RAPID RESEARCH
Defining and implementing a rapid research program at Cisco to cut down study timelines from months to days.
Role: Lead Researcher, Program Manager
Company: Cisco
Timeline: ~1 month for model creation, <2 weeks for execution thereafter
Research Method: Varies
Research Type: Rapid
CONTEXT
The Experience Management and Insights (EMI) research team focused on the customer and partner experiences on Cisco.com and Cisco Community forum (~5M monthly visitors), collaborating with stakeholders to unify the troublehsooting, licensing, and eCommerce experiences across our many products and services.
Up to 2023, the EMI research team was conducting only broad, large-scale research projects taking a month or more each, covering the entirety of a complex topic (e.g., "Software Updates"). There was no method to accommodate time-sensitive research questions that needed fast answers, and/or smaller topics that did not justify months of work.
For example, what if we were building a product this week for launch next week, and needed a single question answered by customers before then?
GOALS
Define a rapid research model that can turn out insights as fast as possible.
Implement the rapid research model into existing workflows.
APPROACH
First I defined what would be appropriate for the rapid research approach:
Time-sensitive questions.
Bite-sized topics.
Lower-impact efforts.
Early, exploratory work to provide direction for a larger-scale study.
Follow-up work to provide clarity on questions raised in previous projects.
Based on this, I determined we could use the rapid research model for both generative and evaluative work, so long as it was small in scale.
Then I evaluated our existing research timeline and identified multiple areas that could cut down on time, including:
Recruitment: we spent a lot of time recruiting in our own channels, which could be saved if we utilized a third-party recruitment platform like UserTesting.
Scale: smaller-scale studies would mean less time spent writing plans, building tests, analyzing results, and presenting findings.
Execution: restricting rapid research to mostly unmoderated formats (e.g., unmoderated tests or surveys) meant less time spent sitting in sessions.
Templatizing: we were spending a lot of time each study rewriting screeners, demographic questions, introductions, and recruitment standards, all of which could be standardized and built into UserTesting and Qualtrics as plug-and-play templates.
Generative AI: Cisco's internal, GPT-powered generative AI tool empowered us to create study guides and recruitment copy at a faster pace than ever. Given its closed-system design, it is secure enough to take in confidential data, so can also assist in trend-spotting during analysis and writing copy for the report.
Next, I defined the timeline (see left) for a single rapid study, estimating <10 working days from discovery to handoff, depending on project scope.
I validated the model and timeline through a proof-of-concept pilot, then documented the entire process in our team knowledgebase and led a training session so that all our researchers could execute rapid research.
I also presented the rapid research model and proof-of-concept to our entire department, so most key stakeholders would understand the new research approach and knew it was a tool they could utilize to inform their work.
IMPACT
The rapid research model compressed a months-long process into 10 business days or less, becoming one our team's core research approaches. Since launching, this model has worked for unmoderated tests, surveys, and even moderated usability testing.
Since its inception, I owned our rapid research program and have led 12+ studies, which have impacted decisions* like:
the definition of global date, time, and time zone formatting standards.
the cancellation of two potential projects, saving valuable resources and time by halting initiatives that lacked customer demand and were deemed to have insufficient impact to justify the effort (see left for stakeholder feedback).
the naming of a new type of product documentation.
*Note: Cisco has strict confidentiality terms for customer data, so this case study is intentionally vague with respect to results and impact to ensure I meet those guidelines.