Friday, January 20, 2012

Should you design websites for the '1%'?

A concern I’ve had for the last ten years is how websites are designed and approved by organisations (both in government and the commercial sector).

In a better-than-average world, when asked to develop a new website or improve an existing one, the web team goes out to discover what users think of the site.

This involves identifying the site’s key audiences and using surveys, focus groups, other research and past feedback to identify good and bad design and usability features. After this the team come up with concepts, tests them on audiences and refines.

 (In a average or worse world the web team isn’t given the time or resourcing to do all this research, so short-cuts the process with their ‘best guess’ design improvements based on feedback and experience. This is far too common but can still deliver improvements.)

When the web team reach final agreement on a few design alternatives, they go to senior management for approval, often with a detailed case explaining all the design decisions.

And this is where the process breaks down.
  • “Can you make the website more blue? I want it to be bluer.”
  • “I like (pick a random site visited in the last day). Can you redesign it so that our website looks just like that one.” 
  • “I don’t use search, I use menus, so can you move the search to the bottom right of the page” 
  • “I don’t believe anyone wants three columns in a webpage, please restructure to two columns.” 
  • “It’s too hard for me to find anything, can you simply list all the main site categories and pages in the homepage.” 
  • “We’d prefer to organise information by our divisions rather than by subject, I’m sure that would be much easier to understand” 
  • “We actually wanted the website to look just like the printed brochure” 
  • “I like the shirt I am wearing today, make the website the same colour”

Suddenly web teams have to reassess what they are attempting to deliver and who they are delivering for.

Their collective expertise and research is no longer relevant.

The audience of the site is no longer relevant.

They are designing for one person, or a small group of people – decision-makers who are often not the target audience and possible don’t even use the website.

This is a source of great frustration for web teams. They are no longer designing for the 99% of their audience, they are designing for the 1%.

Now what if this process was turned on its head...

Rather than having an executive or Minister approve a website, we instead released several near final designs for A/B testing on online audiences (as organisations like Google, Amazon and Microsoft do), a proven and effective technique, or took the final couple of design alternatives and put them online for the public to vote on and thereby approve.

Of course there would still need to be some level of senior executive involvement in defining the organisation’s overall requirements for the website. The site does have to meet the organisation's goals.

However the actual approval would come from the audience, the 99%, people using the website, the people you wish to communicate with, support, engage or influence.

Radical? Maybe.

Effective. Certainly.



  1. Craig,
    Once again great points you have raised here. I would like to point you to an example by the ACT Government during last year where they actually went to potential users of a site - indeed it was the ACT Government's new Open Government site.

    I recall that three sample sites were 'exposed' and feedback sought from potential users. I believe this was the first time that this had been done by the ACT Government and wonder whether they have recorded some 'lessons learnt' from that exercise that they could share with other agencies, both ACT and Federal.

  2. Agreed, there's often a disconnect between project owners' expectations & optimising end-users's experiences.

    +1 to more user research. Consultancies need to get better at tying the outcomes of that research to their rationale for design & interaction decisions.

    Neutral re user research in the form of public polling on visual design mockups. I personally fear the knotty process and underwhelming outcomes of 'design by committee' (where committee includes public forums).

  3. Spot on Craig. The key here is how quickly the collective expertise of a deep engaged web team can be blown away by a bunch of half-interested focus group testers.

    I lean toward the Steve Jobs thinking (but not all the way) - don't ask what they want, just deliver great product

  4. Great post Craig. Agree with Jimi on the design issue. Akio Morita used to say that Sony would never have developed the Walkman from a focus group - it took an idea.

  5. Re Jimi's comment: I think Craig was levelling the blame at executives rather than focus groups.

  6. While I sympathise with the concerns raised here, they raise an interesting tangential point. Many of the posts I read about online activities emphasise some problem with senior management. Of course, it's entirely possible that senior management is at fault. But even if they are (and that's not what I think in every case but I am not disinterested), are we risking creating a self fulfilling prophecy - online complaints about senior management raise concerns among senior management which creates risk averse behaviour by senior management about online activities which creates online complaints about senior management ...? How can we break this vicious cycle?

  7. Craig - often we don't even get to the first step. The client knows what users want and obviously knows more than experienced web teams. User research, web stats, card sorts etc. - who needs all that rubbish !

    Or the Minister needs a website for his/her launch in a fortnight - leaving no time to go through a proper development process.

    And then 6 months after the site goes live, the client comes back and wants an explanation as to why the site is getting such low traffic, or why engagement is so poor.

  8. "Why the site is getting such low traffic" is something I'm grappling with on one of my sites. Unless you set some sort of target at the outset, how low is "low"?


Note: Only a member of this blog may post a comment.

Bookmark and Share