1. Metrics Marketing Group conducted a study measuring the efficiency of opening a checking account online at 13 major financial institutions.
2. Using software to model expert user behavior, they found the average time was 6 minutes 49 seconds, with a range of 4 minutes 37 seconds to over 10 minutes.
3. Long completion times, multiple questions, and usability issues can negatively impact conversion rates and the customer experience on financial institution websites.
1. 1Metrics Marketing Group Phone: 877.332.9222 Web: www.metricsmarketing.com Address: 905 Corporate Way, Suite 250 Westlake, OH 44145
Efficiency, or the amount of time an expert user takes to complete a task, can have a
direct impact on the bottom line. A few seconds off of a call center task can save hundreds
of thousands of dollars in a year. But what is the impact of efficiency on consumers?
Metrics Marketing Group’s Research & Customer Experience experts wanted to explore the
power of efficiency testing and to understand the impact of efficiency on the customer
experience. Over the course of approximately two months in 2009-2010, 13 top financial
institutions were selected to measure the task time of opening a checking account online.
How Efficiency Was Measured
The efficiency of this task was calculated using CogTool , a software product developed
by Dr. Bonnie John’s lab at the Human Computer Interaction Institute of Carnegie Mellon
University. CogTool evaluates a design with a predictive human performance model to predict
the time it takes for an expert to complete a task – in this case, to open a checking account
online. CogTool’s modeling procedure is based on the Keystroke-Level Model (KLM) that
assigns time values to each physical and cognitive action, including reading text on a screen,
moving the mouse, moving the hand from the mouse to the keyboard, and typing.
Findings & Why It’s a Concern
Based on this study, the average online checking account application completion time was
6 minutes, 49 seconds, with the shortest being 4 minutes, 37 seconds and the longest
being over 10 minutes. This is of concern, as this evaluation is based on the prediction
of an “expert” user. The reality of the lack of efficiency of opening a checking account online
for non-expert users may be increased substantially. In fact, research shows that when
performing tasks online, the slowest 25% of users take 2.4 times longer as the fastest 25%*.
The efficiency of performing the task of opening a checking account online is only one
obstacle to overcome. Online usability issues are introduced when:
• Jargon is used
• Users do not know what to do next
• Users do not see site errors
• Users or do not know how to recover from site errors
• Site credibility and security concerns are introduced
• Users experience too many questions to complete
For the online checking account applications reviewed, the average number of total
application questions was 42, with 33 of them required, on average. According to
Forrester Research, 58% of visitors who experience usability problems on a Web site never
return. In addition, the average Web site can double its conversion rate by designing for
increased usability†
.
According to Forrester
Research,58% of visitors
who experience usability
problems on a Webs site
never return.
Improving User Efficiency to
Increase Online Account Openings
white paper
2. Metrics Marketing Group Phone: 877.332.9222 Web: www.metricsmarketing.com Address: 905 Corporate Way, Suite 250 Westlake, OH 44145 2
Sample financial institutions selected to evaluate:
BANK NAME TIME
REQUIRED
QUESTIONS
TOTAL
QUESTIONS
BB&T 4m 37s 31 38
Principal 5m 27s 37 43
US Bank 5m 32s 35 45
Bank of America 6m 1s 19 29
Fifth Third Bank 6m 54s 27 34
Wells Fargo 7m 38s 31 42
CitiGroup 10m+ 37 46
Table 1: : Predicted times are shown for a user to complete an online checking account application.
A full list of institutions evaluated with corresponding rankings may be available upon request.
Previous research indicates that users do not typically fully read disclosures. In addition, almost
80% of users scan the page instead of reading word-for-word online.**
The following chart demonstrates efficiency when time spent reading disclosures was removed.
Sample financial institutions selected with time spent reading disclosures removed:
BANK NAME TIME
Fifth Third Bank 3m 55s
BB&T 4m 37s
Bank of America 5m 8s
Principal 5m 27s
Citibank 5m 30s
US Bank 5m 32s
Wells Fargo 6m 2s
Table 2: Predicted times shown without the task of
reading disclosures. A full list of institutions evaluated with
corresponding rankings may be available upon request.
Previous research indicates
that users do not typically fully
read disclosures.In addition,
almost 80% of users scan the
page instead of reading word-
for-word online.**
white paper
Improving User Efficiency to
Increase Online Account Openings
3. Metrics Marketing Group Phone: 877.332.9222 Web: www.metricsmarketing.com Address: 905 Corporate Way, Suite 250 Westlake, OH 44145 3
What This Research Doesn’t Tell Us
Evaluating efficiency and task times is a helpful way to identify elements of a task that take
longer than others, or as a tie-breaker between design ideas. For interfaces where efficiency is
the most vital evaluation metric, CogTool can be a very helpful tool.
However, it is important to keep in mind that efficiency testing only goes so far.
It does NOT tell us:
• The “Whys” of User Behavior – We might know how long it will take an expert to
complete a task, but we do not know why it takes actual site users longer. We do not
know why a user might be unwilling to complete the application.
• Potential Risks in Design – We are unable to identify places where users will be confused,
unwilling or unable to convert, or have a negative experience.
• Actual User Expectations & Needs – Task times do not provide insight into what actual
users want to do, what they expect, what their goals are, how they go about selecting a
product in the first place, or what overall process this interface fits into. For example,
we can answer the question of how long it would take an expert user to complete the
online checking account application, but not if that expert would want to complete it.
• Non-Expert User Performance – CogTool and the KLM methodology are designed to
measure expert use. However, for most interfaces we interact with – whether it’s a Web site
or a DVR programming screen – actual users often do not use it frequently enough to
become an expert.
Utilizing other user research methods such as Usability Testing, Heuristic Evaluations, User
Personas, Card Sorting, Focus Groups, and Surveys are essential to fully evaluate the user
experience and maximize conversion.
white paper
Improving User Efficiency to
Increase Online Account Openings
4. Metrics Marketing Group Phone: 877.332.9222 Web: www.metricsmarketing.com Address: 905 Corporate Way, Suite 250 Westlake, OH 44145
Points of measurements
used should be based
upon the overall goals
of the interface, as well
as areas of the user
experience that specifically
need to be benchmarked
and measured.
For More Information:
Cathy Zapata,
VP Research & Customer Experience
T: 440.471.6026
F: 440.471.6010
czapata@metricsmarketing.com
About Metrics Marketing
Metrics Marketing Group is an analytics-driven database marketing and interactive firm.
Metrics leverages data management, analytic consulting, automated marketing, interactive
services, user research and direct marketing execution to improve marketing ROI for clients.
Additional Points of Measurement
CogTool provides just one point of measurement to evaluate a site’s usability in order to
improve conversion and the overall customer experience. To comprehensively measure
and benchmark the application process, additional points of measurement should be
evaluated, such as:
• System Usability Scale (SUS) Score
• Task Success / Failure
• Time to Complete Task (of actual site users)
• Steps to Complete Task
• Frustration / Delight
• Number of Errors
• Severity of Errors
• Learnability
• Memorability
• Eye Tracking Measurements, such as:
- Gaze Sequence
- Hit Rate
- Dwell Time
• Subjective Satisfaction
Points of measurements used should be based upon the overall goals of the interface,
as well as areas of the user experience that specifically need to be benchmarked and
measured.
References
1
http://cogtool.hcii.cs.cmu.edu/
* Reference: Jakob Nielsen’s Alertbox, May 15, 2006: http://www.useit.com/alertbox/performance_variability.html
†
Source: IBM
** Source: Sun
John, B. E. (2009). CogTool User Guide. http://cogtool.hcii.cs.cmu.edu/use-today/documentation-and-other-support
white paper
Improving User Efficiency to
Increase Online Account Openings