Bad news for good writing

Rules, even golden rules, are made to be broken. Jakob Nielsen’s latest Alertbox article puts forward a remarkably strong case for prioritising the first two or three words in any headline or paragraph. Usability studies show that the eye scans down the left edge of any text; so you need to get the most relevant keywords into the opening of your sentences. And if that means breaking the golden rule of web writing – active voice, rather than passive – then so be it. He advises:

Words are usually the main moneymakers on a website. Selecting the first 2 words for your page titles is probably the highest-impact ROI-boosting design decision you make in a Web project. Front-loading important keywords trumps most other design considerations.

Writing the first 2 words of summaries runs a close second. Here, too, you might want to succumb to passive voice if it lets you pull key terms into the lead.

I’m used to SEO advice clashing with good writing guidelines, but this is the first time I’ve seen it happen with actual usability.

The death of cutesy copywriting

I’m inclined to agree with Jakob Nielsen’s latest column, which puts another nail in the coffin of cutesy copywriting. The US Census Bureau has a big, javascript-driven Population Clock at the top of its homepage. It shows a current estimate of the US population in 18pt bold red digits. It even moves, for goodness sake. Yet a staggering 86% of users missed it. Or rather, they tended to see it, but didn’t look at it. Nielsen’s recommendation in this case: avoid ‘made-up terms or branded descriptions’. If it’s the current population of the US, label it ‘current population of the US’.

Nine, ten, 11, 12… Nielsen on numbers

It seems to be a UK standard that you write the numbers from one to ten as words, but anything from 11 upwards should be written as figures. I’ve had drummed into me in several workplaces, and I’ve forced it on others; it’s also in the Economist and Times style guides. (The Guardian says numbers begin at 10… typically awkward.) But Jakob Nielsen makes a good case for a different approach, based on semantics and skim-reading.

In essence, he favours using figures in any situation where you’re referring to an exact number, including single figures; but to say ‘hundreds’ or ‘thousands’ as words, because you’re using those words as vague descriptions of a quantity, rather than a precise number. It’s hard to argue with the logic in this; but I wonder if I’m too ingrained in the old way to change now?

Future of press releases?

For a few years now, Steve Rubel has been something of a guru in terms of public relations and the web, and his blog is required reading. He and colleagues from major PR agency Edelman have just unleashed something called StoryCrafter, which:

basically breaks down a press release into its core parts, leaving it up to you – the journalist (citizen or pro) – to decide how it should be put together. Most importantly, every press release gets feeds, tags, del.icio.us/digg buttons, trackbacks and comments.

There’s an example here. The initial comments aren’t very encouraging, but I think they’re missing the point. This isn’t about a new web tool for creating press notices. It’s about a new way – rather, a new way of presenting them. And good on them for that. In my journalism days, I spent too long deconstructing someone else’s prose, before turning it into my own. The bullet-point approach will make PR people feel less creative, but it’s what their customer ultimately wants. It’s still rough round the edges, but this feels right.

Interesting to see the comments facility; but this should be a perfect situation for trackbacks (effectively ‘other sites linking to this one’) to come into their own.

Google owns the internet (continued)

Further evidence of the power of search engines – and why you should be concentrating on search engine optimisation any time you build a site, or write an article. Heather at Hitwise reveals that Google now has a 78% market share in UK searching… and that figure is rising, up 9% on last year. (Yahoo, MSN and Ask don’t make it beyond a single digit.)

But even more interesting, she notes that 35-40% of all traffic to sites classified by Hitwise as ‘Travel Agencies, Appliances & Electronics and Insurance’ – all big players when it comes to e-commerce – comes from search engines. And yes, the trend here is upwards too. ‘In the week to 21st October 2006,’ she writes, ‘Travel Agencies received 39% of their upstream visits from search engines, up 7% in the past six months. For such an established category, that is a huge gain.’

In other words – people are becoming ever more reliant on search engines, and more specifically on Google, to get what they want from the internet. If you’re running a website, and you aren’t giving serious thought to your placement on Google, and how you can improve it – consider yourself warned.

Microsoft reveals search term data

Not sure how much to trust it, given the lengthy disclaimer (‘anecdotal’, ‘proof of concept’, etc)… but it’s good to see Microsoft offering hard numbers on the keywords used in searches. Not quite as great as the equivalent from Yahoo, which gives precise numbers for each keyword – and for each combination using that keyword. Still my favourite data source out there. Still no sign of Google sharing its numbers.

If you do any serious writing for the web, especially in a commercial role, you should be checking these two sites. Use the same words in your headlines and page titles as people use in their searches. This is how to buy your way up the Google rankings.

IPSV taxonomy: ‘a waste of time’

Knowledge management expert Steve Dale reports that ‘the debate on the (ir)relevance of IPSV (Integrated Public Sector Vocabulary) has burst out into the open again.’ He muses:

The schema is uneccessarily detailed/over-complicated, and these days most search engines do not rely on subject metadata alone (or even at all) to classify or organise results. Who will be brave enough to admit that the huge intellectual investment put into developing and implementing IPSV has been largely a waste of time?!

Er, I will.

Steve is absolutely right: it’s far too big to be practically useful, with over 3000 primary terms (that’s before you get into synonyms)… and it’s getting bigger, with an extra 350 primary terms added as part of the April 2006 update. And that’s before several major government departments get round to fleshing out the branches which should notionally be their domain. Which, of course, means the production of further revisions… which, inevitably, means everyone has to go back and review the subject tagging they did against the previous versions. And so it goes on.

I just don’t believe big taxonomies can ever work. I think there’s a reason Google (free text) deposed Yahoo (big classification structure) as the web’s #1 search engine. And if you’re writing decent web content anyway, you’ll have all the important keywords in your important fields – like the page title, the H1 heading, and so on. Just your standard Search Engine Optimisation tactics, which you should be doing anyway. Amount of additional effort required: zero.

But I do think a smaller-scale subject tree can be useful. I’ve recently led an exercise to produce a mini-taxonomy. We set very tight limits: a maximum 100 terms, ideally two levels, but three at a push. And I’m very happy with the structure we produced. It prints nicely on a single sheet of A4; people could keep it by their PCs, and refer to it as necessary. (Mind you, with only 100 terms, I’d expect people to know it by heart fairly quickly.) I’d much rather have a guarantee of a ‘near enough’ match, than a situation where exact matches are dependent on people being bothered to tag exactly.

But I’m quite relaxed about it, really. Yes, it’s mandatory… but, theoretically, so was IPSV’s predecessor, GCL. I don’t remember too many people actually implementing it properly either. And I certainly don’t remember the Taxonomy Police rounding people up.

PS: Note to Steve… nice blog, nice photo, but you need to make it easier for us to tell who you are. It took quite a lot of research to find your name?!

Writing for the web: it’s all about search

Interesting to see Jakob Nielsen’s latest thoughts on the subject of writing for the web:

‘Web users are growing ever-more search dominant. Search is how people discover new websites and find individual pages within websites and intranets. Unless you’re listed on the first search engine results page (SERP), you might as well not exist. So, the first duty of writing for the Web is to write to be found.‘

I’ve written and delivered a few ‘writing for the web’ courses in my time; my more recent attempts have featured a chunk about search engines and SEO. I wonder how many ‘writing’ trainers include that? It’s a fine example of technical and creative skills meeting head-on.

Naturally, the inverted pyramid thing still applies… but I’ve always considered that to be basic ‘good writing’ policy anyway, regardless of the medium. It’s the concept of the search engine which marks ‘writing for the web’ out as being a new and different discipline.

Just as an aside… I notice that Nielsen’s approach to the URLs of his columns changed about a year ago, from a yyyymmdd.html approach to a ‘pretty URL’ based on keywords (chosen manually, I suspect). It’s a reminder that the Address Bar is another area for potential keyword loading. Use it wisely. (The WordPress blogging tool is especially good at this.)

Search intelligence: the web’s best kept secret?

The new Google Trends site is interesting enough, but I still wonder why Google doesn’t just take it to its natural conclusion, and let us see the actual numbers.

Thank heavens for Yahoo / Overture, who tell you precisely how many people searched for a given keyword (and all derivatives) in a given month… if you know where to look. The link used to be fairly prominent before Yahoo bought Overture… but it isn’t now.

With this kind of hard statistical evidence at your fingertips, there’s really no excuse not to research the popular keywords in your area of operation; and to target your efforts accordingly. And not just on the web, either.

Look at Ryanair, who dropped their slogan ‘the low fares airline’ in favour of ‘fly cheaper’. Why would they do that? Compare the Overture data for the numbers of people wanting ‘low airline fares’ with those searching for ‘cheap flights’. Suddenly it’s a no-brainer, especially when your industry is dominated by e-business.

Reading in an F-shape

Jakob Nielsen's latest Alertbox suggests that people read web pages in a shape resembling a letter F: one horizontal line across the top, a second horizontal line lower down, then a vertical scan up (or down) the left edge.

Two of Jakob's three conclusions are just your basic 'inverted pyramid' thinking: 'Users won't read your text thoroughly' and 'The first two paragraphs must state the most important information'. But the third is new and a bit more challenging: that you need to put the important keywords at the start of subheading, paragraphs and bullet points.