Project Bento

How might we help users find critical information quickly?

Google's partner-facing intranet played a critical role to provide customer support agents with compliance policies and product information. Historically, the intranet contents have been written by multiple authors without guidelines, prohibiting users from quick information retrieval.

users and technology

USER - Customer support agents in partner organisations who need to follow the guidelines housed in the intranet to answer customer inquiries
TECHNOLOGY - Intranet

problem

Users of the intranet had to spend a long time to retrieve information critical to their daily work, negatively affecting overall productivity.

goals

To minimise the time spent on locating key information
To save cost while improving customer experience

research design

Contextual Inquiry

I sat with customer support front line teams to understand the end-to-end process in their real work environment. This informed the formulation of hypotheses. We found:

Users are required to read through lengthy narratives to find a piece of information they need
Article titles do not tell users what’s in each article as each article tend to include various information

A/B Testing

Hypothesis was that breaking down lengthy narrative articles into smaller chunks and adding matching titles can help users to locate the information more quickly. We tested this with the A/B testing. We found:

Contents in chunks required shorter time to locate information (-38%).
Search keywords used by participants tend to reflect customers language, contrary to technical terms used in existing contents.

Interviewing

To supplement data from A/B testing, I also asked participants about their experience. We found:

While experienced participants felt legacy content easy to use, novices found experimental content helpful.

Metrics

Working with Analytics and Finance, I also defined target metrics and calculated expected monetary value of the project. This was to gain resources to implement desired changes.

Hypothesis

Users can find information faster with Design A (left) than with Design B (right). See Figure 1 below.

Figure 1: Design A (left: experimental design) than with Design B (right: current design)

design Recommendations

Consider introducing a topic-based, targeted content strategy
Consider reflecting non-technical terms used by customers instead of technical terms

Result

Redesign enabled users to locate information faster by 16 % points, saving over £2.7million per year
The project team received an internal recognition award for a job well done.

RefLection

The heightened complexity of stakeholders across organisations, job levels, languages, cultures, and time zones was well-managed by adopting different communication strategies per group.
A small sample size due to limited resource availability was augmented by explicitly mentioning it as a limitation and by preparing three different estimates: optimistic, pessimistic, and neutral.

Voice of users

“[New design] Easier to search, takes you directly to the exact document that you're looking for.“

“[Old design] Especially when you’re talking to someone on the phone at the same time, it seems difficult to find what you need.”

team

Myself (Planning, recruiting, moderation/interview guide, moderation/interviewing, presentation, hiring, training, project management) [location: Japan/US]
UX writer (User research advisory) [location: US]
Analytics (Business metrics advisory) [location: Japan]
Finance (Cost reduction analysis) [location: US]
Content team (Implementation) [location: US]

Page updated

Google Sites

Report abuse