Are guidelines useless?

Upon reading Employee Directory Search: Resolving Conflicting Usability Guidelines (Alertbox Feb. 2003)

“In recent studies on how employees in a range of companies use their intranets, an important guideline emerged: intranets should provide a dedicated search box for finding phone numbers and other employee directory information. Preferably, this employee directory box should appear on every page of the intranet, and it should definitely appear on the top levels of an enterprise portal and on the main intranet homepage.
That’s one usability guideline.
At the same time, one of the most established usability guidelines for search is to provide no more than a single search box on the homepage. Competing homepage search boxes are confusing, and advanced search should be relegated to a secondary position inside the site to avoid seducing users away from the simple search, which they’re more likely to use correctly.
That’s another, conflicting, usability guideline.

Mr. Nielsen goes on to be puzzled as to what to do, and resolves it by recommending emperical thinking and human observation. I think he’s talking about design….

About a year ago, when Nielsen’s book on homepages came out, a friend told me his employer’s site was 90% compliant, and 100% unusable. How were the guidelines useful in redesigning that page?

I’ve also heard it said guidelines are helpful for beginner designers, before they get experience.

You know what? I think beginner designers are the ones who should be kept away from guidelines, as far as possible. Instead, they should work on looking at usable designs, and designing, and seeing their designs tested and retested (in the words of the great homer simpson, “lather rinse repeat. always repeat.”).

ON the Yahoo intranet homepage, we’ve got a search box, a dropdown with “employee, text, and conference room” as the choices and a submit button. Occasionally I make a mistake and search for text in the employee database, but 9 times out of ten, I’m looking up an employee and for the odd times I’m looking for a form or a policy, I do usually catch the dropdown before my pinkie makes a dive for the enter key (they have a mind of their own, those digits).

Still better would be to run queries against each database and return answers from each, rather like amazon does… most of the time amazon knows if you are searching for a book or a CD, and when they don’t, they offer you choices from each category. Why not apply this tidy solution to the intranet problem?

This is called design. Thinking about a problem, and thinking up answers by looking at the world around us: design. Blindly adhering to guidelines is not design. Looking at guidelines and comparing them to the world and deciding if they are applicable to your unique situation: design.

Don’t get me wrong, I think guidelines are valuable. Often I come up against a usability problem and say to myself, “Hey it’s a proximity problem” and can fix it easily. But those guidelines have meaning to me because they represent compressed experience.

A guideline– such as “Place related items in close proximity to each other”– is simple a mnemonic for the hundreds of times I’ve seen users in a lab not notice something because it’s on the other side of the page from the thing they were looking at. The guideline really falls down when the subtleties of the rule are revealed… I’ve noticed that a line between two items is a hard divider for the user’s eyes. One might think that the proximity rule would keep this problem from happening– if they are close, they are associated, right? But I know proximity isn’t just physical space, but also visual structure of that space. A line between two items is the same to a user’s perception of relatedness as an inch of whitespace.

When I worked on the book, I tried to get around the problem for guidelines needing context by using lots of examples and stories. But I wonder if this can make up for the value of watching users use your designs.

I would say if you are a junior designer the three things you can do to become great are:

1. Look at great design. Collect books, use other people’s websites that have been earmarked as good but also winning book and industrial design, visit museums. All media is relevant.

2. Watch every single session of usability testing you can, no matter how redundant the problems may seem. Sometimes the fifth user will suddenly surprise you. Take notes (it will help keep your attention high) and sketch little design solution-ideas to the problems you see in the margins of those notes.

3. Design all the time. Design for fun, design for work. Design a homepage, Design a page for your mom, your sister, your favorite charity… design design design.

And above all, take every guideline with a grain of salt. A guideline is a starting place, not an ending place for thinking. Ask yourself questions after reading an Alertbox– “Why would a dedicated search box for employee search be valuable? Maybe that is the most common type of search. Hey, maybe I’ll check the logs. Hey, maybe I’ll interview some employees, and see what they search for. Hey, maybe I can design it this way…..” then test and watch.

Mr. Nielsen says “The usability field is one in which empirical observation and theoretical analysis reinforce each other.”

I’d add that the design field is one where thought and empathy lead to more satisfying products. Guidelines are useful when the reference thinking, but dangerous when they shut it down.

21 Comments

Add Yours
  1. 1
    Andrew

    Here’s an interesting test for the value of guidelines for beginners: what other design or engineering fields use guidelines from gurus as teaching tools?

    In some fields, it’s probably the reverse: “do it this longer but clearer and more predictable way until you get comfortable with the technique. Then start looking at the more efficient, simpler, more elegant work of experts, and try to learn from them.”

  2. 2
    Andrew

    BTW, you sound as if you end up agreeing with JW anyway. You write at the beginning:

    “Mr. Nielsen goes on to be puzzled as to what to do, and resolves it by recommending emperical thinking and human observation. I think he’s talking about design….”

    But in the end you come back to just his methods and tactics:

    “Why would a dedicated search box for employee search be valuable? Maybe that is the most common type of search. Hey, maybe I’ll check the logs. Hey, maybe I’ll interview some employees, and see what they search for. Hey, maybe I can design it this way…..” then test and watch.

    What’s the difference between what he proposes and you propose? You’re both trying to solve the same problem of “handling the common need for employee search without adding confusion” by studying users, logs, and testing.

  3. 3
    christina

    I actually started out disagreeing, and ended up 70% agreeing. I do agree that observation informs design, but I’m surprised that he doesn’t think about design and the act of designing. He writes as if usability was all that exists.

    SO I guess my point is that design matters, and guidelines are not to be relied on solely. Is that 50% in agreement? 😉

  4. 5
    Keith

    I agree with, oh about 89% of what you are saying. 😉 No really, I’ve had to deal with this exact same problem for an intranet site and we ended up with a solution to the problem that Jakob speaks about in a bit of a different manner. Similar to what you talk about with the drop down, but using tabs to switch between the site search and phone directory. Read more if interested.

    Thing is dispite all the user testing, designing, guessing, etc. We will still run into problems and more challenges as people start to interact with it. I think there are many times when there just isn’t a “right way” to do something, in our case I think a little user education will be needed, and who knows, maybe we’ll need to scrap it and start over – but that is a beauty of the Web.

    The two things you talked about that I think rang very loud and clear in my mind is taking guidelines with a grain of salt and keeping junior designers away from guidelines.

    I don’t know how many time I’ve heard, “Well, Neilsen says….” or “Amazon does it THIS way…”
    When you are starting out you need to make mistakes, learn to think and recognize that for every rule (guideline) there is probably going to be an exception.

  5. 6
    Jared Spool

    I’ve often thought that stories were far more important than guidelines. In our work, we often tell stories and we rarely dispense anything that looks like a guideline. We always caveat our ‘recommendations’ with “you’ll need to see what happens with your users on your site.”

    The problem with a lot of guidelines is that they are just too subjective to be useful. Consumer Webwatch (a project of Consumers Union — the folks who publish Consumer Reports), for example, has a set of guidelines for “improving your web site’s credibility.” One of the many guidelines is “Site privacy policies should be easy to find and clearly, simply stated.”

    How is a designer to know when they’ve adequately achieved this guideline? When is a policy easy to find? When is it clear and simply stated?

    (Interestingly, they don’t seem to care what’s in the privacy policy. You can say that you pay telemarketers to take the phone numbers you collect and pester them continually — that’s ok, as long as it’s stated clearly and easy to find.)

    And even with guidelines’ inherent problems, the number one request we get from our clients are for guidelines. Whenever we publish some of our results, people write us and say “I was really hoping you were going to give me some concise guidelines to follow. I just want to know that if I do x, I’ll design an excellent interface.” Oh, I wish I could. I really do.

    Someday, we’ll have guidelines that our data will prove always (or mostly) work. Then we’ll publish them. Until then, I’m just going to have to apologize and keep telling stories.

    p.s. Glad you’re feeling better. I was getting worried there… I was about to send over some miso soup. I once heard a guideline that says you’re supposed to drink a cup of miso every day. It heals what ails you, the guideline said.

    p.p.s. So, after you’re sick and don’t post for weeks on end, do you always come back with a 497-page essay? 😉 Some sorta pent-up posting phenomenon?

  6. 7
    vanderwal

    You are nearly dead on with the beginners listening to gurus as a guide Christina. I have a client that held the Jakob Homepages book very near and dear. We were working on a redesign with this client and they were making many right steps, but would take suggestions that we could show were improvement (based on some quick gurilla testing). When the client took the product to a new stakeholder the client could not defend the work, nor the reasons why it was a strong design. The new stake holder has no design background and comes from a print jounalism world, where unfilled pixels are revenue lost.

    Taking guru’s at their word does not give the beginner a base to defend what is good design for the users of the site. More dis-service to the users of the site was done by not learning the craft and reaching for the insta-guru book of wonder. Those of us that want to test the redesign (now done by an ouside print design firm with little Web experience, but knows the word yes) are not permitted to as the client does not want to show up the new stakeholder.

    I now love my other clients much more as they listen and thrive on learning more and working to get it right for the users.

    I believe beginning designers can learn from the gurus as there are some best practices that can be gleaned. The beginner must how ever learn to think and discern on their own. The desingers must also learn how to understand the users for whom they are building the product. If the user finds it unusable I hope it is at least pretty and can be sold as art.

  7. 8
    Lyle - Jr. Usability Guru

    I’ve taught an ‘intro to usability’ course, and really enjoy teaching a section on ‘guidelines’ – I cover “laws” (e.g. Fitt’s), “heuristics” (e.g. Jakob’s 10), and “guidelines” (we have a set for ‘web UI’). I have the students do a quick excercise reviewing a web page based on a few heuristics. Then, after talking about guidelines, I have them do a review of different pages based on the guidelines. These are group exercises, so part of the learning is that other people interpret the ‘rules’ differently. The ultimate take-away is that these ‘rules’ can easily be used improperly if you’re not careful, but can sometimes be useful tools. Part of their value is that they help you look at different aspects of a design from different angles (e.g. low-vision accessibility or form layout)

    I try to emphasize that ‘guidelines’ (IMHO) aren’t as valuable as understanding the rationale behind the each of the guidelines. Guideline: “Don’t talk to strangers” might be useful. Having a child understand why not, and when it might be okay, is often better.

    After discussing guidelines, I then relate it back to the need to conduct usability tests and to adopt ‘best practices’ from proven designs. Still, when it comes right down to it, design is damn hard to teach.

    Of course, when people try to borrow from ‘good designs’ – they may be borrowing from a bad design that they, in their subjective opinion, think is good. People also form their own unwritten personal guidelines based on reading a research report, an article or talking to an ‘expert.’

    The key mistep beginning designers make is acting on opinion only – a better way would be to prove all design decisions right or wrong. Of course real projects have constraints, so not everything can be proven/disproven.

    If people could do good user research, iterative design (prototyping, etc.) and conduct ample usability tests, we’d have little use for guidelines. It’s those real world project constraints that drive people to place their hope in guidelines.

    Guidelines are not User-Centered Design. People can’t build a good house only using a hammer, they need the whole toolbox.

  8. 9
    Chris McEvoy

    Guidelines should only be followed once you have successfully broken them a couple of times.

    My own guideline would be : Use as few search boxes as you possibly can, with none being the optimum number.

    Re : Directories. We have a directory for 12,000 employees that is searched about 50,000 times a week. We use a drop down list that allows the search to be scoped to People, Department, Branch, Area, Intermediary and Agency. The vast majority of searches are for people.

    We allow for first/last name searches by parsing the input. If there is only one word in the search terms then we assume a lastname, if there are more than one term then we assusme a firstname lastname, on the rare occasions when someone needs to search for all people called chris then you can search for chris *, you can search on any partial name so a search for c m would find Chris McEvoy (along with another 119 people).

    In addition we only return one page of results (one line per person), rather than paged results. We display a maximum of 200 results from any search, so that if someone searches for a (all people with a surname beginning with a) which matches 556 people, we only display the first 200 matches and ask the user to “try a more specific search”. Because no-one is actually going to look at 500 results to find a person, no-one has ever had a problem with this method of displaying results, because no-one would ever trawl through all of the actual results. Google also rely on this ‘behaviour’ to get away with only ever giving us 1000 results from any search.

    We found that this option was easier for people to understand than having a separate box for firstname, lastname.

    We are doing some work on making all of our searches work via one search box. We are trying to identify what type of search they want from analysing that they type into the search box. I think it is possible to do this in an Intranet environment where the scope of searches is much narrower than internet searches.

  9. 10
    Mike Steckel

    I wanted to comment briefly on this one part of your essay:

    “1. Look at great design. Collect books, use other people’s websites that have been earmarked as good but also winning book and industrial design, visit museums. All media is relevant.”

    I very much wish there was more conversation in the community about this. I would love to see examples of “great design” and discuss them. We seem to be reluctant to discuss what influences us beyond Amazon and Google. We talk too much about “What is design?” and don’t spend enough time delighting in it.

  10. 11
    Lyle Kantrovich

    I agree. One of the most difficult questions to answer is “what web sites have GOOD designs” — most examples have caveats that go along with them.

    Also, how do you talk about ‘good design’ without talking about the user needs the design is serving?

  11. 12
    Lyle Kantrovich

    I agree. One of the most difficult questions to answer is “what web sites have GOOD designs” — most examples have caveats that go along with them.

    Also, how do you talk about ‘good design’ without talking about the user needs the design is serving?

  12. 13
    Jon Horn

    One point that I’d like to mention is that often in designing for usability, you have to balance certain aspects. There might be a trade-off in aesthetics by designing a perfect 100% usable website (if there is such a thing). So, you have to decide which is more important. I find this very relevant in the “1 or 2 search box debate”. When users are searching for employees 90% of the time, i say screw the other text field, and use the drop down to select which search to conduct.

  13. 14
    Ron Zeno

    Guidelines in the hands of the unskilled and incompetent are useless or worse.

    I think beginner designers are the ones who should be kept away from guidelines, as far as possible.

    And let them be totally ineffective instead as they work hard to reinvent the wheel, poorly? Might I suggest instead they research guidelines?

    The problem is that there are far too many, very bad guidelines perpetuated by those who don’t know better and have other agendas. There’s obviously a great deficit in the field in the ability to identify good guidelines and to point out the bad ones. Physician, heal thyself.

    Luckily, there is usability.gov, where people have done a great deal of the work already.

  14. 15
    Tim Beidel

    I do find guidelines are useful as shorthand for known usability issues. The search problem, for example: The guideline exists because a lot of experience shows that users have trouble with multiple, complex interfaces. I always feel more confortable working with designers who know the guidelines, because the guidelines suggest what to worry about when you “depart” from them. With that knowledge, “depart” you should – to the design that works for your audience and tasks, and with the other elements of your visual design.

  15. 16
    unraveled

    Guidelines and Design

    Jakob Nielsen recently published Employee Directory Search: Resolving Conflicting Usability Guidelines, in which he explains how usability guidelines often conflict…

Comments are closed.