Blog

As a recap, the table below shows current 2008 browser standings:

Top line features include:

• Open source code base
• Based on WebKit used by Safari
• Powered by Google Gears
• Multi process design
• Built in task manager
• V8 virtual machine built from the ground up
• Redesigned user experience based on 2008 web browsing habits
• More robust error management
• Safer browsing through process sandboxing

So what will the above features mean for the rest of us? Well, reading Google’s cartoon pitch we should be seeing some impressive results including:

• Faster browsing
• Quicker JavasScript performance, so quicker web apps (i.e. gmail, netvibes, etc.)
• Less chances of the browser crashing because of unresponsive applications
• The ability to close an unresponsive web page using a task manager just like on your computer
• Privacy mode to clear your browsing history as you surf
• Better pop-up manager, which can contain pop-ups within individual page tabs
• Safer browsing by using sandboxing techniques as found in operating systems

We haven’t tried it yet, and will update you once we’ve tested it thoroughly. At first glance Google Chrome sounds like a revolution in the browser space, and it comes just after Google renewing their deal with Mozilla to support Firefox.
So far it’s Windows only, but Google are hard at work building versions for Mac and Linux.
So what are the implications of Google releasing a web browser?

Apart from extending its reach beyond the search engine, advertising, video, social, knowledge, medical, financial, and technical space, Google Chrome sounds like an ideal complement to Google’s suite of analytics and tracking solutions.
Now for the Big Brother thing and conspiracy theories… If Google were to track users’ web visiting habits, bounce rates, visit times, and interactions with websites from within its own browser, than it would be able to provide the ultimate model of web behaviour, and base its algorithms and advertising solutions on this data. It would also be able to provide granular information on trends, and insight based on very accurate metrics.
Another aspect is the integration of the Omnibox... Similar to Google Suggest, it replaces the URL bar, which will undoubtedly blur the difference between typing in URLs and searching which will undoubtedly affect keyword bidding on the likes of Google Adwords.

So we look forward to reviewing Google Chrome, and how it will affect our clients.

We’ll update you with our feedback soon!

 "Whether or not you are a fan of Bill Gates, it is impossible to deny the impact he has made on the spread of computer technology across the planet during the past three decades."

Five lessons we have, without a doubt, learned from Gates....

1. Software matters
2. Arrogance breeds failure
3. Computing will spread everywhere
4. You don’t have to be first to win
5. Geeks can be businessmen, too

A more detailed analysis and the original blog post about these lessons can be found at http://blogs.techrepublic.com.com/hiner/?p=761&tag=nl.e101

Adobe's FAQ about this is here. Hat tip for CL for finding this.

To quote:

"Adobe is working with Google and Yahoo! to enable one of the largest fundamental improvements in web search results by making the Flash file format (SWF) a first-class citizen in searchable web content. This will increase the accuracy of web search results by enabling top search engines to understand what's inside of RIAs and other rich web content created with Adobe Flash technology and add that relevance back to the HTML page."

OK...so, where's Microsoft and Ask in this? Ask don't have a lot of search volume, but MS are definitely trying to grow theirs. Do I smell a conspiracy - have Adobe decided to not invite MS to this club, because of Silverlight, Redmond's attempt to compete with Adobe Flash? Not sure the regulators would like that...

Also - will this really add a lot of value to web search results, at least to begin with? Most Flash files don't contain that much information to aid results quality, and designers aren't the best people to be adding this sort of textual information to help with SEO - will we see amateur attempts at old fashioned keyword-stuffing?

I'm glad our design team, Open, work closely with our Steak SEO team, and think about SEO, usability and good design at all times - and remember that sites exists to generate a return, not just look good. We like things that look good, though!

I'd like to think this a major step forward, but it doesn't address a more important issue: designers and agencies who build sites in 100% Flash, and expect SEO teams to fit a layer on top to give the site rankings. That's still bad practice. If this was an announcement of image recognition technology, now that would excite me.

IMO a good site design will use Flash in moderation when it offers functionality or a visual experience that no other technology can offer. Not because Flash is the easy way out or their only skill. So as Google say "Try to use Flash only where you need to".

Updated: typos 

Your website is the core of your online presence. If it is not converting visitors at the maximum rate possible, it is holding your business back. It's that simple.

If it's not working as hard as possible, your ability to buy the most competitive keywords or display advertising placements simply doesn't exist. As surely as having your marketing budgets cuts, you are locked out of these opportunities.

Your competitors however...need I say more?

So:

• Make sure you can track results from click to sale - and report on the drop-off points in-between to the site can be improved
• Test your landing pages and the copy of ads that link to them using paid search (to control spend and allow for fast changes)
• Test page copy and form design / shopping carts, too, where possible
• Test, refine and test again until you are 100% sure the page(s) are working as hard as possible

Now you may be thinking "that's fine in theory" but I don't have the time or resources - or nobody will understand internally, anyway, and my efforts will go unrecognised.

That's going to change. Here's why: Google.

Google offer a free landing page testing solution, Google Optimiser. It's not the most in-depth on the market in terms of features - but it is a good starting point and it is free.

Google know that if websites work harder, advertisers can spend more and the audience they send to websites should be better served, adding to their loyalty to Google.

Expect to see a lot more press coverage of Google Optimiser - including with a joint case study with ourselves. Use the contact form here to requests a copy of this when it is available.

In the meantime, other case studies can be found here at the Google site.

Updated: typo corrected.

This is a simple but effective way to make it easier for consumers to order across brands and sites - and no doubt raise sales. Found via oldnavy.com.

 

Cookie Length

Many tracking tools set cookies for 90 days. This is crucial for verticals with long consideration periods before purchase. Travel is the prime example - consumers might research a holiday destination over the Christmas period, but not book until much later in the year. Car insurance is another example, where consumers research ahead of their renewal date.

However some sites set them for a short period - one week, for example. This may be because they do not want to pay affiliates for repeat visitors making repeat purchases (for the affiliate angle on this, see this blog post). Or it may be the site owner thinks consumers normally buy within 7 days and if they return after that period they can be viewed as a "new" visitor motivated by brand loyalty or other adverts. Ideally different traffic sources should have different cookie lengths set.

Personally I think 7 days is too short in most situations - what if the consumer needs to wait after payday and bookmarks the site until then, or are researching a gift for a relative and have a longer consideration time to buy? 14 or 30 days seems more appropriate.

Repeat conversion window and view conversion window are two other important settings to consider with ad tracking tools.

The former ensures that if a consumer clicks twice on the buy button - e.g. because their connection is slow - conversions aren't double-counted when in fact there is only one sale. An example setting is one minute.

View conversion window determines the time frame in which conversions are recorded after a view of a banner; so if this is set to 14 days and a consumer visits the site without viewing/clicking any other ad link and completes a tracked action within 14 days, then that banner receives the credit for the conversion. This opens up a discussion on first click vs last click conversion tracking, but I'll save that for another day...

Advanced Reporting and Cookies

There's another problem with setting short cookie lengths - one which may grow. As tracking tools provide more visibility on the overall path of a consumer - which banner ads they saw/clicked on, which PPC ads they clicked on, which natural search links they clicked on, which emails they viewed/clicked on, and which affiliate links they clicked on - so the cookie length becomes more important. Consumers will typically be exposed to these different ads and their tracking links over a longer period of time than if the tracking tool and cookie only recorded data from one type of campaign. If the cookie length is too short, E2C (Dart) and Engagement Mapping (Atlas, now MS Advertiser and Publisher Solutions) will be next to useless.

Ideally, reports should use long-length cookies and record first visit, second visit etc so site owners know their volume of repeat visitors as well as truly new ones, the buy rates of both, and the ads they saw/clicked on along the way - providing enhanced visibility, leading to more informed buying decisions.

Analytics vs Tracking

The above "ideal" ad reporting blurs the line between analytics packages and tracking solutions. To my mind, the two can be defined as:

Ad Tracking (Atlas, Dart etc) - these solutions track adverts that lead visitors to the website, and may then record if they complete a defined action (like a sale or email subscription). They only work for sources of traffic where the URL can be controlled to use a redirect URL or have a code appended to the end, resulting in the visit source being recorded and a cookie dropped during redirect or by the destination site.

Analytics (Google Analytics, Webtrends etc) - these packages record all site traffic, can determine its source of URLs are tagged up accordingly, and record site activity - not just sales or similar actions, but also page views, visit lengths etc.

All too often, these two sets of tools create silos of overlapping, sometimes contradictory data. We often have clients ask why tracking tools produce different numbers to their analytics software. Cookie length is one factor - followed by different counting methods, de-duplication, JavaScript availability, browser security settings, conversion window settings, 1st vs 3rd party cookies etc.

So, cookie length is important - and needs to be seen in the wider context of the convergence of analytics and tracking solutions (an end to data silos!) alongisde the mergence of advanced reports. Especially now that Google and Microsoft both own ad tracking and serving companies, search engines, display ad networks and analytics packages...I think report convergene may be a hot topic in the later half of 2008 and into 2009.

Good for them I say - the more users are warned about security risks, the better. To quote the Computing.co.uk article:

In a white paper, PayPal said that "letting users view the PayPal site on one of these browsers is equal to a car manufacturer allowing drivers to buy one of their vehicles without seatbelts".

Site visitors using old browsers may just seem like an oddity of the web. However, as PayPal have noted, they can be a threat to security - they have bugs that may leak the user's personal data (including passwords and credit card info).

I started to ask myself why users still have old browsers like IE3 and IE4 - and thought of all those old PCs donated to Africa. Do they still have old browsers, because newer versions are too resource-intensive? Probably. What will happen to these users if this becomes a common decision by security-conscious sites? Are there poorly-funded libraries with old PCs still using IE4 (a colleague just told me about one using IE 5.5 in Central London)?

I still feel that blocking old browsers is the right thing to do on sites requiring a high level of security (e.g banks) - but attention needs to be paid to older computers used in other countries, too.

The web has come a long way since I first started working in online (back then we all called it New Media, capital letters and all).

As an experienced industry expert (read: cynic) I sometimes feel the web has moved away from its grassroots days, when anything felt possible and communities of interest would spring up on Geocities, Yahoo Groups, or newsgroups.

Yesterday I used a service that reminded me of the early days of the web in all the best ways. It was mentioned by a colleague in passing, I found it via a quick search, it felt a bit rough around the edges, but the idea was simple and it worked.

I think there is an interesting point here. Can websites look too polished? Can they look too Web 2.0?

I think so. I've worked with a client in the past who targeted customers in a low demographic. They had previously launched a new site. It looked great - far removed from the late 1990s throw back site they had (tables, animated GIFs, no CSS on the horizon...)

I think you can guess what comes next...their conversions bombed. Through the floor. The site looked too slick, too professional, too good - it didn't say to their potential customers: "We are cheap enough for you". It said: "We're out of your price range".

A quick roll back to the old site, and conversion levels began to return to normal. When they next launched a site working with us, we encouraged them to gather feedback from their customer profile before going live, and carried out split testing.

So...here's to websites that aren't 100% slick and are rough around the edges, here's to new ideas that are more important than the latest Flash affect or Ajax...here's to the pioneering days of the web.

PS the service that inspired this post was Freecycle.

Following up on my posts about IE8 launching in beta and browser usage stats for the UK, I decided to look at the screen resolutions people were using today on the web.

The best source I found was Onestat.com (an old favourite for analytics stats) from April 2007:

• 1024 x 768 dominates - no surprise given the number of TFTs - 55.34%
• 1280 x 1024 - again TFT variation -17.23%
• 1280 x 800 - 8.23%
• 800 x 600 - old CRTs (also known as bricks!) are disappearing as IT departments and home users replace them 8.18%
• 1152 x 864 - again, squarish TFTs - 3.67%

Hopefully they will publish 2007-2008 stats soon - if anybody has a more up to date source, please leave a comment / use the General option on the contact form (link at top of this page), starting FAO Duncan.

Browser statistics can, like all stats, be a bit of a minefield. On the surface the web offers the easiest opportunity to collect anonymous usage data.

However...it's just not that simple (you saw that coming, didn't you?).

The user agent - the bit of a browser that identifies it's make and model - can be "spoofed" to make the browser to appear to be a different version or browser entirely (often by less-than-savoury visitors like email harvesters). Add to this that browser data from any one site can be skewed by the target audience - for example technical sites have a much higher number of visitors who are not using IE compared to most sites, because IT and Web Dev staff are more likely to install FireFox or use Macs with Apple's browser, Safari.

Not so easy, is it?

So...how to get a reasonable idea of browser usage in the UK?

One way is to look at aggregated statistics from multiple website. Onestat publish these at intervals every year in the PR section of their website; they gather them from 75,000 sites globally. The latest stats:

United Kingdom - February 2008 (brackets show % change since June 2007)

• Internet Explorer 86.15%  - increased by 0.15%
• Mozilla Firefox 11.52% - increased by 0.30%
• Apple Safari 1.59% - down 0.02%
• Opera 0.52% - down 0.01%
• Netscape 0.12% - increased by 0.02% (NB: Netscape has now been killed-off by AOL)

So...IE and FireFox are the obvious winners, but what about different versions of each browser?

Browser Version Stats from Wikipedia (for 2007)

• IE7 at 45.5% in 2007
• IE6 at 32.64%
• FireFox Version 2 has 16.30%
• FireFox 3 (beta) 0.18%

Internet Exporer entry; FireFox (Netapplications.com stats).

So whilst IE7 is replacing IE6, it is still going strong - and Firefox 3, from these stats at least, is still in the minority.

The conclusion: no surprise- designing for IE6, IE7 and both FireFox versions, as well as Safari, remains important to provide cross-platform and browser compatibility.

Oh, and there's Opera too...and for some sites, mobile access is becoming more common...

 A developer's work is never done!