Background
Vendr is a SaaS procurement platform that helps companies negotiate better deals on software purchases. In late 2024 we launched a redesigned marketplace experience where visitors could request Price Estimates or Price Checks for software they were considering.
The company needed to understand:
- How visitors discovered the marketplace
- What expectations they brought with them
- Whether the experience met their needs
- What prevented conversion to paid customers
The Challenge
Despite launching an improved marketplace experience, Vendr faced a classic product-market fit challenge. New visitors were requesting Price Estimates, but the team lacked insight into what happened next. Were visitors getting value? Did they understand Vendr’s capabilities? Would they return?
The marketplace team needed rapid, actionable insights to guide immediate improvements. They needed to know not just what users did, but why they did it—and crucially, what would make them come back.
This presented several research challenges:
- Timeline pressure: The team needed insights quickly to inform Q1 2025 roadmap decisions
- Scale requirements: With 115 potential participants, traditional analysis would take weeks
- Depth vs. speed tradeoff: The team needed rich qualitative insights without sacrificing velocity
Research Innovation: AI-Accelerated Analysis
Rather than choosing between depth and speed, I leveraged Google’s NotebookLM to compress analysis time by approximately 50% while maintaining research rigor. This approach allowed me to:
- Process hundreds of pages of interview transcripts efficiently
- Identify patterns across 26 user sessions systematically
- Create a democratized research asset which let anyone on the team query the data themselves through NotebookLM
- Maintain the human judgment necessary for strategic recommendations
The AI-enhanced workflow didn’t replace the researcher’s role. It amplified it, freeing cognitive resources for synthesis and strategic thinking rather than mechanical transcript review.
Problem Statement
Marketplace visitors didn’t understand what Vendr could do for them. They arrived with varied and often inaccurate expectations, received Price Estimates that sometimes lacked specificity, and were left uncertain about their next steps.
The core product question was: How might we transform first-time marketplace visitors into engaged, returning users who understand and trust Vendr’s value?
Our driving problem statement became:
How might we…
- Help visitors form accurate expectations about Vendr’s capabilities from their very first interaction
- Build sufficient trust for visitors to share proprietary procurement data
- Deliver Price Estimates that feel valuable and actionable
- Create clear pathways for visitors to engage deeper with Vendr’s services
- Design an experience that makes returning to Vendr feel natural and valuable
The research needed to answer: What’s preventing visitors who get value from a Price Estimate from becoming advocates and repeat users?
My Approach
Working closely with Vendr’s product team, I designed and executed a mixed-methods research program that balanced qualitative depth with quantitative benchmarking.
Research design decisions:
- Recruited 26 participants from ~115 recent Price Estimate requestors
- Conducted 30-minute semi-structured interviews exploring their end-to-end journey
- Embedded Product-Market Fit (PMF) scoring and experience rating questions within conversational interviews
- Used retrospective protocol to capture authentic reactions rather than hypothetical preferences
- Analyzed discovery patterns, expectation formation, value perception, and intent to return
Key methodological choices:
- Rapid recruitment via LinkedIn: Bypassed email noise by reaching out directly on LinkedIn, improving response rates and avoiding sales email streams
- AI-enhanced analysis: Used NotebookLM to systematically code patterns across transcripts while preserving nuance in strategic synthesis
- Democratized insights: Created a queryable NotebookLM project so any team member could explore the data themselves
- Dual-format deliverables: Produced both a comprehensive written report and a presentation-ready readout for different stakeholder audiences
The approach emphasized speed without sacrificing rigor—getting actionable insights into product leaders’ hands within weeks rather than months.
Results
The research revealed critical insights that reshaped Vendr’s marketplace strategy:
The positive signals:
- Visitors derived genuine value from Price Estimates and expressed clear intent to return to Vendr for future software purchases
- Organic discovery mechanisms were working—most users found Vendr through Google search, LinkedIn, and word-of-mouth
- The marketplace experience itself was generally smooth and intuitive
- Strong potential for viral growth through user recommendations
The critical barriers preventing conversion:
- Unclear value proposition: Visitors fundamentally misunderstood what Vendr could do for them, forming inaccurate expectations about services offered
- Trust deficit: Users hesitated to upload proprietary contracts and quotes due to unclear data usage policies and concerns about how Vendr monetizes
- Inconsistent data quality: Many Price Estimates showed overly wide ranges or lacked product specificity, reducing perceived value
- Ambiguous next steps: After receiving estimates, visitors were uncertain what to do next with Vendr—no clear conversion path
Quantitative benchmarks:
- Conducted product-market fit (PMF) scoring across the participant cohort
- Gathered experience ratings on key interaction points
- Measured intent to return and likelihood to recommend
Strategic recommendations delivered:
- Set accurate, aspirational visitor expectations through consistent messaging across all touchpoints
- Build trust through transparency about business model and data protection
- Improve journey signposting with clear next-step guidance
- Enhance Price Estimate specificity and actionability
- Double down on nurture campaigns to maintain engagement
- Consider supporting visitors earlier in their buying journey (solution research phase)
The research directly influenced Q1 2025 product priorities and established a replicable AI-enhanced research workflow for future studies.
Impact & What I’d Do Differently
The Vendr marketplace research demonstrates how AI-enhanced workflows can accelerate insight generation without compromising research quality. By compressing analysis time by ~50%, I delivered actionable recommendations within weeks of data collection—fast enough to influence immediate product decisions while maintaining the depth necessary for strategic guidance.
What made this effective:
- Speed to insight: Rapid recruitment and AI-accelerated analysis meant findings reached decision-makers while the product questions were still fresh
- Democratized access: The NotebookLM project allowed non-researchers to explore the data, building organizational research literacy
- Multi-stakeholder design: Dual deliverables (comprehensive report + executive readout) ensured insights reached both tactical and strategic audiences
What I’d explore next time:
- Earlier integration with product metrics: Connecting qualitative themes to quantitative funnel data would strengthen the business case for recommendations
- Broader recruitment channels: While LinkedIn worked well, expanding recruitment would test whether patterns hold across discovery mechanisms
- Longitudinal follow-up: Tracking whether participants return after receiving recommendations would validate assumed intent
The real value wasn’t just the insights themselves…it was establishing a scalable, AI-enhanced research practice that Vendr’s team can replicate for future questions. When you can compress analysis time by half while maintaining rigor, research becomes a continuous product input rather than an occasional deep dive.
Want to see the full writeup for this case study? Contact me for access.