How to evaluate your website, Part 1

I have done shedloads of website usability in my time.  Some projects are straghtforward, others are very complicatedThis is my guide to thinking about your own assessment – you might still need to do user testing, but this should start to surface the glaring problems.

1. Define your website’s goals

Your goals may be many and varied.  List them all out before you start.

  • Enabling people to do their banking online
  • Informing prospective students about evening classes
  • Persuading visitors to come to your gallery
  • Providing up-to-date information about genetics research
  • Helping people find out about your company
  • Selling clothes/books/DVDs
  • Showing site visitors that you are cool and happening

Brainstorm them, and then group them roughly into the hard goals (the concrete things that you want people to do next, like order something or make a call, or bookmark the page); and the soft goals, like the image or values that you want to convey.

You might want to put them in rough order of importance.  Even now, you may be able to able to spot some obvious flaws. ‘We want people to sign up for our newsletter, but we’ve buried the link in tiny print at the foot of the page.’

2. Think about the typical website visitor’s journey through your site

This is usually called ‘the user journey.’  There will be a number of these.  Some of them will be very important to you, others might be less important or less obvious.

So.  What do people do here?  What different journeys can you identify?  Here are some to get you started.

  • Browsing for information (about products/costs/opening times)
  • Choosing something, or making a shortlist
  • Downloading a report
  • Buying or booking
  • Ordering a prospectus
  • (Often forgotten) Looking up your postal address
  • (Often forgotten) Finding travel instructions before a meeting

Also think about the difference between the journeys taken by the first-time visitor and the returning visitor.

3.  Check out your website’s performance on the basics

If the basics don’t work, you can forget about everything else.

Navigation (menus, tabs, headers, footers, links)

  • Can your visitors find their way easily around the site?
  • Does your navigation look like the navigation on most websites? (Hint: Horizontal scrolling is unusual)?
  • Can visitors find everything they want (products, Head Office, privacy policy, vacancies etc)?
  • Can they find their way back to the Home page? Even if they don’t know about the convention of the header doubling as Home link?

Navigation is often a real problem for visitors.  It can also be the hardest for site owners to understand, because they know where everything is.


  • Does everything work? (links, downloads, video)
  • Does it work across all the important browsers/platforms?
  • Does it work quickly? If it works slowly, does the process keep the user informed?
  • If it doesn’t work, does it degrade gracefully? (Hint: Have a look at your 404 page)


  • Is the language clear and appropriate to your users?
  • Are your instructions friendly or nasty? (Hint: Fill in your forms and leave out some of the fields. How do you feel about the error messages?)

Basic Accessibility

  • Is the text easy to read (reasonable size, good contrast, pleasant background colours)?
  • Is text size adjustable?


  • Do the colours, layout and images reflect the message you want to send to visitors?
  • Does it look…designer, professional, ordinary, folksy, edgy, homemade…?
  • Is that what you’re aiming for?

4. Assess how you are doing with the main user journeys

You can score them if you like. Nothing complicated: excellent, room for improvement, poor, entirely missing.

Focus in particular on what your visitors do when they come to look around, and then what happens when they take action.

5. Look back at your master goals list

You should have enough information by now to understand whether you are meeting some of those goals.   You may find that some are easily ticked off as excellent or circled as having problems that need solving. However, there will be a few left where you don’t have enough information to proceed.  You may not know how real users view the site (in which case you need to get some feedback from real people), or you may feel that the basics are in place but the site is underperforming.  Now you need to start the hard thinking about how to improve the website design so that you meet your goals.

In Part 2 (up early next week) I’ll run through my process on solving the thorny areas where the site is clearly underperforming.

Adding beauty to market research

go on, do something different!

I’ve been reading various posts about market research and social media, which tend to focus on the usual self-hating stuff about the market research industry’s vulnerability.  I agree, pretty much:  some of the space that research took up is now being eaten away by other specialisms (data mining,  search engine optimisation, and web analytics), while the rest of what’s rightfully ours is taken by DIY tools such as SurveyMonkey.  (Can QualMonkey be far behind?)

Personally, I would argue that MR agencies have neglected to develop certain 21st-century skills in-house.   Market researchers tend to focus on data collection and analysis technology like Confirmit or SPSS  (if you’re lucky).

Things market researchers don’t bother with:  design.  Graphic design, web design, information design, whatever.   Design is right down at the bottom of the pile.   You buy it in, or you manage without it.

Researchers writing presentations huff over the latest critique of death-by-Powerpoint and insert a couple more company-approved clipart images into the 80-page deck.    Somewhere, a designer is weeping.

So.  If it were up to me,  I would not only run shedloads of training on statistics and experimental design, but I’d include these:

  • Essentials of graphic design
  • Digital photography
  • Photoshop
  • Using stock image libraries
  • Web design and an introduction to CSS

And if I were running a big research agency, I’d invest in some graphic designers and programmers to create some nifty and beautiful interfaces for running surveys and online communities.

Why shouldn’t people expect loveliness in a survey?

What skills would you like to see?