As an example: on Monday 12th December 2005, I wanted to buy a copy of Jamie Oliver's new cook book Jamie's Italy from amazon.co.uk. So, I went to the "Books" section of their website and searched for "olivers italy" and these 9 items appeared on the results page:
1. "The American Tractor" by Patrick W. Ertel
2. "A Garden in Lucca: Finding Paradise in Tuscany" by Paul Gervais
3. "History in Exile: Memory and Identity at the Borders of the Balkans" by Pamela Ballinger
4. "Oliver Tractors" by Jeff Hackett, Mike Schaefer
5. "Wyoming (Moon Handbooks S.)" by Don Pitcher
6. "Wines of Australia (Mitchell Beazley Wine Guides)" by James Halliday
7. "All Music Guide to Jazz: The Definitive Guide to Jazz Music" by Ron Wynn (Editor), et al.
8. "Larousse Gastronomique: The World's Greatest Cookery Encyclopedia" by Prosper Montagne
9. "The Teacher's Calendar: The Day-By-Day Directory to Holidays, Historical Events, Birthdays and Special Days, Weeks and Months" by Holly McGuire (Compiler), et al.
Jamie Oliver's book didn't appear anywhere on the results page, even though it had been Amazon's 3rd best-selling book in the previous 24 hours.
The problem was that I had typed "olivers italy", instead of "oliver's italy" (which would have returned Jamie Oliver's at the top of the search results list). That single missing apostrophe was all that it took for Amazon's expensive search engine to splutter, fall over and fail.
So - if Amazon can't do it, it must be impossible, right?
Wrong - here are some things the boys & girls at Amazon could - and should - have thought about.
Two types of problems
There are two basic types of problems that a user can experience when they are searching for something:
- User-error - the correct search term is entered incorrectly (i.e. the user intends to enter a search term that would cause the search engine to return results that are relevant to their needs, but they enter it incorrectly).
- Search engine error - the wrong search term is entered (i.e. the user enters a search term that the search engine does not relate to their needs).
User error
People generally enter the correct search term incorrectly because they either:
- Don't know how to spell it.
- Have made a typing error
It's important to realise that there are millions of potential customers who can't spell very well. For example, a 2003 survey of the literacy (i.e. reading and writing) estimated that there were 16% of English adults (aged 16 to 65-year-olds) had literacy levels no higher than those expected of an 11 year-old (source: The Skills for Life Survey).
Also, let's not forget that according to the British Dyslexia Association around 4% of the population are severely dyslexic and a further 6% have mild to moderate dyslexia problems.
This means that your search engine has to account for people making basic knowledge-based spelling mistakes.
Your search engine should also account for people who know how to spell what they are looking for, but make typing errors. The main categories of typing error are:
- Characters close to one another on the keyboard being entered erroneously (either in place of - or in addition to - the correct letter). For example: wrong/wring ; for/dfor.
- Characters being omitted. For example: missing/missng ; oliver's/olivers.
- Characters being entered too many times. For example: impossible/imposssible.
- Characters being entered in the wrong order. For example: disaply/display ; being/ebing.
Your search engine should allow people to make these mistakes and still return useful and relevant results.
Even though we have named these types of issues ‘User error', if your search engine fails to return information that that the user is looking for it is, of course, your fault and not theirs!
Search engine error
When people enter the wrong term into a search engine, it is only wrong because you have not anticipated it. You should aim to cover as many bases and anticipate as many different search terms as possible.
What to do
The next steps for making your search engine perform better are really simple:
- Sit down and make a list of all the spelling errors, typing errors and alternative search terms that you think could possibly be relevant to your site (e.g. actually look at your keyboard and think about what letters are close to one another).
- Ask other people in your organisation to make similar lists.
- Do some research into what search terms people are using on your site (e.g. interviews, questionnaires, check your search engine logs, etc.)
- Apply everything you learn to your search engine.
And that's it. You now have the knowledge you need to begin improving your site's search engine.
Other thoughts
- Improvements in word processing software have made people lazy typists. Software that auto-corrects many spelling and typing errors means that people are no longer forced to review and correct their work to the same extent as in the past. This means that many people are getting out of the habit of precise spelling/typing. So, when they move out of an auto-correcting environment (and onto a website, for example) they are more likely to make - and less likely to notice/correct - mistakes!
- Search results pages should display the search term the user entered in large text (e.g. 28pt). This would help people spot any inadvertent errors.Results pages should also provide the telephone numbers for customer enquiries/assistance.
Submit Your Site Search
One method that web designers can use to design and structure a website is to use frames. However, if your web site utilises frames then you could have major problems getting indexed in the search engines. Although site design using frames has become less prominent with the rise in popularity of CSS, there are still sites out there utilising framesets. But there are ways to sidestep the problem.
In the past, frames were considered a great way for site designers to quickly and easily display content whilst maintaining a structure throughout the site (e.g. by having a title, or navigation bar). They allow more than one HTML document to be shown on a page by displaying each one within its own "frame", which are defined by the "frameset" HTML tag. This tag defines which pages to show and the size and position of the frame it should appear in. Although this sounds great in theory, it creates problems with both navigation (and therefore search engine indexing) and usability.
As web usability expert Jakob Nielsen has documented on his site, http://www.useit.com/alertbox/9612.html, there are several usability issues associated with frames:
They can be disorientating to users ? if they click within one frame and this affects another frame it can make navigation confusing.
The address bar doesn't change as the user navigates between pages, because the pages load inside the frameset. Again, this can confuse and disorient users.
The loading time increases because there is more than one page to load. If the site takes too long to load then visitors are going to go elsewhere instead.
If a user bookmarks a page within the site, they will be sent to the default frameset when they revisit rather than the bookmarked page.
On top of this, there are also issues with search engines finding and indexing all of a framed site's pages. The most fundamental problem is that search engines find and index pages by following HTML links in a document, and because framesets reference a page rather than linking to it, the pages within the framed site cannot be reached. Ultimately, this means that no matter how large a site is there's a possibility that only the frameset page will be indexed.
There is a way round this problem. You can place a "noframes" tag in the body of your frameset page to provide alternative content that will be displayed if the browser is not frames compatible. Fortunately, search engines can also read this tag, so if you include normal links within this tag the search engines can spider them like a normal site. Visit http://www.w3schools.com/tags/tag_noframes.asp for more information on the noframes tag.
So now the search engines can find your pages, but what happens when your visitors find them?
Because the frameset dictates which pages should be loaded, if an internal page is accessed directly through a search engine then it will be loaded outside the context of the frameset. This means that the page will be viewed on its own without any of the intended accompanying pages specified by the frameset, such as a navigation bar. This type of page is known as an "orphan" page. They are confusing for visitors because once they find the page they may not be able to navigate the site, meaning you may end up losing the visitor, or worse, a customer.
Again, there is a way around this issue. You can use JavaScript to force the page into its framed context, and although this causes problems with JavaScript incompatible browsers it does neatly avoid the orphaned page issue. There is a great tutorial available at http://www.webreference.com/js/column36/forcing.html showing how this is done.
There are alternatives to frames that allow for similar functionality, the most popular would be to position elements on a page using Cascading Style Sheets (CSS), or if you are designing a dynamic site then utilising Server Side Includes (SSI) would be a good option. Although the above shows how to avoid the complications caused by frames, for the various stated reasons it would be better to avoid them altogether.
Both Tim Fidgeon & Craig Broadbent are contributors for EditorialToday. The above articles have been edited for relevancy and timeliness. All write-ups, reviews, tips and guides published by EditorialToday.com and its partners or affiliates are for informational purposes only. They should not be used for any legal or any other type of advice. We do not endorse any author, contributor, writer or article posted by our team.
Business Plan For Photography Cost evaluation is extremely important and you must be aware of what your bottom line is. If you operate with the bottom line always in mind then the investment should be less stressful
