Sorting it out

So much for constructive criticism and demonstrations of positive alternatives. But how might this work in practice for the UKAPW? This page gives one step-by-step approach.

A fast track and low risk method of testing the water

Here's a simple way that we could judge whether the UKAPW or an alternative was preferred by actuaries:

  1. Invite 100 actuaries at random to sign up to be reviewers of the UKAPW.
  2. Put up a mock alternative home page (if the webmaster is too busy or the profession has limited appetite we could use the Society of Actuaries home page or myFIA's alternative home page)
  3. Take a vote and act on the result
  4. Extend to other doorway pages.

Bask in the glory of being user-responsive etc. Another approach is to nominate 10 actuaries to write to the top level ("landing page") content, and agree the content by committee.

[1] Adopt a genuinely positive attitude

The first step may be the hardest. Websites are not built in isolation and their users expect them to behave in certain ways, even if the website holds some monopoly power. Much of the work behind a website can and should be done before any coding is carried out; the visual part of this might involve pencil and paper in a process web designers refer to as "wire framing". Good practice has developed.

We can solve this quickly if we want to do so.

[2] Think content, structure, content again

This following is a robust way of thinking about websites, before the coding begins:
  1. Think about the content of the website: have many areas of content will there be?
  2. Then think about how a consistent folder and site structure supports this.
  3. Think in more detail about layout.
  4. Finally, write the content. The content is a "use test" for site structure.

[2.1] Content [nature, amount, grouping]

The purpose of the website and the target audience should be clear; coding before this would be ridiculous! One assumes that the UKAPW is mainly aimed at student and qualified actuaries or those aspiring to be such.

As in many areas (e.g. mathematics, science, literature) it will help to break down the content further into categories. This will be particularly beneficial for a site with a lot of content. A small amount of content may not justify a separate area?

Sometimes the site's scope grows or contracts over time: grouping can help clarify the purpose of the site. The UKAPW does a reasonable job here, but could drop its "News" area.

[2.2] Structure [areas, folders, layout]

The site's folder structure might mirror the areas/groups; other things being equal this is a good idea. The UKAPW is a little strange in this regard; most areas are under the folder www.actuaries.org.uk/FolderName – for example the news section is at http://www.actuaries.org.uk/news – but there are exceptions:

Content also interacts with site layout; a significant amount and depth of content might suggest that a part of the layout (a column?) should be dedicated to "related information" which might comprise a selection of internal and external links. The UKAPW attempts this, but the links aren't really related. More effort is required.

[2.3] How do we decide on an appropriate layout?

We might consider questions such as:
  • What would a typical page look like?
  • Is there any advertising?
  • Are there any related links?
  • If the content is "rich" is it explained further e.g. in a side panel?
  • Where will the navigation be?
  • How much of the site is accessible directly through the navigation?

Each page should have a consistent "look and feel" with the items placed consistently, making life easier for the site user. This gives rise to the concept of site structure.

[2.4] Content again [quality, clarity, length]

Finally, of course, the content has to be written. It's a good idea to write the introductory pages ("landing pages" in UKAPW terminology) early on. Those pages should explain the purpose of the section and how further content is to be found. This is a good test of whether a separate section is really necessary.

[3] Integrate the navigation and sitemap

As outlined above, the site structure and navigation depend on the amount and grouping of content. A bit of thought here goes a long way.

[3.1] Navigation theory and practice

There are various satisfactory methods of navigation, and whole books have been written on this. Navigation theory is developing, centred on the user (customer!):
  • Items should not be hidden deep within multi-level menus since this means a game of "hunt the content".
  • Megamenus are acceptable; they typically group items effectively, striking a balance.

The theory is being driven by science, since the effect of (e.g.) different menus can be tested, using a concept such as A / B testing as well as being debated by passionate webmasters and usability gurus

[3.2] Consistency of navigation and sitemap

It is possible to ensure consistency between the navigation and sitemap. Typically the navigation will reveal only the items within the current area, whereas the sitemap will show everything (but see the "trap" below). This can be achieved using the same underlying file, so that updating the file simultaneously updates the sitemap and navigation: one way is to use CSS to hide the navigation outside the current area.

[3.3] A sitemap trap for large sites

Consider not putting all of the content into the navigation or sitemap. Of course, for the largest sites (e.g. Amazon) this is impossible, but even for sites with a few thousand (hundred?) pages such a sitemap becomes impossible to navigate, often exacerbated by multiple "levels" of site structure. This can be seen on the UKAPW. Sites such as QFINANCE have multiple sitemaps.

[4] Provide robust links and a feedback mechanism

Bad links are the bane of a website and are largely unnecessary. Each link on a small website can be tested by hand, although this is not recommended. Larger sites may use free automated link checking tools: one example is Xenu link sleuth For the largest websites, with tens of thousands of links, automated testing should be combined with extra automation and easy user feedback.

[4.1] Automation

The site should be able to monitor whether a link is "bad" i.e. the page linked to does not exist. The best sites:
  • Supply a custom error page
  • Automatically submit the bad link and its "calling page" to the webmaster

After the webmaster identifies the intended page, a good site will then be updated using quality "search and replace" tools, so that the bad link is removed from all affected pages.

[4.2] User feedback

An effective means of user feedback should be incorporated, with a link ideally available from each page. The feedback should enable the user to select from a range of options e.g. "bad link", "link to wrong material", "poor quality content", "grammatical errors" etc. Obviously space for further explanation should be given.

[4.3] Feedback *to* users

If users are good enough to feed back, get back to them. Either email them or provide a list of user-driven site changes over the last month. This is the way to improve credibility, content quality and link accuracy.

[5] Cut out the CV-driven design

CV-driven design is where the writer of the website incorporates a technique because it builds his skills (or enhances his CV) rather than because it makes life better for the site users. Regrettably the UKAPW contains various examples of this, typically resulting in irrelevant and overly-complex functionality; instead of providing a simple link, a box opens and the user is invited to click a further link, often with the same title!

[6] Engage the community

The most successful modern sites have huge amounts of user engagement:
  • A famous example is Amazon where the user reviews can be key to buying decisions.
  • A less well known example is TechSupportAlert which recommends freeware.
Two ways that the Profession might engage its members in improving the UKAPW are to:
  • Set up a bad link reporting facility. This should have some tracking facility to keep the web team "honest" (aka a stick). This is one thing that myFIA might otherwise provide.
  • Let members propose content. This could start with the "landing pages" and progress to other areas. Competing pages could be set up, with users voting for their favourites.