After 25 years of user experience (UX) consulting, I can easily state that pretty much everyone skips the critical step that guarantees success: task analysis. Some folks claim to do it, but what they do is not task analysis.
First, a definition: Task analysis is a step-by-step analysis of the users' task, from their perspective.
Many folks mistake use cases for task analysis. Use cases are system-centric, describing how the actors interact with the system, not how the user performs their tasks.
Another common mistake is to perform time and motion studies of the current design. The objective of a task analysis is to capture and understand the user's perspective of their tasks, not of the existing technology.
It's vital to remain solution agnostic and avoid documenting the process of using the current technology. Otherwise you'll end up merely automating current frustrations.
For instance, when we conducted the user research for an online florist, we didn't watch users buy flowers online. We visited brick-and-mortar flower shops to watch how buyers selected flowers. What we uncovered radically changed our design approach and made Proflowers a market dominating success.
The goal is to understand the individual steps in the task process that flow from trigger to outcome. You also need to identify the decision points in the task flow, as these are the critical points requiring specific knowledge by the user.
You'll find that users often lack the specific knowledge necessary to accurately make the right decision. Your design should avoid relying on some knowledge users will not likely have and instead find ways to either impart the knowledge they need or optimize the task based on the knowledge you know the users will have.
Define the Problem, First
The reason task analysis is so critical to the success of a product is that it accurately defines the users' problems. You can't accurately solve a problem until you just as accurately define it.
The typical mistake most product teams make is to ardently believe that they already know the users' problems. In my 25 years of experience on over 250 projects, not one single product has been focused on solving the right problem. All of those folks who thought they knew their users' problem were dead wrong.
Remember, if you don't know the problem, the best you can hope for is to solve the wrong problem, very well. An objective task analysis will help you realize just how far off your design really is.
The Ubiquitous Shopping Cart
Pretty much every ecommerce site includes a shopping cart, but few do it right.
Breaking down the task, users find something they like while shopping and then engage in one of several potential task sequences:
- They are looking for only one item, find it, and want to purchase it. If the user is looking for one thing, they may likely want to skip the shopping cart and go straight to the checkout sequence – express checkout.
- They are looking for only one item, but realize they need other things that go with that item and need to purchase the two items. (Suggestive selling – nothing goes better with your flower purchase than a box of chocolates.) Proflowers has done very well without a shopping cart and has a very successful suggestive up-selling conversion rate.
- They are looking for several items. They find something they want, flag it, and continue looking for other things. Once they find all their intended items, they purchase them. This is the only task that a typical shopping cart serves well. However, this is one of the least common tasks.
- They are just browsing. When they find something interesting, they need some way to flag that item or put it in a wish list. When they do decide to buy something, they may have several things in the wish list that they would like to keep in the wish list. The problem with using the shopping cart for a wish list is that the users have to remove some items in order to purchase the ones in the cart that they need right now, thus losing all the other wish list items. This increases shopper anxiety since it creates fear of losing their other saved items. This is one of the most common shopping tasks. You can easily see how a typical shopping cart fails at this task.
- They are just comparison shopping and need a way to keep track of various options and switch between them to determine which is the better option. Again, this is a common task and the shopping cart is ill equipped to support this, well.
The typical shopping cart is often used for many things other than just holding your intended purchases. The typical shopping cart is used more like a temporary storage device or shopper wish list than the "shopping cart" it was intended to be.
Your objective should be first to determine which scenario represents the most common usage pattern of your users and then to optimize your design for that pattern. You may not even need a shopping cart, at all.
For instance, Proflowers is one of the most successful ecommerce sites on the Internet and it doesn't have a shopping cart. Shoppers tend to buy one bouquet at a time and we incorporated suggestive up-selling as part of the checkout sequence.
Other online florist sites copy the general Proflowers design approach, but include a shopping cart. These sites aren't nearly as successful as Proflowers. You do the math.
Start With User Research
First and foremost, begin by conducting some observational user research, You can't effectively conduct a task analysis without actually watching the users.
If you just gather a team in a room and start analyzing the users tasks based on your knowledge, you will fail. You know too much about your product and are already biased by its approach. You will merely end up mimicking your current approach.
You only need to observe a few users to get an idea of their perspective. Avoid analysis paralysis and don't go into too much detail.
Which users to observe is always the question and most companies that do user research tend to research their existing customers. The problem is, those users have already drank your Kool-Aid and are biased by your existing design.
The best insights come from people who don't use your site. For that matter, you may want to observe users not using any site to solve their problem.
For Proflowers, we watched men buy flowers in flower shops, not online. Try to avoid biasing your research by watching users using existing technology.
When your task domain is unavoidably cluttered with existing technology solutions, making it difficult to observe the tasks independent of any technology, you might have to resort to conducting your observations within an analogous task scenario.
For instance, when designing a medical charting device for in-patient medical procedures performed by nurses, the nurse environment was polluted by various electronic devices and made it nearly impossible to capture their true needs. However, we were able to observe similar charting responsibilities performed by in-patient physical therapists. Though the medical procedures were different, the charting tasks were close enough to accurately portray the true needs of the nurses.
The five key things to look for in a user observation are:
- Trigger: What gets users to start their task.
- Desired Outcome: How they will know when the task is complete.
- Base Knowledge: What will the users be expected to know when starting the task.
- Required Knowledge: What they actually need to know to complete the task.
- Artifacts: What tools or information do they use in the course of the task.
One of the keys to a great design is being able to embed knowledge into the product instead of relying on the users to have that knowledge. Determining the knowledge you can expect users to have and the knowledge required in order to succeed identifies the knowledge gap that your design must bridge.
Diagram the Task Flow
Once you've completed some observations, use sticky notes to create a flow diagram of the observed tasks.
Start with a high-level overall task flow, then create more details task flows for each of the separate tasks. I use different colored sticky notes to represent different aspects of the user's task.
- Green represents the actions that users need to do.
- Yellow represents a step the system can do.
- Purple represents objects, tools, or information that the users need.
- Orange represents questions or issues about the task.
Basic task analysis only describes how users currently solve their problem, now, not how they could solve it. The real genius of the task analysis process is in optimizing the task. The real goal of task analysis is finding ways to remove or automate specific task steps, thereby helping the user achieve their desired outcome with fewer steps or less dependence on user knowledge.
Optimize the Task
Optimizing is the key to differentiating your site from the rest of the herd. The optimization process identifies the real opportunities and unmet needs in a task domain.
For instance, while everyone else is relying on a shopping cart to address the other tasks in a typical shopping domain, you may find that your shoppers need a wish list with comparison tools. Once you solve for that need, shoppers will be more likely to use your site, instead of the competitors.
The trick to optimizing the users' task is to start the process with the desired outcome and work backwards up the task flow. The goal is to find ways to eliminate user steps, or as we like to call it, turning the greens to yellows, getting the system to do more of the work for the user.
A key approach to automating some of the task involves applying your user personas to determine where the task requires information or knowledge that the users are not likely to have, but need. This is an opportunity to embed your company's knowledge of the task to aid the user.
Common methods to designing knowledge into your site are to include templates, intelligent defaults, or suggestions, or to optimize the flow along a "best practices" approach.
You know more than your users about what your site can do to help them. Optimize the site to guide them down a single best-practice approach.
A common mistake most sites make is to add more functionality or choices. In reality, users perform much better if they have fewer choices about product or functionality.
To paraphrase Antoine de Saint-Exupéry, your design is done not when there is nothing left to add, but when there is nothing left to remove. This is the objective of task optimization, to reduce user steps and decisions.
For instance, with the Proflowers site, most of the users were men buying flowers for women. Let's be honest, men know next to nothing about building bouquets, so asking them to build a bouquet (the de facto standard approach, back then) failed miserably.
The task analysis identified that men needed to pick a bouquet, but knew nothing about which bouquet was the right one for their occasion. By organizing the bouquets by occasion, rather than by flower type, we eliminated the need for the men to know which bouquet was right for their occasion. They merely needed to select their occasion, then choose from any of the bouquets grouped in that occasion page.
Although you may think you know your users and their tasks, just give this user research and task analysis approach a try and see if you don't notice any glaring differences between what users need and what your design provides. It takes a little practice, but not a lot of time.
Once you've completed some of the task analysis, walk through the task flow with your persona and current design to see where your design doesn't fit the user or the task. Don't be surprised if you find new, unmet opportunities that will leapfrog your competition and dominate your market.
If your new design looks just like everyone else's, you've done something wrong. Done right, this analysis always uncovers novel design approaches. I've used this process to create many market-dominating designs over the years, and none of them looked like the competitor designs.
Search and traffic sourcing are both crucial to luring shoppers to your website. In this article, "2 Successful Holiday Strategies for Online Retail", you'll learn how to use a two-pronged approach for your holiday search campaigns that combine top keywords with the best referral sites. Data in this article comes from SimilarWeb.