Users are in Africa and stakeholder is talking to me to design a system. It is a ecitizen system where users can apply for a passport, driving licence and other citizen related services. What usability methods i can follow to make it more usable.
I would first push back to the client on this and try to arrange access to some test users. This is critical to success, particularly in a scenario where you probably aren't familiar with the users' culture and have little idea about how they will respond to the system.
If this is not possible, I would:
- Try to learn as much as I can about the system users. In particular, I would try to find out about any other computer systems the users may use and how they work. Then you might at least have an idea of their expectations/ability.
- Use a different set of test users. Which will at least help you improve the design in a basic way, though not context-specific.
- Have someone try to stand in as a test user based on what you know about them.
- Plan for a bumpy roll out with problems, support issues, and requests for changes. Because even if you do your best it is still going to be very limiting.
Any user testing or user centered design is better than skipping it in general due the users being difficult to get access to. I worked in the defense industry in the past and while the distance to the user was pretty frustrating(geographical and security clearance), using the following alternatives gets you a step closer than giving up.
- Get a proxy - can you mail an ipad a laptop to someone with boots on the ground to hit your users? Can you get them to hit an remote desktop session from some community session? Can you just mail fax or send some paper protoypes to someone with boots on the ground near your users?
- Test other users. Can you get users similiar to these people? Maybe people who have went through the process in the past and now are living near you? If not test against any users NOT Familiar with the project, ie user testing against your QA people or Project Manager is silly because they have hints into what the workflows should do. Test against people who have never seen it before. Want something with a lower barrier of entry throw something simple on Amazons mechanical turk (ie should i do login screen 1, 2 or 3?).
- Ethnographic studies - [Read more about how to make the most of them here], however you may get this wrong and i would try to do the other things mentioned first.1
In short user testing is incredibly valuable. Hitting people who will use your product is the best, have similiar demographics is better, and testing anyone vs no one is better than nothing. Also the barrier of entry is not that great as One can get meaningful results from doing user testing on only 5 people.
This is a tough situation. Ideally, every product/application would have some type of research with target users that informs the design, but this isn't always possible. And it's much harder for a new product because there would be no existing analytical data. So, what to do then?
For the purposes of discussion, let's assume that you're working on a new product and won't be able to speak directly to target users.
Step One: In-depth Stakeholder Interviews Start with interviewing as many stakeholders as you can. The crux of these interviews will be to identify the problems they want to solve, not their idealized solutions. For example, since they want a digital ecitizen service, ask detailed questions around the reasoning, what they hope to accomplish, where the idea came from, etc.
At the end of your interviews, you should be able to accurately describe the problems they want to solve, who they want to target (audience), the business goals, and, possibly, what they feel success would be. More information on stakeholder interviews.
Step Two: Secondary Research Just because you can't talk directly to your target audience doesn't mean you can't find valuable information about them. In order to obtain good information, have a detailed view of who you are trying to target. For example, "everyone in Kenya who needs to update their identification" is far too broad to understand the nuances needed. Something along the lines of "18 - 30 year old Kenyans who have high familiarity with the internet - use [x] times per day/week" gets you in a better place to create a targeted design.
Research public information on demographic information, such as rate of internet usage, technological devices, as well as restraints, will help inform bigger design decisions. For example, if you find that your target audience is often in areas without constant internet access, the ability to fill out forms offline and upload them later may be important.
Step Three: Web Standards Understanding user and technological constraints (device, internet access, development stack, etc.) will help you determine the UI patterns needed. Applications are often heavy on inputs and forms. I suggest consulting LukeW - both his book (Web Form Design) and blog posts about mobile form standards.
Step 4: Test If you can find participants that might somewhat match your target audience, then fantastic. If not, still test your design. Using something like Usertesting.com would allow you to, at the very least, get different perspectives.
Just do usability testing with regular users rather than those trained in your particular domain. Ordinary humans uncover usability issues as well as your particular audience. Have them take on a role and perform common tasks as they think aloud.
There might be some types of issues that non-experts won't find, like flaws in labeling and proprietary procedures. But in most systems I've seen the main problems are those generic usability problems that non-experts can discover.