Search Engine Musings

There seems to be a lot of confusion in the minds of many webmasters, both new and experienced, about what SEO actually is and what tasks it is useful to spend time on. You see forum questions asking ’should I worry about code validity’ (I’ll return to that one in another posting soon), ’should I put in a better shopping cart system?’, ’should I worry about Firefox/Macs/disabled people?’. You soon get the impression that these webmasters aren’t looking at their web sites with any sort of rounded view.

A web site isn’t there just to get rankings, just to get traffic, just to sell stuff to IE users. You have to look at it as a whole - how do the various parts fit together to attract users, attract genuine links, satisfy the demand for the product, service or information you provide, abide by the law, allow search engines easy access, and a dozen other areas.

It’s not just that SEO and usability go together - everything about a web site should be optimised to make visiting it a quality experience. Navigation should assist users to find what they want. Content should be laid out and structured to be easily scanned while providing maximum information. Images should be sensibly sized and add to the information rather than just acting as generic filler. If you’re selling something then it should be possible to order it with the minimum of fuss and with as many payment options as possible.

There are no shortcuts to quality. Think total web site optimisation and you won’t go far wrong.
To me that’s what true SEO means, but the terminology in our industry is so fluid that it can be taken to mean just about anything and many people see it in much narrower terms. If you’re hiring an SEO company then find out beforehand what they think it means and define how wide their remit should be.

Earlier this week over at AltSearchEngines the title of alternative search engine of the year was awarded to Quintura. Strictly speaking it doesn’t seem to be a search engine in its own right but rather a search interface, and that’s where its value lies.

Quintura is of Russian origin, and it makes use of that design feature beloved of Web 2.0 bloggers - the cloud. But instead of the static and to my mind fairly pointless version we’re used to seeing, it uses the cloud idea in a dynamic way, interfacing it with the search box via some fairly heavyweight artificial intelligence algorithms. It currently uses a Yahoo xml feed for its main web and image search, plus Blinx for video search and there is also an Amazon search facility, but from the description of the associated desktop search it looks as if it can work with others too. They have a Russian language version which works with Yandex results.

Searching by Clouds

Basically you enter an initial search as normal and the usual results list appears in the right half of the screen. On the left (or optionally above if you reposition it from the settings menu) the cloud appears with your initial search terms in bold red in the centre. Around it are clustered some other related terms and if you hover over one of them that word is temporarily added to your base search phrase (you’ll see this happen in the search input box) and the search results update accordingly; as does the cloud. Hover over another word and that one replaces the first addition, while if you click on it the word is added to the search query as a persistent addition so you can then refine further without losing it. Hovering over a word also causes a red x to appear beside it and by clicking the x you can exclude that word in the same way as using a minus sign would in the search box. Double-click in a blank area of the cloud and a text box appears so you can add a word that doesn’t appear in the cloud.

All this takes longer to describe than to do - once you get used to using the cloud you find it’s a fast way of refining results without taking your hand off the mouse.

Because the cloud interface is so intuitive for children there is a separate childrens version and the founders of the company are apparently planning a version specifically aimed at mothers.

Although the cloud interface takes up a fair bit of screen space, given that most search results leave a lot of white space anyway it is less intrusive than you might expect, and all in all this seems like a useful progression in search interface design and it’s one I’ll be investigating further.

I certainly find it more useful than the various attempts at ‘universal search’ that we’ve seen so far - most of those seem to get in the way of seeing normal results by cluttering the space up with other media or somewhat doubtful local results. Quintura’s beauty is that, like all the best ideas, it is essentially simple in operation for the user. At the moment it’s still sporting a Beta tag so I look forward to seeing whatever future developments they can come up with.

As a consequence of the paid link controversy and whether you should use the nofollow attribute on paid outbound links, there is now another one about whether webmasters should use it on internal links. On one side there is the original suggestion from Matt Cutts in his interview with Rand Fishkin of SEOMoz, and followups from such as Dan Thies, while on the other side are people like Michael Martinez who find the idea of trying to manipulate PageRank by this method dubious at best and downright dangerous in most cases.

In the meantime there is also the question of using nofollow on blogs as an anti-spam measure, or not, depending on your perspective.

Nofollow on Blogs

Lets take the easier one first - blogs. This is what the nofollow tag was invented for, to stop people spamming the comments sections of blogs and forums with pointless messages containing embedded links back to their sites. Even then it was a bit controversial and many argued that there were better ways of combatting link spam. The debate has raged on and there is now a “dofollow” movement that advocates getting rid of nofollow tags from blogs (some blog software adds them automatically). Of course you then need other methods of defence against the robot blog-spammers but this can be managed - Askimet does a pretty good job in Wordpress blogs and you can pre-moderate if you have low levels of comments or only allow people who have already had comments approved. I already make links within my posts carry full weight and I’m inclined to go the dofollow route on comments too - just need to decide on the best method of doing it.

Nofollow on internal links

On the thornier topic of using nofollow to manipulate PageRank within a site there are a few arguements that I find persuasive.

Firstly I don’t believe enough people actually understand PageRank enough to start trying to fiddle with it. Anyone who reads the SEO forums will know that they are full of questions which show that people believe the most nonsensical rubbish about the subject and pick up on old wives tales at the slightest opportunity. The sort of mess that these folk could make of their sites with nofollow doesn’t bear thinking about.

Secondly the kind of pages that are being suggested as candidates for downgrading - About Us and Contact Us pages for instance - are actually perfectly useful pages that often can be made to rank well for important terms. The potential gains are far outweighed by the likely losses.

Thirdly we have the problems that would be caused to the usability of sites. Many websites use Google’s own search system to provide site-search facilities, and studies show that many users will use the search boxes to navigate a site. If you close off some of your pages with nofollow then those pages won’t show up in these search results. Why would you want that to happen? Golden rule - build sites to serve your users.

Fourthly there is something very fundamental here which I think needs to be addressed. Google have always said that you should show users and search engines the same things. That’s why hidden text and cloaking is so disliked by them. If you show a user a link then you are telling them that it’s worth following it. If you use a nofollow attribute on it then you are telling the search spiders that it isn’t worth following. Isn’t that rather dishonest? Isn’t that against the very rules that Google want us to adhere to? I think it is, and for that reason as well as the other listed above, I won’t be using it.

I feel a rant coming on… The other day I was preparing an email to a client about why it wasn’t a good idea to have lots of ‘click here’ links all over her site and thought to check a few of the usual places like the W3C and accessability sites to give her as references. My eye was caught by a description line in the search results which took me to a book review about usability. I won’t name the author as I’ve heard him make perfectly sensible points before but some of the stuff apparently in this book had me shaking my head in disbelief.

Now remember that I’ve been building sites since 1994 and have spent much of that time insisting to anyone who’d listen about the need for logical well structured navigation and ease of use, while as an SEO I insist on that as well as well-structured text in digestible chunks using attractive and comprehensible language. What I don’t advocate is dumbing down.

This author however clearly does, if the summaries and reviews are to be trusted. One of his suggestions was ‘halve the amount of text on the page, then halve it again’. Astonishing. Presumably he doesn’t want any search rankings - text is fundamentally what search engines index. Presumably he also doesn’t want well reasoned and informative content either.

We are informed that people don’t read web pages, but only scan them. Certainly a lot of scanning goes on, but when you find something useful then you read it. The scanning is largely part of the human search routines - we are presented with multiple possible sites when we make a search and we then visit them, scanning quickly through to see if they are relevant to what we’re looking for. But we are looking for good quality sites that have useful information, not for dumbed down summaries with no real value.

Another headline of the book was ‘don’t make me think’, along with the suggestion that it doesn’t matter how many clicks you have to make as long as it’s a mindless choice. What an appalling indictment of the assumed intelligence of users and a dreadful waste of the largest collection of information the world has ever seen! It seems to me as if a lot of this is driven by a view of the internet as just a massive selling machine with SEO seen as just a way to cheat your way to easy money. Given the true value of international communication and the hopes that were invested in the net in the early days, I fervently oppose such a view. Make your sites as good as they can be with your SEO’s help and you’ll have lasting value that will deserve to rank well.

While I’m on a roll, does anyone take Jacob Nielsen seriously? At the start he made some useful observations but now most of what he says seems more designed to maintain his guru status, and his examples have always been awful. He keeps telling us to keep it simple - good within reason - but then says do it like Amazon!! Only one of the most hopelessly cluttered sites around. It sells a lot of books because it sells them cheap - not because of the design.

Which brings us back to navigation - make it logical, not mindless. And never mind the usability gurus.