04 September 2009

Measuring the User Experience

This weeks post is a review of the book Measuring the User Experience by Tom Tullis and Bill Albert. From time to time other book reviews will follow.

Why a book review

The current state of books on UX is deplorable. Many UX books can’t make up their mind if they are about a given subject or the UX world according to Garp. Just looking at my UX bookshelves, I notice there are, for example, many books with authors who have a narrow or focussed expertise. These authors write books supposedly over a narrow subject, which they sustain for about a chapter or two before they deteriorate into their own homemade version of the User Centered Design process that has little if anything to do with the subject of the book they intended to write. The result is a book with grains of truth in a stew of platitudes. A review of just three books one claiming to be on prototyping, one on designing and another on UX communications, reveals that all of these books cover more or less the same material such as user research, task analysis, persona’s and prototyping; but it does it in such a way that they use both conflicting terminology and conflicting methods.

My more ideal UX books are those on a subject and stick to that subject. They explain their topic in a way that is process independent so that they can plug into whatever processes companies or organizations utilize. The fact of the matter is that no two organizations adopt the same software development process. What they all have in common whether they are called agile or waterfall, iterative or serial, is that they are all machiavellian. Therefore if a book's material cannot fit into the current machiavellian software development processes, then the book is largely worthless; even if entertaining (though probably not as entertaining as E.M. Forester).

I think one of the best services I can do then is to help people navigate around these literary cataracts and start a series of book reviews. These reviews will try and highlight the best of the UX literary corpus.

Measuring the User Experience: Collecting, Analyzing, and Presenting Usability Metrics by Tom Tullis and Bill Albert

I want to start with one of the brighter lights in our industry Tom Tullis. I have often wondered why he had not earlier written a book, given the high quality of contributions he has made to our profession. Well the wait is over.

It's true it is a book on usability metrics. Now I realize there are some people who hate metrics. These people particularly hate any accountability for their design work. I can’t tell you the hate mail i received, even from large design firms, when as Interactions Editor we did a special issue on measuring usability that was guest edited by Jeff Sauro. Well, I purchased Measuring the User Experience (MUX if you will) expecting a more thorough version of that special edition that went into the statistical significance of usability testing. I was in for a very welcomed surprise: this book does not just cover summative usability statistics but many different ways to collect user experience metrics and the also discuss proper analysis techniques.

The book empowers the user to make the right decision regarding what methods you can use and what you can expect the metrics to be able to tell you or not tell you. As the book states metrics can help you to answer questions such as:

  • Will the users like the product?
  • Is this new product more efficient to use than the current product?
  • How does the usability of this product compare to the competition?
  • What are the most significant usability problems with this product?
  • Are improvements being made from one design iteration to the next?

This is a refreshing change from just looking at time on task, error rates and task success rates. Though of course these play a role they are but ends to the means of answering these larger questions. Furthermore, the book also points out that there is also an analysis step that can greatly alter the seemingly obvious findings.

I cannot tell you the amount of time and money I have seen wasted as perfectly reasonable and wonderful user research was conducted, only to have its results obfuscated and mutilated beyond use. This book will not just enable the usability tester or researcher to avoid such mistakes it also empowers a project manager to see to it that a development project designs the solid usability study that will fit in the goals and needs of the development team.

In their discussion of designing the right usability study. The authors guide you in choosing the right metrics.

First you need to establish if the goal of your study is what the goal of the user’s are. Then on that basis you can look at which metrics, the authors identify 10 common types of usability studies:

  1. Completing a transaction
  2. Comparing products
  3. Evaluating frequent use of the same product
  4. Evaluating navigation and/or information architecture
  5. Increasing awareness
  6. Problem discovery
  7. Maximizing usability for a critical product
  8. Creating an overall positive user experience
  9. Evaluating the impact of subtle changes
  10. Comparing alternative designs

Then, a key issue they discuss is looking at the budgets and timelines, aka, the Machiavellian business case for the study. Then you can tailor the type of study: how many participants, will it be tests, or reviews or focus groups or a combination thereof.

In the conduct of these studies it is also important to track the right metrics. Tullis and Albert identify the following types of metrics:

  • Performance Metrics -- time on task error rates, etc.
  • Issue-based metrics -- particular problems or successes in the interface along with severity and frequency
  • Self-reported metrics -- how a user can report their experience with questionnaires or interviews
  • Behavior or physical metrics -- facial expressions, eye-tracking etc.

It handles these metrics as they should be as part of an overall strategy not favoring one over another as being innately superior. All too often usability testing consultants are one trick ponies, prisoners of whatever limited toolset they happen to have learned.

This book allows the user to assemble all the needed metrics across types to achieve a more holistic view of the user experience, or at least sensitize them that they are not looking at the whole picture.

What is also amazing is the focus and discipline in the book. I think many other authors would not be able to fight the temptation to then expand the book to include how to perform the different types of evaluations, usability tests, etc. These authors acknowledge there are already books that cover these other related aspects and keep their emphasis purely on the subject matter of their book: measuring the user experience.

Yes the book does also get into statistics and evens hows you how to do simple straightforward statistical analysis using that panacea to the world’s known problems’ excel (but that is next week’s topic).

And just in case your wondering the usability score for Amazon is 3.25, while Google’s is 4.13 and the Apple iPhone is a mere 2.97. While the web application suite I just finished designing got a perfect 4.627333.