Any inputs on how to conduct an usability evaluation of a site (when you dont have access to user data) [duplicate]

Question

Possible Duplicate:
How to find out usability problems myself?

A lot of times I have been asked by friends to have a look at their sites or apps which they designed and give feedback from an usability perspective.Though I can give some general feedback on what confuses me and what should be the potential process flow or information architecture ,I am often at a loss as to what would be the steps taken to perform an usability evaluation when you don't have access to user data such as information from A/B testing,click data or heatmap information.

How do you handle situations like this?

Note: This might be a question more suited for Meta UX ,so please migrate it if needed.

Answer 1

This turned out to be a long question, and several considerations needs to be taken into account in order to give a proper answer, but I think the essence of the question is good and I believe that all of us have faced situations like this.

The problem here is two-folded. First of all that the project started out on the wrong foot - with no UX plan. How is it possible to get the project back on the right track? The second problem is about getting access to real users. What can be done without real users on hand?

The clients attitude

The first step is to bring the right point of view to your friend. To succeed, it is quite essential that the right person has the right attitude with respect to UX. So this is one of the first considerations. Is your friend just small talking, as in "Oh, you're an UX expert, hu? What you think of this spare time blog of mine?". Or is your friend really serious about the effort that should be put into this, as in "I really should have focused on this before, but one thing led to another and here I am - and I really want you to help me to deliver a killer solutions!".

To succeed, it is essential that client/contractor is in your boat.

---

First aid - DIY

Then, as a first aid "do it yourself" solution for the client, you should ask some judicious questions. By asking the right questions to your client, he is forced to make some considerations and thus be pushed towards an UX centered thinking. Having pointed his mindset in the right direction can be critical when design decisions are made.

Eg:
1) Who is the main user group?
2) What is the main purpose of this site/app?
3) What is the main task for the average visitor?
4) Can you describe the purpose of the system in one sentence?
...etc, etc

Give the right person the right mindset.

---

But what can I do

Well. There are tons of things that can be done - even without any users. You have a bunch of methods that are called "usability inspection methods". They are carried out by an expert (or a team) and they don't require any users - just a product (fully functional product, competing product, deprecated product, prototype, sketches, mockups, wireframes, etc).

Take a closer look at:

• Heuristic evaluations
• Cognitive walkthrough
• GOMS analysis (if you really want to bring out the scientific UX geek)

To conduct the first two, it is very handy to have some guidelines to work with. The GOMS-analysis is very interesting, but requires some training and experience.

Perform inspection methods to avoid the most common pitfalls.

---

Some things should be done anyway

It is quite common to perform several UX-related activities after the product is released. And there is no reason to believe that these activities become any harder to conduct. You just need to pick out the appropriate ones, adjust them and adapt them to the situation you're in.

Eg:

• Questionnaires
• Follow-up interviews
• Logging of user interactions
• A/B-testing
• Documentation of support and userfeedback
• User testing

Do whatever you would have done with any other released product.

---

OK, let's take one step back as well

The obvious problem in these situations is that there is no UX plan. The customer believes that UX is a wrapping that can be applied to the solution right before it is released. We know that this isn't true, but we also know that it is never to late to put UX on the agenda. Therefor, you still need to establish a plan. Conduct a quick-and-dirty analysis phase where you gather some basic information and establish some essential goals. In the eyes of the owner, the product might seem completed. But as an UX expert you need to do a proper job and enter the project with blank sheets and a very thoroughly prepared prototype :-)

Consider the first, released version to be the first iteration of the design phase and carry on towards the second version as you normally would during a second iteration.

Answer 2

Besides the heuristic evaluation (which is pretty good, if you have an objective/unbiased view), there are some other ways to evaluate without users.

One method is the "cognitive walkthrough", which you might already use; it's basically thinking about how a user would use the interface to reach a specific goal. You could think of some scenarios and play them through, simulating every step on the site or app and looking for possible problems during the whole scenario.

Depending on the complexity of the site or app, or a system or design in general and the time you want to spend on it, a literature review can be helpful. You could search for similar designs and userstudies that were done to evaluate them, or directly for studies that you would setup yourself if you had access to users. This gives you the results, without having to do a userstudy. Of course, this only works with systems based on known and researched elements.

Then there is "model based evaluation", which is pretty interesting. There are tools and models for measuring interface efficiency, for example information efficiency ((minimal information needed)/(information provided by user)). KLM (Keystroke-Level-Method) is a model based on human processing/perception, to measure a sort of input efficiency. It adds up the average times people need for specific actions, the exact actions for a scenario depend on how the interface is designed. The wikipediapage for KLM is pretty good. If you search a bit for model based evaluation, you will find other models and formulas you can use if you want to measure the actual efficiency of the designs.

Answer 3

From this point you can do these things,

• Start User testing by asking real users to use the app or website

• Insert some analytics (google ) and wait for minimum 2 weeks and use that data

• Make the competitors analysis and see how better are they

• Heuristics or Expert review by yourself

• Content analysis

Answer 4

Umm no users, no data mean no usability test to me. Why not do a heuristic evaluation? (http://www.useit.com/papers/heuristic/)