Wednesday, May 25, 2005

Caller tunes feature usability evaluation

I like caller tunes (or hello tunes) as much as I like powder coated metallic blue hand cuffs.

1. Poor Feedback - The earlier tones give clear feedback about the state of the call. In caller tunes, the caller does not get any feedback. They don't know if the called phone is ringing or busy. The result is quick and multiple hangups by callers.

2. Mental Model Mismatch - The caller expects a familiar ring or a busy ring. Any other type of ring is a surprise and frustration.

Mobile companies' revenues may go up. There are two reasons:
a. More missed calls to the called number as callers hang up in surprise. The called person then rings back the "familiar" numbers thereby increasing the mobile company's revenues.
b. Revenues from caller tunes feature subscription.

Mobile companies' revenues may also go down due to lost revenues in calls that never connected.

There may be a user population that will use this feature - there is enough literature available on gift giving behavior among young adults. However, I have not seen any ethnographic studies or usability tests published on this. Any links?

Age-old practices in the ‘New World’: A study of gift-giving between teenage mobile phone users
by Alex Taylor and Richard Harper presented at CHI 2002

The Gift of the gab?: a design orientated sociology of young people's use of "mobilZe!"
by Alexander Taylor and Richard Harper Journal of Computer Supported Cooperative Work (CSCW) 12(3), 267-296

Monday, May 23, 2005

Key usability issues (and solutions!) in customer service software

1. A very high percentage of outbound calls land up at an answering machine. Leaving standard messages on the answering machine is routine and takes a lot of user's time.

This task can be allocated to a computer -- an automated message delivery will drastically reduce the call time. In certain cases where the customer decides to pick up the call, the computer can seamlessly give the control to the Customer Service Agent (CSA).

2. Most calls are transferred to another CSA one reason or another. These calls are typically routed through a regular 1-800 number. Each transfer takes between 45 seconds and 2 minutes and this frustrates the customer. Some CSAs use this time for documentation by gaming the called system manipulating their placement in the queue. This helps CSAs to gain official work time. Customers also loose valuable time as the CSAs need to tell the called CSA about the manner in the customer was verified, the account number (or identity information) of the customer, and the reason for the transfer. The called CSA then starts with the routine "hello" that again frustrates the customer.

Find out customer queries that are regularly transferred and ensure that one CSA can answer all queries. This will enhance the customer experience and save expensive call costs. If the call must be transferred then ensure that:
a. The call bypasses IVR menu tree. This will ensure that the CSA doesn’t have to key in choices.
b. The customer verification information is passed transparently to the transferred CSA.
c. The transferred calls take precedence over the customers that are waiting now. This reduces CSAs idle time and reduces double queue waits for callers.

3. Organizations try to ensure that CSAs document everything that happened during the call. CSAs, on the other end don't want to be punished for not documenting. This ensures that whatever actions CSAs take using the system, they also document that. This documentation takes extra time, the CSA needs to remember what actions were taken (some are forgotten if the call was long).

Document only what was talked with the customer. Don't document the actions that were taken using the computer (the computer already knows about it!). The computer can automatically document this information for the next CSA to see. By doing all this considerable amount of effort and time will be saved.

4. Most customer service software is still Unix-based (remember Green Screens?). To navigate to a particular piece of information, the CSA must recall one of about eight hundred codes and key it in correctly. All this must be learned over three months.

Drastically reduce learning time by providing software application that does not require CSAs to remember unrelated codes. Provide software that does not need too much of navigation - for customer service applications 25 to 30 screen applications are enough. And, provide software that does not need much navigation as 80% work could be handled by the Main screen.

Sunday, May 22, 2005

We dont require costly usability labs

We may all agree that most big-ticket usability problems lie in poor navigation - insufficient scent of information, poor organization, poor understanding of tasks and subsequent task design, etc.

To test navigation, in most cases, we do NOT require an elaborate lab setup. There are many low cost testing techniques that can be effectively used to understand navigational problems. These techniques do not require costly lab set-up or monitoring equipment.

Lets look at the list of activities that common usability tests have, and see if they really require an elaborate setup.

Pre-Test

  1. Participant Screener - No
  2. Participant Consent Form - No
  3. Participant Demographics - No

Tests

  1. Test of Self Evidence / Structure Evaluation - No
  2. Label Wording / Task Wording - No. In cases where the users may have to initiate a task among many tasks, you may require a computer.
  3. Card Sort - No
  4. Task Completion - No. For hi-fidelity prototypes, you will need a computer for simulation.
  5. Task Efficiency/Performance - YES an elaborate setup is needed.
  6. Test of Brand Values - No
  7. Test of Affordance - No

Post Test

  1. Comments - No
  2. Likert Scale Rating of usefulness, ease of use, learning, efficiency, and satisfaction - No

Saturday, May 21, 2005

A tree on the left 2

  • A tree typically provides poor navigational clues - does not tell where the user is now, what is clicked, and what is open. To provide all these clues, a software developer spends a lot of energy.
  • A tree helps software developers put artifacts as in the systems model. This model is usually very different from the user's mental model.
  • The tree typically takes more than 20% of screen space. This amount of space for navigating from one point to another is a waste of precious screen area.
  • Software developers believe that each "leaf node" in the tree must be associated with a corresponding screen. These corresponding screens typically dont contain any content and are shown blank or with content that users never need.