Information Architecture and Design - Visual Design Don'ts

There are a myriad of issues when it comes to visual design, many more so than any of the other elements of designs mentioned so far. There is a single reason for this, when it comes to visual design, everyone thinks they are an expert! For the other elements of design, most people are happy to leave it to true experts, but when it comes to visual design everyone has an opinion. The most heated debates in conference rooms regarding websites will almost always be about visual design. There are a number of common approaches to visual design, each with their own strengths and weaknesses. Although you may not be able to change the approach taken, at least by understand it, you can help to shape the outcome and rather than being a part of the problem.

Design by Default

Every site has a design. It might not be very good, but it will exist. When design fails to be considered as element in it’s own right, a default approach is taken. What this means is the design will follow an existing approach

Existing Technology

This is the easiest of the “design by default” approaches. It’s when the design mimics an existing technology, more often than not, it’s a copy of windows user interface elements. eg. windows style icons, drop down boxes…etc. Although not the worst approach, it’s also not the best. Using known approaches, such as drop down menus will be familiar to users but it may not be the best approach for your particular site. What this approach does is take a short cut. It’s indicative of a lack of effort and will result in a lack of quality in the end result. That’s not to say using conventions is a bad thing, it’s about making sure you use the right conventions in the context of what your site is supposed to achieve.

Existing styleguide

Many corporates will have an existing style guide for print. Some of these corporates will expect that the print styleguide will apply to the website. It can be used as a basis for a web style guide but should not be applied exactly as the web is a very different medium to print. Overall, as an approach, at least there is a styleguide and hopefully it will be applied consistently. Where this approach falls down is in areas that there is no equivalent in print. You can’t click a brochure, select a drop down box, rollover a button…etc All of this requires a web styleguide and the common result is a hybrid of a print approach and design by default as the designer tries to fill the gaps.

Exisiting Business Processes

This is similar to the existing styleguide but goes further. In most businesses there will be processes that are represented by a printed form or electronic interface (eg. Database). Rather than adapt these to the web, the approach is to copy them exactly. The end result can be a form that on a piece of paper fits neatly on one page but scrolls a long way down the page on a website or in the case of existing applications is a database entry screen that is mimicked on a website that also has to deal with navigation elements that the existing application doesn’t have to compete with. The inevitable result is a hodge podge that might seem easier for the business but is harder for the user to use.

Developer’s interpretation

This is probably the greatest of all design by default sins. Letting the developer design the end solution. A talented eZ publish developer, Bruce Morrison taught me this lesson the hard way. Even after repeated requests for a visual design to be applied to a part of the site, none was forcoming, so, Bruce created a visual design that was so ugly that I was forced to do something about it. Bruce understood the importance of visual design and also that he wasn’t the best person for the job. Not all developers are this smart and in the absence of a visual design, will make one up. If it’s not dog ugly, it will probably make it to the final solution, ie. design by default. We would never expect a graphic designer to come up with a database design and we shouldn’t let developers do visual design.

Design By Lead

Design by lead is less obvious but can produce equally as poor results as design by default. With design by default, there is a lack of forethought and something is put in place to fill the gap. With design by lead, the visual design has been considered but rather than come up with the right visual design for the project, the approach is to copy someone else’s design. This is wrong on a number of levels, first it’s lazy, second it’s cheating and third, it may not work.

The common examples of design by lead are following what your competitor is doing or following the current trend. It’s basically following the herd mentality. Just because someone else does it, doesn’t mean it’s a good idea for your site. There was a time when many sites had a flash intro, now we now they get in the way and flash intros have thankfully almost entirely disappeared. Design by lead, ie copying an approach is a bad thing, however that’s not to say we can’t learn from existing visual designs, it’s a matter of adapting elements and applying them to your own visual design – not simply copying them.

Design by Fiat

It’s a hard to pick the worst offender of visual design issues but design by fiat has to be close to the top of the list. Design by fiat is when one person dictates the visual design based on their personal preferences. If the person dictating the visual design is an expert, then it can work, if not, the results can be disastrous. On a recent project, after the visual design had been approved and implemented, the CEO demanded that drop down menus be added. When asked why, the answer that came back was “because he likes them”.

It didn’t matter if useability studies had shown drop down menus can cause frustration in older users, this site was going to have them. It also caused problems with making such the drop down menu would appear over flash elements on the page and problems with the information architecture. None of this mattered because the CEO was employing design by fiat. When this happens, user needs aren’t considered and best practice is ignored. It can also cause problems with the existing design creating rework. It’s impractical to think we can prevent design by fiat, all we can do is hope to educate our clients enough to trust us to do what we do best.

Design by Committee

This is along the same lines as design by fiat with the blame shared across a group of people rather than a single individual. The end result is often a compromise that tries to keep everyone happy except for the most important person, the user. The reason the results are often poor is that expert views are not considered and the project objectives can be obscured by committee members arguing over personal preferences. At times it can combine design by default, by lead and by fiat all in one! All I can say about design by committee is to quote this well known saying… “ In all the parks in all the cities you’ll never see a statue of a committee

Duelling Departments

The duelling departments scenario is a politically tricky one that actually goes deeper than just visual design but only comes to the surface when people actually see what is going to be produced. It’s at this point one department finds out they don’t have any space on the homepage, marketing thinks the logo is too small and sales demand that space be made for banner ads top, bottom and in the middle of the page! There is only one thing that you can be sure of, someone is not going to be happy. That’s why it’s so important to have a clear project objective to resolve such issues to and avoid the political debates.

Religious Debates

Religious debates are not refined to just visual design, you can find them occurring between developers over technical design – the only difference with visual design is everyone feels qualified to pitch in. Here’s a typical example of a meeting over a website design

John – “I like drop down menus”

Jill – “Drop down menus are awful”

Simon – “Drop down menus could cause performance issues”

Dave – “Company X uses drop downs”

Grace – “Drop downs save space”

You can see in these statements a range of personal views, all bar one lack any basis in research. The main problem with these statements are they are concerned with what these particular individual like, not what is best to achieve the project objectives and what the user will like.

The Average User

As a part of religious debates, the concept of the “average” user can be used to ustify a particular argument. Eg. “users prefer drop downs”. This is a flawed argument because, and I’m sorry to upset anyone new to web design, but there is no such thing as the average user. Such a creature is no more real than the abominable snowman or Santa Claus. Sure, I’d be happy to sustain the myth of the average user if it mean we could have another public holiday in celebration but it’s of no use when it comes to visual design.

Web users are unique and idiosyncratic, we all have our own likes and dislikes, expectations and assumptions. Rather than thinking in terms of the average user, we need to appreciate the particular audience for the website we are building. Know their individual intentions, experience, cultural backgrounds and age. To realize that it’s not about the average user but appreciating the complexity of people. This means that not everyone will like the design created, what matters is as many of the right people do, ie, the people that the site has been created for.

The Right Questions

Like most things, visual design comes down to asking the right questions and avoiding design by default or fiat. For example

Will using drop down menus with these items and this wording in this context create a good experience for our target market?

Will using drop downs better help us achieve the project objectives?

If All Else Fails

If you can’t avoid any of the design issues outlined, then the only answer is to put it to the test. Hopefully that’s in a controlled environment with expert testers and a formal report, if not, as happens too often, the visual design is tested by putting the site live and seeing how people respond. Both tests work but the latter is more expensive and potentially embarrassing.