Back to the User: Creating User-Focused Websites
Published on October 15, 2002
Since our earliest interviews with Web users, enormous progress has been made in shaping sites that inform, engage and build lasting customer relationships. Imagine searching for information without Google; finding a rare collectible without eBay, figuring out what ails you (at 3 A.M.) without WebMD, or getting the news you want delivered to your desktop without your favorite online newspaper.
We've all come a long way. That being said, when we observe target users trying out Web sites in our consumer lab, some common problems persist. This article summarizes five key lessons learned from listening to and observing all kinds of users (from teens to seniors to doctors) try out all kinds of Web sites at various stages of development. Our goal: to provide some overarching guidelines about bringing a customer voice to site design.
Lesson 1: If you want to know what people want, ask them.
Many marketers--even of well-established brands--forget the basics of marketing and usability research and treat the World Wide Web like the Wild Wild West, not exploring up front:
- Who will be using the site--customers, prospects, investors, etc.
- How these users speak and think
- What content and features drive their interest
- The relationship between the on- and off-line experience
Over the years, we have conducted post-launch research for many organizations that launched their sites without conducting any marketing or usability research--a treacherous process known as "Launch and Learn." The outcome: costly revisions implemented after alienating and frustrating untold numbers of customers and prospects.
We have also worked with many organizations that first expose target users to their Web sites in usability testing--after considerable time, money, and emotional investment has been made in a particular site architecture, feature set, and look and feel. This often occurs right before launch when there is little chance for user feedback to be integrated--and at great cost. Sound familiar?
To fast track the development process and infuse initial design ideas with a user mindset, we suggest conducting a few carefully constructed focus groups at the very outset. These are some of the questions we've found that focus group research can help answer:
- How to create a "front door" that will successfully "invite in" each of your various audiences (customers and prospects; doctors and patients, etc.)
- How to "bucket" the main content areas and label them intuitively
- What the site can do to reinforce a brand's strengths and overcome its shortcomings (e.g. beef up customer service, offer products there is no room for in your stores)
- What features and content are worth building or acquiring
If done well, focus groups should generate insights that help designers build a robust prototype of the site that comes pretty close to nailing what users want. It is at this point that a site is ready for usability testing.
Lesson 2: If you want to know if they can use your site, watch them do it.
Many marketers launch sites without knowing if anyone other than their developers can use the site to:
- Buy a car
- Build a stock portfolio
- Send a gift
- Find out about a drug or disease
- Get the right cell phone plan
Despite all the progress we've made on the Web, how many of us still today experience:
- Error messages that don't help us
- Registration processes we can't complete
- Products we can't find
- Purchase paths that don't lead to a sale
Either these sites have not done usability testing or the testing they've done has not accomplished its goal--to make sure the people who use the site can easily and successfully do what they came to do.
The following are tips for how to conduct usability testing that ensures that your users can--and want to--use your site.
- Interview people who would actually use your site. If you want to know if a gift finder service works, there is no substitute for observing the folks that would choose to use it.
- Start early--before you've built out the entire site.
- Even if you use a slide show, show test participants screens on a computer, not paper. We believe strongly that all of us react differently to a computer screen than we do to a printed page. Test in the medium you ultimately will use.
- Give people tasks to do that let you see them travel the key pathways you want to make sure are successful.
- Word the tasks so they don't use the language of the site.
- Don't just observe people as they are trying out a path: ask them what they are thinking. If queried at the very moment they click on a particular button, test participants can tell you exactly what was going through their mind--and offer invaluable help about how to fix navigation, labels, etc.
- Iterate. After the fourth person tells you they don't know what a key label means, change it. Modify the prototype based on their collective feedback and try a new label with the next four people. Iterative design and testing makes the best use of research dollars and accelerates development.
- Document your findings on video. Reviewing tapes of the user experience is very helpful to developers. Integrating video clips of real users into your presentations to management provides powerful support for your recommendations.
Lesson 3: Your homepage is a 30 second window of opportunity. Don't be shy.
Many homepages are so cluttered that it is difficult for site visitors to figure out basic information such as what the site is all about and who it serves. One of the things people tell us about Google is that the sheer simplicity of the homepage gives users the impression that their search will be successful.
While your homepage may need to contain a great deal more information than Google's, there is a lesson to be learned from their approach.
We've found that if you're lucky, users will stick around and try to figure out your homepage for about 30 seconds. Therefore, their first impressions are critical. Users should immediately be able to:
- Understand what you offer and feel like they've come to the right place
- Select the path that is right for them--HMO plan member, HR Director, online banking customer, loan applicant
- Identify the range of activities they can do--get recipes, shop, download forms, chat, etc.
- Intuit what is most important from where you've placed content.
We find that users consistently correlate where you place "stuff" on the homepage with its importance:
- Left, center, and above the fold = important.
- Below the fold = unimportant.
- Anything with a logo or ad-like visual, next to a banner or on the far right of the page = advertising which = "ignore" me.
Just like with a newspaper, we as Web site users have been trained to select "in" some information and select "out" others--and we've created strategies for navigating lots of information. It is critical in laying out your homepage not only to eliminate what is extraneous (or can be introduced later) but to put the most important content and functionality in the high rent district!
Lesson 4: People don't read, don't make them.
We figure about 10% of the population at most are true "readers." You occasionally encounter them in usability testing. They read every instruction, caption, copy point, etc. They make informed decisions about what button to click on and what path to take. It is important to note that they are a very small segment of Web users. The rest of us:
- Click first, read or think later
- Go directly to a bolded word, icon, or button that looks about right
- Skip over directions, help pages, or text that gets in the way of where we think we want to go
- Take a lot of inadvertent pathways and need an easy way to get back on track
In order for your site to succeed, it is important to design for the "non-reader":
- Any words, instructions or data crucial to a user's success should be embedded in a visual icon so it can't be missed.
- People will tend to fill in the first field they see even if it isn't for them. So, if you have two alternate paths, try presenting them side-by-side vs. one after the other. This will protect non-readers from filling out a field that is not intended for them.
- If a task requires several steps, number them so users can easily track where they are and what to do next.
- Don't tell them anything until they need to know it.
- Many users are very literal so make sure your words cannot be misinterpreted (e.g. "Click anywhere to begin"; "Buy now.")
- Provide enough information up front (e.g. article summary, product overview) so users can quickly assess if they want to take a path--but not so much detail that they have to wade through lots of copy before getting to other options
- Use hyperlinks effectively so users control their pathway--and the level of detail they seek. The beauty of the Web is its "non-linearity."
Most importantly, when designing your interface, consider that non-readers will not always take the path you intended for them. So, save them from the garden path. Your overall site architecture should be consistent with a clear breadcrumb trail that lets them easily and intuitively get back on track.
Lesson 5: Search and You Shall Find... Hopefully!
In many a usability lab, you give a user a task to find or buy something and the first thing they do is to search for the Search box. That is, if they can't immediately find what they want on your homepage and, for some, even before they've given it a shot, they start typing what they want in your search box.
As such, Search is probably one of the most critical things to get right on your site. Here are some guidelines we've developed based on what users have told and shown us about what makes or breaks the Search function.
- If a search fails, offer tips on how to get better results--or point users to a path that keeps the dialogue going (e.g. a way to email you and tell you what they are looking for.) Witness the search term function Google added for misspellings: "Did you mean...?". Even if none of the choices they provide help you, this feature builds good will and encourages the user to try again.
- Make sure that the Browse and Search functions are clearly separated and mutually exclusive--or users will think they can browse and then refine their choices by searching within the category they've chosen
- Lose the Boolean logic--only statisticians want to define their search with strings. Offer a way to get a pop-up window that shows users how to enter words in the search box to optimize results--e.g. Chicago-style pizza in Los Angeles--and doesn't require them to leave the page they are on.
- There are many ways to define a label or word--rug vs. carpet, PDA vs. handheld, sofa vs. couch. Any way you say it should get results.
- Offer Advanced Search only after Basic Search has failed. Limit the options you offer so it isn't too advanced for the user.
- Make sure people can search on the dimensions that are most important. Focus groups are a good place to learn what criteria are most important to search on (e.g. for sweaters, which criteria are key: color, fiber, price, size, style or all of the above?)
- This isn't English class. All of us spellers, the good, the bad and the terrible, should still be able to get the results we seek.
- Make it clear what universe they are searching--your site or the entire Web.
As much as people love the Web for the access it offers to a huge database in cyberspace, most people will tell you that their Search experience is, on the whole, a very frustrating one. And, what is worse, when Search fails the user, they often equate the fact that they can't find something with the perception that you don't offer it.
The silver lining here is that you can truly differentiate your site and build a relationship with users by offering a Search function that provides them with the results they seek.
What we've seen consistently over time is that companies--small and large--who bring target users into the development process at key junctures get tremendous payback for their investment. The result:
- Sites that build lasting customer relationships
- Sites that "beat" the competition by being first in market to identify and address unmet needs with new features and content
- An accelerated development process
Most importantly, having users provide information about their needs, expectations, language and logic enables developers to think like users--vs. copywriters, graphic designers or programmers--and, in so doing, create powerful user-driven site experiences.
The concepts from this article are taken from a book by the same name: Back to the User: Creating User-Focused Websites by Tammy Sachs and Gary McClain, Ph.D. (New Riders Press: 2001). Also see Digital Web Magazine's review of Back to the User: Creating User-Focused Websites.
Tammy Sachs is President of Sachs Insights, a qualitative research consultancy whose goal is to bring a customer voice to Web site, software, and interactive voice response design. Tammy leads a staff of skilled strategists and researchers who conduct focus groups, usability research, ethnography, and site assessments. Prior to founding Sachs Insights, Tammy was a researcher and developer at Citibank working on delivery of online banking, yellow pages and shopping, way before the Internet. Prior to that she was a senior planner at Ogilvy & Mather. When not trying to make technology "fit for humans," Tammy searches the globe for antique dragons and listens to old time music.