s direct influence. Technically, to recover from Panda you should:

• try to keep the users more time on your website; try to engage them somehow (maybe with movies?… online games?… p%rn pictures?…);
• try to make the users to visualise more pages; more page-views / user is better (you should split the content into several sub-pages – I know this sounds totally anti-Panda 1.0, but it seems that Google now has a different view than 1 year ago);
• try to make the user to return to your website several times after the initial visit coming from a search; with every search visit, Google places cookies in the user’s browser and then tracks the returning rate, which then uses as an indicator of a website’s value;
• try to lower the bounce-rate.

Don’t forget: Google has a lot of usage tracking mechanisms. Analytics, Chrome browser, search cookies, social media sharing monitoring etc. … you really can’t trick Google into thinking that your website is more beloved by the users than it really does.

Good luck with your Panda 2.5 recovery!

Well, after carefully reading a lot of personal experiences and industry reports (such as this one from SearchMetrics) I think that I’ve found a pattern of the affected websites, and this is usability-related.

More affected were the following websites:

• websites with a low pageviews per user average;
• websites with a high bounce rate;
• websites with a short time-on-site metrics.

More favoured were the following websites:

• websites with a high number of pageviews;
• websites with low bounce rate;
• websites with nice time-on-site metrics.

If you scan the Alexa data related to the websites listed by SearchMetrics blog, you can see that the affected websites are all struggling with these parameters: bounce rate, pageviews per user and time-on-site, while the most favoured websites are displaying, all, strong data. However, some losers display good numbers.

For example ActressArchives.com is cited with 62% positions loss after Panda 2.5. In Alexa, the usage parameters are quite good: ~6 pageviews per user, ~50% bounce rate and ~29 minutes on site. This is much better than PerezHilton.com or TMZ.com, but PerezHilton has a very nice bounce rate: 39%. And PerezHilton is cited on the winners’ list. So, it seems that the bounce rate was the essential metrics to be taking into account.

[However, I suspect that Alexa's data on ActressArchives.com are erroneous. On Google AdPlanner they are much different: 3,9 pageviews per user and only 3 minutes time on site.]

If confirmed, it is a clear sign that Google started to melt the search data with data gathered from its other products: Analytics, Chrome, Plus etc. Since not all the webmasters are using Google Analytics, I suspect that the Chrome browser’s usage statistics are the main factor that triggered this Panda 2.5 update. But I am also sure that Google will release very soon an update to this update, simply because it’s quite stupid to link the relevance of a search with the users’ behaviour. That’s why we are seeing now a lot of excellent, unique websites that are hit by this … failed Panda Update 2.5, so don’t despair. I’m very confident that the Panda 2.6 update will come in 3-4 weeks and will repair the damage.

Scotland won several times the Six Nations tournament. During the World Cups, Scotland always qualified for quarter finals and in 1991 they even qualified for the last 4.

Unlike football, rugby union is a sport with very cemented hierarchies. Since the very first World Cup, which took place in 1987 in New Zealand, rugby’s big forces are always the same: from Europe – England, France, Ireland, Scotland and Wales; from Southern Hemishpere – Australia, New Zealand and South Africa. Every World Cup  semifinal was played between teams chosen from these 8, with one, noticeable exception: Argentina in 2007.

These teams form the rugby’s perenial elite. There are also some contenders, equally cemented. These are Italy, Argentina, Fiji and Samoa. From time to time, these teams win a match against the worst teams from the big 8 and they even manage to play in a World Cup’s quarter finals. But in general they always lose. Only Argentina in 2007 managed to win matches against France, Ireland and Scotland and finally get the 3rd place. But this was rather an exception.

Then come “the others”. Teams that never – and in rugby never is really NEVER – win against the big 8, and rarely win against the secondary 4. Romania is such a team. In general we win one match per World Cup, against the weakest team. In 2007 we won against Portugal and in 2003 against Namibia.

That’s why today it would have been a massive shock if Romania managed to keep the Scots to a 24-24 tie. Unfortunately the Scots scored two tries in the last 5 minutes and won the match 34-24. But for us the display was excellent and full of hope: the hope to lose honorably against England and Argentina and win the final match against Georgia…

Now, after a week of usage, here are my findings:

• Don’t buy it if your room is really hot, because this small appliance will not improve the room’s temperature. I’ve use it with a big block of ice but the inside temperature remained absolutely the same. Not a single degree in minus. It just doesn’t cool the air.
• What this appliance really does, is to increase the level of humidity in the room. In my room the level of humidity was quite low, 37%, and thus the air felt very hot and dry. When this appliance started to work, the air became instantly fresh and pleasant. Although the real temperature wasn’t lowering, the body was feeling much better.

Some people say that this EAC is nothing more than an overpriced cooler, but they are wrong. A cooler doesn’t improve the level of air humidity (relative humidity). I would compare the evaporative air cooler’s effect with the effect of a waterfall in the mountains, in a warm, sunny day. I suppose that anyone has seen at least once in a lifetime such a waterfall, and knows how pleasant the air is around it.

So, in conclusion, if your needs are limited, you can try the evaporative air cooler. I would say, if the temperature in the room doesn’t go over 26-27° C / 78-79° F, and the humidity is low, you can give it a try.  Otherwise, buy an air conditioning – it’s better to be safe than sorry.

1. Why not a Kindle?

I searched and looked at a lot of Kindle reviews on Youtube and other video sharing websites, but since the recent discovery of Apple and Android security failures, I simply don’t trust anymore the big brands. I know that Kindle looks impressive, but the hidden marketing strategies that must be incorporated in it and the limitation of reading formats were enough and strong reasons for me to drop any choice regarding a Kindle purchase.

2. Why a Prestigio? What is Prestigio?

Prestigio is a brand originating from Cyprus. In fact, the parent brand is Asbis. Asbis is one of biggest IT hardware distributors in Eastern Europe and Prestigio is the brand-name for their e-readers and tablets. Most of their products are researched, developed and manufactured in China, but definitely not by any of the so many low-profile Chinese producers.

Currently, Prestigio sells three models featuring an e-ink display (I didn’t even bother to take into consideration buying an e-reader with TFT display – a tablet would be a better a choice if one needs to watch movies or color pictures). The low-end device is Prestigio PER3362, which costs around 100-125 euros and it is strictly an e-reading device, while the high-end device is PER5062 which has a touch-screen display, wi-fi connection, touch-screen keyboard and is sold for around 180-200 euros.

In between lays Prestigio PER5162 which looks much similar to Kindle 3 and is sold for around 150-170 euros. PER5162 has a 6” display, wifi connection and built-in mini-keyboard. Good to know, every Prestigio model, no matter the price, comes with a free leather cover and a pair of headsets, which are a very nice and useful add-on to the package.

PER5162 comes also with a 2Mb storage card.

3. Technicalities

There is little to be said about this e-reader. I think it is one of the best.

It can read the most popular formats: epub, fb2, pdf, doc, txt, rtf. When using the fb2 format, one can take notes and markups. The pdf reflow function is excellent, very convenient and it works smoothly. This e-reader can also reflow docs, but with long docs the processing time it’s much longer, so my suggestion is to save the docs as pdfs and only then put them on the e-reader.

The books can be organized in folders, with no matter how many levels in depth, for example “Authors – A” > “Aristotle”, “Authors – B” > “Bakunin”, “Baudelaire” etc. Thus, you can really organize an entire library into the card. The e-reader supports mini-SD cards up to 32 Gb.

Of course, the e-ink display makes it easy to read in plain light or in the sun. There is no screen reflection. The drawback of this technology is given by the fact that you cannot read without an external source of light.

The e-reader is very easy, it weights around 300 gr. including the leather cover. The buttons are very soft, but firmly. The main input control is a mini-joystick very similar to the joystick from PSP. I’m curious to see how long it will last, but it seems to be very solid.

The wi-fi and browser function is very convenient. The wi-fi receptor is very powerful, while the browser (a Linux browser) is excellent for basic tasks, such as checking emails or reading a newspaper frontpage. Of course, everything will be seen in black and white. However, from what I have seen in videos over Youtube, the Kindle’s web browser is more powerful. Luckily I don’t need it.

This e-reader also features two little games, a built-in radio, a photo viewer and an mp3 player. Honestly, I didn’t use so much these functions, but when I tested it looked ok. According to the specifications, you can read and in the same time listen to the radio or music player.

The reading options are very diverse. No matter the format, any document can be zoomed in and out (and there are at least 4 levels of zoom). The page can be fitted on the screen for the best reading experience. As stated before, with some documents, such as pdfs and docs, you can reflow the text, then zoom in and out until you have the best option. In the gallery, you can see a page from a pdf document, in initial mode, then in reflow mode (with the maximum zoom out level). As you can see, the reflowed text looks great and the page is much easier to read.

A notice about the gallery pictures: I did the pictures using the flash light. As you can see, the e-ink display looks exactly like a paper page. It doesn’t reflect light, it doesn’t “kill” your retina.

Although you close a book, the e-reader remembers your settings so the next time when you open that particular book, it will open exactly where you leaved it, and how you leaved it. All the reading activity is stored under the “History” menu.

I didn’t have until now any trouble with special Romanian characters, such as ş, ţ, â, î etc. I think that this e-reader can display virtually any european language. When taken out of the box, the storage card is filled with free books in a lot of languages, from English and German to Russian and Polish. I was curious to open one such book from every language, and I didn’t see any trouble.

There is also a voice reading option, but in my opinion this needs to be improved. There are only two languages, Chinese and English, and the English voice is quite dubious. When reading a literary work it really sounds bad. It works much better with journalistic texts.

4. Conclusions

Definitely, Prestigio PER5162b is one of the best e-ink e-readers currently on the market. It is a little bit more expensive than Kindle, but the package includes the leather cover and it reads virtually any e-book format. I’m highly recommending it.

As a final word I’m also recommending to download and install the firmware upgrade from the Prestigio’s website. It adds several options in the reading menus and solves some bugs with the wi-fi and the browser.

Later edit: other names / brands of this e-reader:

- Boeye G6 e-reader
- Delyca S600k

In short, I wanted to link within a conditional comment a special stylesheet inside the head of a document. Usually, this is a very simple task.  I made it a lot of times before, but I never noticed the strange requirement explained below.

My code was something like this:

<!--[if lte IE8]>
<link type="text/css" related="stylesheet" href="blabla.css">
<![endif}--> 

To my surprise, IE did interpreted correctly the code from "blabla.css" but displayed the following line on the top of the page:

<!--[if lte IE8]><![endif}-->

which was a totally non-sense. It was a triple non-sense, because a) this expression is a comment, b) anyway, this comment was inserted in the head section, and c) the IE interpreted the css code.

If you don't believe it, just use the following code and see it for yourself:

<!DOCTYPE html>
<head>
  <title>Test</title>
  <!--[if lte IE8]>
   <style>
   h1 {color: Red;}
   </style>
  <![endif]-->
</head>
<body>
  <h1>This should be red in IE8</h1>
</body>

The example above will output a red H1 in IE8, preceded by… a strange line that displays the conditional comment’s expression.

After scratching my head and trying to find a solution to this uncomprehensible situation, I finaly managed to deal with it. The problem was the name of the browser. When writing the comment, one must carefuly obey the syntax recommended by Microsoft and write <!--[if IE 8]> instead of <!–[if IE8]> (with a space between “IE” and “8″). This is all! The following example will output a correct page and will stop triggering the abnormal behaviour, despite the fact that the only difference is given by a space between IE and 8.

<!DOCTYPE html>
<head>
  <title>Test</title>
   <!--[if lte IE 8]>
    <style>
     h1 {color: Red;}
    </style>
   <![endif]-->
</head>
<body>
   <h1>This should be red in IE8</h1>
</body>

When we had to cope for the third  time in a row with such an incident, I decided to study the log files and I was amazed to discover that in the hours preceding every single incident the server was “attacked” from 3 IPs, coming from the same C-class: 216.104.15.* There were a lot of automated tools that were requesting non-existant pages from our website, by running into an infinite loop that was requesting something like this:

www…/list-services/js/functions.js
www…/list-services/js/js/functions.js
www…/list-services/js/js/js/functions.js

It was obvious that these requests were coming from some sort of search bot too stupid to understand the <base> tag.

Digging more into this, I discovered that this C-class pertains to Trend Micro, a shitty anti-virus producer. Then, searching more on Google, I found out that this shitty anti-virus installs on users’ computers a bullshit tool that “verifies” if/whether a website spreads malware or not. Unfortunately, their tool seems to be quite stupid, since it cannot correctly parse the HTML code of a page.

That’s why I decided to ban with .htaccess all the requests coming for the said C-class. And digging even more, I discovered that they are using a second class, too: 150.70.64.* and that a lot of people are starting to complain about them.

Everybody knows that high authority backlinks are the essential ingredient to obtain good positions on search engines. Unfortunately it’s very hard to obtain authority links, however there are some links that are easy to get.

I had to find recently some authority backlinks for Pinbud. We all know the theory: high authority links can be obtained only from:

• educational domains
• government-related domains (including foreign governmental websites)
• high PR websites
• established, old domains with content related to your website’s content

The theory is easy to learn – but the practice is a nightmare.

There are only 2 ways to get a backlink from an educational or a governmental website:

• the legitimate way: you get the backlink because you’ve done something and you deserve the backlink. This is very hard to attain, because 99.999% of websites have nothing in common with any educational resources;
• the “hacking” way: you can find a website that’s vulnerable to XSS (or better) so you can place a hidden backlink. While this is a feasible scenario, it requires tremenduos efforts. You must check websites one by one and do tests. In my opinion, it’s simply worthless.

Governmental links are very, very powerful. 2 months ago, intrigued by a post from WebmasterWorld, I’ve made some research and I discovered some highly positioned websites that had only 4-5 backlinks. But all these links were coming from some Chinese secondary governmental websites, and that’s why these websites were occupying really good positions. It was amazing!

High PR websites can be anything, from news to free classifieds, but the most common are the news-related websites. A news-related website will always feature at least PR 4-5, but the most important websites, even the websites of local, minor, newspapers, easily reach PR 8-9. Links can be paid or obtained for free.

After carefully reviewing a lot of possibilities, I finaly decided to pay some bucks, but to ensure some trusted backlinks. Remember, your time costs money: the timeframe required to obtain for free such backlinks can be so huge, that it’s simply worthless. So, I’ve decided to pay links from:

• Yahoo! Directory. Yahoo! Directory has PR 8. A link from Y!D is $300 / year. Choose carefully your category page. Choose a page related to the content of your website, preferably featuring some competitors, and with a good PR (at least 3-4).
• BOTW directory. Best Of The Web is another old, trustful directory. It has PR 7 and a backlink costs $299.95 (one time fee).
• Business.com is the third major directory, featuring PR 7. Price: $299 (one time fee).
• JoeAnt. JoeAnt directory has PR 7 and a backlink costs only $39.99. However be very careful, because JoeAnt doesn’t seem to be fully indexed by Googlebot. You should submit only to a category that’s indexed by Googlebot.

Thus, paying only $940, we’ve got 4 authority backlinks. But this is not all: Yahoo! listed PinBud not only on the main directory (http://dir.yahoo.com) but also on several subdomains. The same with Business.com. I submitted the website to a category and the editors placed our link on an additional (but relevant) category. In fact, the $940 price paid us not only 4, but 8 authority links.

Beside these we’ve got some backlinks from some minor directories that are probably getting information from Yahoo! So, the money were well spent.

Another precious backlinks that we’ve got in the last weeks are 2 links from Microsoft. As anyone knows, PinBud is powered by Microsoft Bing Maps technology. The coding itself is not Microsoft related (the platform is a classic LAMP – Linux/Apache/MySql/PHP) but the core function, companies listing and positiong, is build on top of Bing Maps integration. So we were entitled to submit our company description to Microsoft Bizspark and now we have a live profile on Bizspark and a link from Microsoft Platform Ready.

And finaly, another worth to mention authority link source is the DMOZ directory – but, unfortunately, it’s very hard to get there. Just submit your website and forget about it. Come again in 4-5 months and, if you don’t see your website, submit it again or try to find on Elance or another marketplace an editor that will agree to let you in for few bucks.

We have seen almost immediately the effects of these backlinks: the traffic on PinBud.ca went up like crazy. And there are enough signs telling us that PinBud.com will be soon out of the infamous “sandbox”. So, the conclusion is crystal-clear: start to build authority links with the most easy to get!

In short,  Google’s engineers have read a story from New York Times that raised some (good) questions: knowing that link popularity is the basis of the Google’s algorithm, is it fair to rank high a website that has a lot of negative reviews?

That’s why they have decided to  perform a sentiment analysis and to position websites according to these results. So, now it’s more difficult to position a website that only enjoys negative reviews.

Unfortunately, in my opinion this approach – sentiment analysis – opens  largely the door to the most unorthodox tactics for search positions manipulation. In fact, to unfair competition.

One of  Google’s most respected principles was that “there’s almost nothing a competitor can do to harm your ranking or have your site removed from our index.” According to the official Google Help Guide,

“Can competitors harm ranking?
There’s almost nothing a competitor can do to harm your ranking or have your site removed from our index. If you’re concerned about another site linking to yours, we suggest contacting the webmaster of the site in question. Google aggregates and organizes information published on the web; we don’t control the content of these pages. [source]”

But now, if Google’s takes into consideration the so-called sentiment analysis, it’s clear that anyone wishing to destroy your website can make it very easily, just by posting some negative reviews on review websites.  Let’s look at this page featuring reviews for DecorMyEyes, the online shop that triggered Google’s reaction: it’s really full of negative reviews! And this page too! And this page! The reviews for DecorMyEyes.com are really awful, but, by commenting on this issue, Google simply showed to everyone the way of harming the competition.

I think that we’ll experience in the next months a growing demand for SEO’ers specialized in reviews writing !