These days banks try to minimize client visits at branches through their pricing policies set in favour of using online services. For active clients internet banking is one of the key channels for carrying out banking transactions. Its usability and the overall quality of user experience is therefore very important.

According to Eurostat, the EU’s statistical office, 15% of the Slovak population used Internet banking in 2007 – a higher share than in the neighbouring countries. These clients are lucrative for the banks, since Internet is accessed disproportionately not only by the young but also by people with higher socio-economic status
.
In early 2007 Pizza SEO began working on a large rating of Internet banking applications of Slovak banks. Our methodology is based on the so-called user personas technique – we created three fictitious but representative users who attempted to complete specific tasks in the various Internet banking systems.

As a result we defined 60 indicators used to test applications of Slovak banks. The list of indicators along with our resulting rankings of usability of Slovak Internet banking applications will be published at the end of this exercise.

Our check-list focused on two aspects of usability: efficiency and errors since these were the easiest to measure objectively. We evaluated to what extent the functions we identified as necessary can be easily accessed by a regular user and whether the application can withstand user errors (ability to enter data in varying formats, etc.). For selected tasks we tracked the so-called click distance (number of muse clicks required to complete task). We based our methodology on several international research reports on the importance of individual functions a features of Internet banking applications.

Pizza SEO if financing this test – we had to refuse the option to be supported by banks themselves to avoid a conflict of interest. This is one of the reasons why the ranking has been so long in the making.
However, while testing each application our consultants have come across a variety of interesting aspects, pleasant and unpleasant surprises. We are therefore publishing a brief videocast (in Slovak) on each of the banks. Some of our findings may be of interest to developers of online banking applications but many are relevant to other types of applications and will therefore be of interest to a broader audience.

We released the first videocast on Komercni Banka on 7th May 2008 on our Slovak blog, several others have followed and we will release them with approximately on a weekly basis. The ranking table of all banks will come at the end and will be provided in Slovak and English. We look forward to your comments and experiences with Internet banking in Slovakia and in general.

In addition to Slovak, OpenPhrases.com offers English, German and Russian. I have little to say on the latter two versions as we rarely if ever get to use them. On the other hand, we often get to use the English version both in our client work and the work we do on our own projects (to supplement, rather than replace other tools, as I explain below).

Unlike with Slovak, this is where OpenPhrases.com enters into a competitive arena with several heavy weight competitors. How does it stack up? Where does it fit in a search marketer’s tool box?

Keyword Analysis

Keyword analysis is bread and butter of any search marketing effort – both on the organic search optimization front and for paid search advertising. In performing keyword analysis SEOs usually look to:
1. Gather keywords and phrases to use in organic optimization and paid ads in search engines,
2. Assess importance of words and phrases (absolute and relative) in terms of search volume,
3. Estimate competitiveness for a given word in organic or paid search.

There are a number of tools usable for simply gathering keywords (e.g. to create a list of words for a PPC campaign) - you can mine logs, use suggestions from search engines or even a simple thesaurus. You can easily load tonnes of keywords in a pay-per-click campaign and see if they receive any searches.

Competitiveness of search phrases can be estimated using Google itself (popular method simply looks at the number of results for the given phrase, although this is only a proxy for competitiveness at the top of the SERPs). Tools within online advertising systems give an indication of competition in paid search.

There are only a few tools to look at search volumes and assess the absolute and relative search frequency of various phrases (especially since the highly popular Overture Inventory Tool went dead in December 2007). Three major tools are used: Keyword Discovery, Wordtracker and Wordze. For a serious search marketer these tools are a de rigeur regardless of their prices – some marketers swear by one of them but many use several to cross verify the info they get.

Having said that these tools can actually be quite costly – monthly subscriptions run from about $45 (Wordze) to $69 (Keyword Discovery). Lots of people do subscribe and there is obviously value to each one of them (see Copyblogger for a great review).

But OpenPhrases.com is completely free (”at least for some time” according to the creators’ blog). Therefore I think it only makes sense to compare it with other free tools or free versions of paid tools.

So I ran OpenPhrases.com head-to-head with the Google Adwords Keyword Tool, Free Search Term Suggestion Tool by KeywordDiscovery.com and the Free Keyword Suggestion Tool on Wordtracker. I chose the search for ‘real estate’ as one I expected to have reasonable search volumes. I also played around with other searches (’shirt’, ‘Bratislava’, ’sex’) but my impressions were very similar.

For Google I turned the Synonyms option off, for Wordtracker I left the default Remove offensive (results) option selected and for Open Phrases I unchecked Synonyms but left Stemming on.

‘Real estate’

For real estate Google’s results looked very reasonable (although Google does not give any specific numbers, just green bars and you can therefore only see the order by volume). ‘Real estate’ as a phrase appeared first, followed by ‘real estate listings’ and ‘commercial real estate’. The first geographically specific phrase and the only one in the top 10 appeared in the seventh position – it was ‘florida real estate’. Overall, the results looked quite US-centric.

The Keyword Discovery tool also showed ‘real estate’ in the top position, with an overwhelming 431,375 searches (to 23,033 for second-ranked ‘Las Vegas real estate’). The other geographic phrase in the top 10 was again ‘florida real estate’ in the 8th position. Several real estate agents (prudential, century 21) also showed. All in all the first 100 results (showed for free) appeared pretty sensible, too.

Wordtracker started off with ‘Marbella real estate’ and seven out of the first 10 searches were geo-specific (places like Portland, franklin tn, fort myers). ‘Real estate’ itself came sixth. Interestingly, search volume on Marbella was reported at 23,526 compared with 18,504 for real estate itself.

Results returned by Open Phrases were somewhat suspect: ‘brava costa estate real’ came first with 568 searches (about 300 of them on a single day according to the detailed info I got when I clicked on [DETAILS] next to the phrase, which brings up a cool 92-day histogram) followed by ‘califormia real estate ray and peggy pierce’ with 310 searches. All the other top 10 searches were also geo-specific, ‘real estate’ on its own appeared nowhere in the top 50.

Digging into the results gave some further clues on their quality: clicking on the third placed ‘crete estate real’ with 253 searches showed there were also five for ‘real estate crete’ indicating these could be drawn from distinct data sources (one that preserves the order of words and one that does not). Clicking on ‘real estate agents’ showed multiple variations with the word kanpur: ‘real estate agents at kanpur’, ‘in kanpur’ and ‘at kanpur.’.

Fun with keyword analysis?

OpenPhrases.com shows promise – nice interface with some great features such as the three-month search history, limiting searches by phrase length or frequency or looking for synonyms (as a side note, what I found quite annoying was the auto-suggest feature which showed phrases as I typed – when I hit enter using Firefox, the tool searched for the top auto-suggested phrase and not the on I had typed in).

At the moment, however, it lacks a database of sufficient quality to be a reliable source for keyword analysis. This is not surprising - finding usable and representative search data to play with is not easy in a world 80% dominated by one search engine. Say the creators: “Current database contains more than 45.000.000 unique keywords (unique search queries) and more than 500.000.000 search queries. We got them from world-wide used search engines, if you ask which, sorry, we keep this information secret.”

As a free tool OpenPhrases.com is great for individual marketers or amateur SEOs to supplement the Google Adwords Keyword Tool and free Wordtracker – i.e. it is great, as its tagline says, “…when keyword analysis is fun” (though maybe not when it is supposed to be serious work). It can provide extra info to anyone who does not need or cannot afford to pay for full access to one (or several) of the heavy duty tools.

If and when the creators get their hands on a more extensive search data feed and sort out minor usability and data parsing niggles, the Open Phrases tool with its friendly interface can become very useful.

Not much, except for a post by Michael Martinez at Seo Theory blog. I know, Michael meant optimizing for “seo”, “pizza” and “Britney Spears” separately, but I guess optimizing for those three terms together could bring some SEO-targetted traffic as well (readers of Michael’s blog searching for these three phrases, welcome!).

Since we already rank in top spots for phrases such as our company name - Pizza SEO, SEO Pizza and other related keywords, it shouldn’t be much harder to rank well for Britney Spears in connection with Pizza SEO.

DISCLAIMER: We do not have ANY other connection with Britney Spears!

The sound samples for songs on music CDs are only available in WMA (Windows Media Audio) music format as well as in RA (Real Audio). Microsoft does not support Linux, so you are unable to listen to Windows Media Audio legally. Real Audio Player v10 is available for Linux systems, however Amazon uses the old unsupported RA codec that the current version of real Audio Player does not support.

Now as a Linux user I do not have any options of trying before buying - Amazon does not take me into account when accessing their website. Linux users are intentionally left before the gates of listening to music samples. Amazon website thus have accessibility issues. This is the email I have received from Amazon after multiple requests to clarify the situation (mostly receiveing the standardized FAQ replies):

Currently, our Music store does not offer a version of titles that are Linux compatible. Please note that, currently Linux users couldn’t able to listen to audio samples on Real Player/Windows Media Player.

We do offer many software titles for Linux systems. Just click the “Browse Categories” link at the top of our Software home page (http://www.amazon.com/software). On the next page, click on the “Linux” link to browse through our list of titles that will work on your system.

We are aware of this issue, and our developers are working on a resolution. At this time, they have not provided us with an exact date when they expect this situation to be resolved, but I would suggest you check back in the near future. We appreciate your patience in this matter.

Thanks for visiting Amazon.com

Keepon dancing robot created by Hideki Kozima and programmed by Marek Michalowski reacts to the music vibrations (or sound vibrations to be correct). The viral video of Keepon robot on YouTube has been getting millions of viewers lately. Keepon is accompanied by its inventor and dances to Spoon’s Don’t You Evah song. Great idea to popularize both Keepon robot and the “Don’t You Evah” song by Spoon.

Watch Keepon dancing for yourself:

Recently I searched for GAIM in Google using the keyword “gaim”. Usually, the GAIM website comes up as the first result, but not this time. Something had changed, the first result did not contain theword “GAIM” anywhere in the title. So, my eyes moved below, making their way to find a GAIM link. After checking a few more results, it became clear to me that I would have to search more. I used “gaim download” now , but large software download website was the first result. I wanted to visit the original GAIM website.

I tried searching for “gaim” only again. I checked the first result thoroughly this time. Andjust as well that I did - the title was “News - Pidgin”, but then in the description Google displayed the following text: “AOL naturally complained, and Mark Spencer changed the name to ‘Gaim’. … Adium uses libgaim, the backend to gaim and gaim-text behind a slick OS X native …“. The four mentions of “gaim” caught my eyes. I clicked the link and finally was transferred to the official GAIM website, same logo with new name Pidgin.

Although, no message describes why GAIM has been renamed to Pidgin on the homepage (except the long text deep down titled “Important and Long Delayed News”), the logo and the unchanged look of the website reassured me that the website is that of GAIM. Now to the SEO and usability aspects of changing product names.

From a SEO point of view, GAIM developers mastered the basics and used proper redirection via 301 Moved Permanently status from the old gaim.sourceforge.net website to the new www.pidgin.im. Google and other search engines (and their robots) picked up the redirection soon and changed the webpages in their indexes.

From a usability point of view, however, the GAIM (now Pidgin) developers made a mistake - they did not keep the original name “GAIM” anywhere in the title, nor had they informed visitors of the name change via some large text placed near the logo on all pages.

Be sure to use common sense when chaning names or using redirection from an old to a new domain - don’t forget to inform your returning visitors (and returning customers) about the name changes. Explain, why you moved the website from one domain to the other. Use large bold text displayed in a highly visible position (above the fold and close to the logo) informing about the change. Think as your visitors do and allow them to digest the change.

What is Pidgin? “A pidgin is a simplified language that develops as a means of communication between two or more groups who do not share a common language, in situations such as trade. Pidgins usually have no native speakers, but are learned as second languages, and they usually have low prestige with respect to other languages.” (Wikipedia)

Orbitz

Starting at 4:35 pm I log in to Orbitz.com to find the best ticket for my trip. Searching brings up the price of USD 831. OK, I click on that. In next step, however, Orbitz says that it’s my lucky day and they have even better price for me. It’s my lucky day, indeed - “better” price is now only USD 1321. Probably some problem, so I click back to get the results and click on the USD 831 flight again. No, the problem is not between the computer and the chair. It’s on the Orbitz side, I am shown USD 1321 price again.

Struggling some more, I find and get the price of USD 852 finally. Happy with the price, I fill in all the required personal details on two pages and continue to the third one filling in the credit card details. But now, I am displayed only few countries for credit card address and Slovakia is not there.

Calling customer support and they confirm that you can only book through Orbitz if from USA, UK, Canada and approximately five other countries. Welcome to the internet, boy!

Travelocity

Moving to Travelocity.com to book my flight, but they just redirect to Expedia. Bad luck, I’ll try Expedia later on.

Opodo

Trying this British flight tickets website after all, Opodo.co.uk offers good prices, but fails at the same stage as Orbitz - not providing an option to book the flight if you are from outside of UK, US and few other countries.

Expedia

Expedia features a nice interface, displays good flights, although is clearly missing some data - they don’t have the flights that Orbitz offered. Unfortunately, the flights they offer are not the best option for me all the way to the US. So, I leave for low-cost SkyEurope to book first part of my flights from Bratislava to Amsterdam. I then return to Expedia to finish my flights reservations from Amsterdam to US. It’s 8:05 pm.

Usability Above All… or is it?

I will summarize that usability thing for you: Orbitz and Opodo allow you to book if you are from US, UK, Canada, Mexico, Germany, Belgium and maybe France. Travelocity just redirects to Expedia. Expedia does not have access to all flight data and provides only part of the flights available. I have spent more than four hours tryring to book my flights, but all of the websites (except of Expedia partly) have miserably failed to sell (or convert me to customer).

Most of them failed because they don’t recognize internet as a global marketplace. Instead they sell only to small groups of Chosen Ones.

Tell me, what’s your favorite flight booking website then?

It’s really easy on April 5. Why this exact day? Because April 5 is the annual CSS Naked Day organized by Dustin Diaz.

Dustin gives you a chance to earn free PageRank 7 backlink, if you make your website naked. By “naked” he means unlinking any CSS styles from your pages, so that pages are served as clean unstyled content.

Jill Whalen of HighRankings.com provided a list intended to bust myths about factors affecting search engine rankings in her latest e-mail newsletter. The list covers a number of widespread beliefs on things that supposedly hurt rankings but do not, in Jill’s view, make any real difference.

Some of the “myths” are simplifications arising from legitimate claims: e.g., the claim that large percentage of code should change between pages to avoid them being seen as duplicate content by search engines arises from a legitimate, albeit distorted belief that duplicate content across pages will be penalised.

Several alleged myths concern frequent coding conventions. These procedures have become broadly accepted best practices, while their original justification has been often forgotten: eliminating comments from code has no effect or only tangential, moving CSS and scripts to external files may help spiderability but generally should not affect rankings.

With one myth, however, I believe Jill has gone overboard. Her last debunked myth is that The site needs to be browser-compatible and screen-resolution-compatible. She writes: “This is another thing that’s nice to do for your site visitors, but it has no bearing on search engine rankings or relevance.”

Although seemingly primarily of concern with respect to usability and accessibility, I would argue that this factor is also a part of a prudent, robust and sustainable SEO strategy.

Specifically, there has long been speculation in the SEO community (backed by some evidence from patent filings) that “stickiness” - most simply defined as the length of time a visitor spends on a webpage, is an easily exploitable factor in determining how well a set of results served in response to a query meets the user’s requirements.

If Google serves up 10 results and detects (using whatever specific technological feature) that users return after a short visit from some of the results, this provides clear indication that the result has not satisfied them. This stickiness is certainly strongly influenced by this “mythical” factor busted by Jill: potential accessibility and usability problems caused for a subset of users by the site NOT being browser-compatible and screen-resolution-compatible will reduce the time spent on site by some visitors, indicating to the search engine that the result is less satisfactory than it would be without these problems.

Does Google take “stickiness” into account? A SEO blogger’s experiment’s results published in June 2007 point in that direction - a page’s ranking was improved by playing around with stickiness. SEOmoz describes a patent application, that would inter alia allow Google to measure aggregate user behavior, including the “amount of time one or more users spend accessing the document”. In forum discussions at WebmasterWorld.com a valid point has been made by numerous members: it is hard to know whether and to what extent Google (and of course, other search engines) use this metric but it is clear both that
1. there are ways to obtain the metric easily, using existing technology such as the Google Toolbar and other forms of tracking,
2. the metric provides some ranking-relevant information, which can be used to improve SERPs.

With this in mind, I would argue that robust and sustainable SEO-aware design will take this into account even given the possibility the metric is not in use today. A focus on screen resolution and browser compatibility may well have an effect on ranking now, later or both.

UPDATE: SERoundTable references a fresh discussion on WebmasterWorld stressing that the use of user behavior data is likely to increase with the advent of personalized search.

We have found the ultimate tool supporting collaboration - a wiki. DokuWiki in particular. DokuWiki is an online collaboration tool allowing anyone to add, edit (and delete) any information required.

We have been using DokuWiki as the internal knowledge base or pool of information instantly available for everyone in Pizza SEO.

What was that page that provided us with that PageRank 6 link? Do you remember the new website you have used for keyword research last time? How do you improve the clickthrough rate for Google Adwords campaign?

Since we got DokuWiki installed, our search for the answers (or the best solution) has ended. Now, we just open Pizza SEO DokuWiki in our browsers and find the exact answers to our questions. And, if it’s not there, just click on “Edit this page” button and add the new information you have found.

DokuWiki requires a server with PHP installed. We recommend using Apache server and having Apache’s mod_rewrite module enabled for nice, clean URLs.

PS: The best thing is that DokuWiki supports UTF-8 encoding and we are able to use our first language (Slovak) that uses the accented characters. Many other PHP wiki applications have pitifully failed this requirement.