Some things you learn the hard way.

As a web developer with nearly ten years experience – quite a lot in such an immature industry – I have made a number of mistakes, and learnt from other people’s.  I’d like to share the most common of those mistakes with you now, because if your site still makes them, you are handicapping yourself unnecessarily.

Splash screens

These are those screens, often animated, which you see before you get to the main website.  Somewhere between 10 and 60 seconds of wiggly lines; words like ‘experience’, ‘confidence’ and ‘specialist’ (why not add ‘onanist’?) fading in and out; quite possibly a drum n’ bass soundtrack. 

If you are not a student nightclub promoter, you have no excuse for annoying your visitors with this unnecessary garbage.  Tesco don’t force their customers to sit through a 20 second video about how they are ‘established market leaders’ when all they popped in for was a pint of milk and a flat screen telly, so neither should you.  Plus, it plays havoc with your search engine ranking.

Drop-down navigation

Sites which have many pages often resort to drop-down navigation, which is where you hover your pointer over the buttons, and more options appear below.  If your fine motor control is not brilliant, you may find this sort of navigation frustrating or even impossible to use, even is if it is implemented perfectly.  Most of the time the coding is far from perfect, so even users with perfect hand-eye coordination can struggle.

Add in the fact that other page elements (like Flash or video) can cover the dropped-down navigation, making it impossible to click on, and you would need a very good reason indeed to insist on drop-down navigation.

Internal scrolling

Fortunately this is less common now, but in the past people often had the idea that they wanted their site to fit inside a single screen, with no scrolling.  This meant that the pages couldn’t be too tall, which was a serious restriction on the amount that could be displayed.  When the text was too big to fit on the page they added scrollbars to the block of text alone,  not the page as a whole.

This is unpleasant for many reasons.  First of all it is counterproductive; if people don’t like scrolling down the main page, they will really hate scrolling through individual page elements.  Second, if you try to print a page with these internal scrollbars you will probably find that the text gets cut off.  Third, the usual keyboard shortcuts, which are a godsend for people with less-than-perfect fine motor skills, don’t work properly, and finally: web pages are documents.  They are designed to scroll. 

If your site still makes use of any of these techniques, it might be worth talking to your web developer to discuss alternative ways to achieve the same thing.  It needn’t be expensive or time-consuming, but your visitors will love you for it.

This is probably most obvious when you look at how often a website is updated.  Search engines will give more weight to a site which is updated regularly, on the grounds that newer content is more likely to be relevant.  Your potential customers will probably make a similar judgement; a site which contains obviously out-of-date content – such as details of events which happened months ago -  can do you more harm than good.

I appreciate that it is not always easy to keep your site up to date.  On a practical level, you may not find it easy to make changes, or the ability to make changes may be in someone else’s hands.  If you are not in control of your website, then you have no power to influence whether it generates a useful return on investment, which is an uncomfortable position to be in.  If your web developer is worth their fee, they should be able to explain the different ways in which you can maintain your site yourself.

One of the most common ways to keep a site up-to-date is to develop a news database, which lives on your website.  You typically add news using a simple web form, and as part of the process you give the news an expiry date.  That way, the website knows when the news is no longer relevant, and can hide it automatically once the date has passed.  You never have to worry about obsolete content again.

Another benefit of a news database is that one simple form submission can automatically trigger changes on many pages; for example the front page could be updated, listing the headlines of the top 5 news items, as would the news list.  There would also be a brand new page for each news item, which is full of relevant text.  All of this is exactly what a search engine likes to see.

Finally, if you categorise your news, you could show different news on different pages, which is hugely useful for your human visitors.  As they browse your site, they are always being offered news which is recent and relevant.  If you look at the website of Lynx Networks (lynxnetworks.co.uk), you will see that each service and sector has its own news, so that as you visit the pages which are most relevant to you, you are always being offered news and case studies which tie in with your interests.

By making really intelligent use of a news database, you will enhance your site for both visitors and search engines, which makes your chances of getting a return on investment that much greater.

The key word to keep in your mind when discussing search engines is relevance. The only thing the humble search engine wants to do is return relevant results.  With the web being as large as it is (tens of billions of pages at least) this is a job that can only be done by a computer, which is a shame, because computers are about as good at analysing intangible concepts such as relevance as they are at windsurfing.  What computers really like to do – pretty much all they can do, in fact – is count stuff.  This means that they have to use fairly crude methods, such as counting the number of times a search phrase appears in your text, in order to make an educated guess as to how relevant it is.

So, the more often the search phrase appears, the better?  Well, only up to a point.  Search engines are smart enough to realise that bigger is not always better, so an unrealistic phrase count will lead to your site being penalised.  They also care where in the page the key words appear, so constructing your web page properly is essential.  This is where a web designer who really understands HTML is going to prove their worth.

What else can computers count?  Well, Google pioneered the concept of counting the number of inbound links a page has, and storing this as a value called ‘PageRank’ (PR).  Each link into you counts as a vote in favour of your site, and the more links in, the greater your PR, and therefore search engine position, will be.  Even better, a link from a site with high PR carries more weight than a link from a low PR site.  Large sites like the BBC and Yahoo tend to have a high PR, so a link from any of these will send your search engine position soaring.

Anything else that computers can count?  The number of days since your content was updated, for one.  Search engine ‘bots’ will visit your site regularly, and what they really want to see is fresh content.  It stands to reason that content which was written last week is more likely to be relevant than content which was written months,  or even years ago (you know who you are!).

Next time, I’ll talk you through some simple techniques to improve your score on each of these metrics.

However, it is quite possible that a decision you made weeks, months or even years ago will have a big impact on the success of your website, and certainly on the speed of its growth.  Choosing the right domain name (the bit that comes after ‘www’, so prominentmedia.com or google.co.uk for example) is incredibly important.  As a professional web developer, I am only occasionally asked for to advise on domain name selection, which is a shame because in many cases I could help my clients pick a domain name which makes growing their website even easier.

Here are a few handy hints (with notable exceptions):

Stick to the main TLDs where possible

The top level domain (TLD) is the final part of the domain name (‘.com’, ‘.uk’, ‘.net’ and so on), and is usually country-specific – in Germany they have ‘.de’, in France they have ‘.fr’.  The domain name system was invented in the USA, and as a result they don’t have their own geographic TLD; .com, .net and .org are theoretically intended to be US-only.  In reality everyone should register their chosen domain name as both ‘.com’ and ‘.co.uk’ if possible; it makes it much more likely that people will type it correctly.

Although there are a large number of TLDs available, at present I would advise you to stick to ‘.com’ and ‘.co.uk’ for businesses, and ‘.org’ and ‘.org.uk’ for voluntary organisations.  In my experience people simply don’t remember the less common ones such as ‘.biz’, ‘.eu’ and ‘.info’.

Exception:  Are you a TV company?  If so, then you may be able to join gm.tv and five.tv in making some use of the ‘.tv’ TLD, intended originally for the island nation of Tuvalu. 

Your domain name should ideally be short, memorable and easy to spell

Internet lore has it that all common single English words have been registered already, in .com form at least, so accountants.com is long gone.  All is not lost, however, because there are plenty of two-word combinations available, and there will be for some time to come.  Try to avoid hyphens if possible, because when you are telling people your domain name the convention is to run all the words together, and it just takes too long to say “Milton hyphen Keynes hyphen Plumbers”.

Exception:  Tech advice site Experts Exchange were sensible to register experts-exchange.com.

Your domain name should contain your key search phrases

Search engines know that you have to pay for domain names, so they give some weight to any terms found within them.  For an accountancy firm in Cambridge, then, the domain cambridgeaccountants.co.uk would be a good one. 

Exception:  If your brand is big enough, your name is all you need.  Amazon, Google and Yahoo! do fine using just their name (and a multimillion dollar marketing budget).

You can have more than one domain name

A common solution to the problem above is to have one domain name for your business stationery – usually your business name – and one which you submit to search engines and build links towards.  In the example above, the firm could register hughesandson.co.uk and cambridgeaccountants.co.uk and have them both point to the same website. 

Beware of having too many domains pointing to the same site, though, because some search engines may take a dim view, and downgrade you if they think that you are trying to unfairly disrupt (or ‘game’) their system.  One good quality ‘keywords’ domain, alongside a ‘business’ domain,  is unlikely to cause any alarm, however. 

Exception:  B&Q do all of their online trading at diy.com.

Iain Row runs Prominent Media, a web design an development company based in Milton Keynes.  They specialise in building websites ‘that work as hard as you do’.

When navigation is not permanently displayed, it makes moving through a site difficult.

One of the keys to making a site easy to use is giving visitors a clear indication of where in the site they are now, have been, and should go to next.  Good navigation will provide all of these things, but it’s impossible to accomplish using menus which only appear when the mouse is over them. 

It can be hidden by embedded elements, making navigation to some pages impossible.

Most dynamically-displaying navigation goes over the top of content which is already visible.  However some content, like embedded video and Flash animations, will not be hidden by the navigation bar, but stay on top, with the result that parts of the navigation will be impossible to reach.

It’s not guaranteed to work in all browsers.

Some browsers have stricter implementations of JavaScript than others, which means that while your code may work perfectly well in IE, you’d better be prepared for some serious testing to ensure that it works equally well in FireFox, Mozilla, Safari, Opera and other minority browsers.  If your navigation is broken then so is your site.

It requires fine motor control and hand-eye co-ordination.

In common with most dynamic menu systems, like Windows’ own Start menu, getting around it requires practice and some skill with a mouse.  If you have ever done a usability study with elderly users you will find that not everyone finds this easy.  Windows does get one thing right; you can use it fairly easily using just the arrow keys – not often the case with JavaScript navigation.

No major site uses it.

Google, Amazon, YouTube, BBC, MSN, Yahoo!, MySpace, eBay, Wikipedia, Craigslist, IMDB; you name them, they don’t use it.

There are better ways to accomplish the same thing.

One good thing about JavaScript is that a single script file can contain the navigation for the entire site.  When you update that file, the whole site’s navigation updates.  However, this is also the case with any server-side scripting technology, outputting standard HTML, with none of the drawbacks listed here.

JavaScript can be easily turned off – then your site is broken! 

Since JavaScript powers many of the things that are most unwelcome on the web, such as pop-ups and windows that refuse to close, some people turn it off (it’s easy to do by accessing the ‘Security’ tab in Internet Options).  In that case, most dynamic navigation stops working, and in some cases the navigation bar disappears altogether.  Statistics from W3Schools show that 10% of the visitors to your site will have JavaScript disabled.  Since there’s no really good reason to use JavaScript for navigation, why even run the risk that your site will be unusable by 10% of your visitors?

Iain Row is the lead developer at Prominent Media Ltd, a web development company based in the UK.  Their SiteNav product is an intelligent website navigation solution.

In previous years, the consensus was that the main function of website navigation was to take visitors from your front page to the area of the site they wanted to visit. However in today’s search-engine-dominated World Wide Web, this model is unlikely to be the most accurate.

Overwhelmingly, visitors go directly to content-laden pages, lower down the navigation hierarchy. This is because these pages are the ones that search engines direct them to. From there, site visitors will make a decision about whether to browse further, contact you directly, or leave the site. If your navigation design makes it difficult for people to move through the site, they will simply leave.

This is the main reason why the use of frames has largely died out, but it also bodes badly for JavaScript driven dynamic navigation, where elements are not visible until the mouse is over them. If your visitors found one of your Web pages from a search engine and didn’t use your home page site navigation, how will they know which area to roll over to get to other useful content? Those hidden elements that guide the user through your site are all but invisible. It’s likely that the user will hit the back button and start another search for content that may be elsewhere on your site.

Breadcrumb navigation can help by letting the visitor know how far they are down a certain path and helping them to come back up, but it provides no clue about what other pages might also be appropriate.

The best solution is to have context-sensitive navigation. When visiting one page, you can see at a glance the other pages in the same section, as well as the other main sections of interest. 

For an example, take a look at www.cranfieldaerospace.com/gsuits.aspx.  This site has more than 150 pages, but even jumping straight into a content page four levels down we can tell immediately where we are, just from the navigation.  We can also tell from the links on the left what other pages are ‘siblings’ of this one, and might be worth a visit, as well as which other sections are available.

A key factor is the use of a different colour to highlight the current page or section.  This is vital for helping the visitor to see at a glance where they are, although it doesn’t need to actually be a different colour; bold or italic text can work just as well, depending on the site’s design.  The key thing is that the path to the current page must be visible.

In the case of Cranfield Aerospace, their entire site’s navigation is run using a piece of .NET code called SiteNav (www.sitenav.co.uk).  This lets the team running the website write a simple XML file which maps the navigation like a family tree; the SiteNav code then reads this file and works out what links are appropriate for each page.  It can use the same data to produce breadcrumb navigation and a site map page, and the key benefit is that you don’t have to be in any way technical to write the XML file or implement SiteNav – two lines of code are all that is required.

You don’t have to use SiteNav, of course – but whatever method you use, try to avoid the false promise of JavaScript and instead implement navigation properly.  If you do, you’ll give your site the best chance of both retaining visitors, and encouraging them to spend as much time exploring your site as possible.

Iain Row is the lead developer at Prominent Media Ltd, a web development company based in the UK.  Their SiteNav product is an intelligent website navigation solution.