The Business of Usability

Preliminary Discussion

I am concerned with TCO; value add is killing products.

If we shave off 30 seconds per call we saved M$200, which is more than they spend on our products. Can this saving actually be realized, e.g., lay off a certain percentage of call center staff? Does it matter? Depends - if we make the sale, vs. ongoing ROI for a corporation.

One of our challenges is when the product we are hired to work on is only part of an entire business system. We have built in a needs analysis as a saleable service.

People don't buy enterprise applications for the UI.

We need to identify what the company needs from us: customers care about TCO; development teams care about deadlines, impressing their boss, increase quality, decrease UI bug count; identify the right products to work on.

Usability is the nexus of development, management, design - this can defocus you. We have a multidisciplinary team so people can contribute according to their skills: design, development, PM selling.

Our success is due to the fact that customers are increasingly satisfied with our website.

We target total cost of sales to our customers to make it easy for them to order things from us.

Our job is to make products easy to use.

I decided not to emphasize numbers because I couldn't win that discussion. I focused on design.

Our company does not put a premium on sales.

Our product is about creativity and exploration.

What is quality? Number of bugs? We work upstream, so our stuff never hits the bug tracking system. It's hard to count this and I don't want to make work for myself.

We enable the development organization. Developers will try to design, so we can support that.

We measure how many features each of my staff supports. I track weekly meetings staff attend. This might be perilous/not-useful.

80% of what is useful to our customers is not the UI.

We speak the developer's language.

The audience that matters to us is engineering, not marketing, so we need to cater to them.

We learn the language the people are speaking and we pick our metrics to address what our audience values.

It is not always obvious to others why better things are better. We have been hired by engineers, business analysis, and marketing people and we change our tune depending on who hires us.

We create stories/demos/visions that sell decision makers to commission/approve something.

We couldn't completely redo their development process, but we could fit UCD processes into their development process.


Metrics. Metrics are business-dependent.

There are stories that you tell your managers.

What does your audience value.

Task domain:

Metric purpose:

A story plus a metric is greater than the sum of its parts.

Metrics Identification Activity


Traditional approaches to ROI need to be extended to target audiences with specific messages.

Constituency Constituency Values What am I trying to achieve? Marketing Message to Constituency
Developer quality of work, bug count   UI spell checker
Manager confidence that they can meet product ship dates and quality   100% task completion rate without assistance
Customer tangible reasons; buy it to save money. How is my boss going to reward me for buying this product? How can I show I'm saving money?   Reduced call handling times.

Target high-return projects. What will make the company the most money or save the company the most money.

Avoiding vanity projects.

Where do corporate goals fit into this model? We assume corporate goals drive constituency values and what I am trying to achieve.

What metrics do I pick?


Mary Beth

Constituency Constituency Values What am I trying to achieve? Marketing Message to Constituency Our Company's Products
Engineering team leads Reduced bug counts

Standard development process

Do I have enough staff? Ratio of U-staff to work.   MathLab


Executive management  team of development Standard development process Are we making any difference?