Rahul
Vidyarthi

GlobalFair is a Series A B2B construction materials marketplace. I owned end-to-end product for the global buyer app — the interface through which international buyers discovered, evaluated, and sourced construction materials from suppliers worldwide.

GMV ($0.3M → $1.3M in 8 months) moved through three product levers:

• Discovery: Search and filtering rebuilt for B2B intent — buyers want to spec by material type, lead time, and geography, not just keyword
• Trust: Supplier verification signals, sample request flows, and RFQ conversion improvements
• AI Matching: Pattern & Colour Matcher — buyers upload reference images, AI returns matching products from the supplier catalog (20K+ images, 77–95% accuracy, cut match time from weeks to hours)

CSAT improved from 11% to 23% through a structured feedback mechanism (JIRA integration, Google Forms, Data Studio dashboard) that created a closed-loop improvement cycle across the product.

Construction material buyers often have a reference image — a tile they saw in a showroom, a surface from a competitor project — and need to find it or something similar in the supplier catalog. Text search was useless for this use case.

Built an AI-powered Pattern & Colour Matcher using image segmentation and visual feature extraction. The system analysed 20K+ supplier images, extracting visual signatures that enabled similarity-based retrieval from uploaded reference photos.

Accuracy improved from 77% to 95% through iterative tuning based on buyer feedback sessions — watching real buyers evaluate results told us exactly which visual dimensions to weight more heavily. Match time reduced by 15 days (from manual supplier communication back-and-forth to near-instant AI match).

Conversion rate increased 16% — buyers who found products via visual match had higher purchase intent than text-search arrivals.

PrintVenue customers bought business cards — then three weeks later, letterheads. Then again: visiting cards for a colleague. The purchase sequences existed in the data but weren't being exploited. The platform showed one-product-at-a-time with zero cross-sell logic.

Applied association rule mining (Apriori algorithm) across 36,000+ transactions. The resulting rules were clear: product A buyers very frequently also bought product B within 30 days. Those rules became the post-purchase recommendation engine.

Revenue impact: +15% from ML-powered recommendations alone. The recommendation engine worked because it was based on actual purchase behaviour, not category adjacency guesses. Average order value increased. Repeat purchase rate within 90 days improved.

MAXXIS is the world's 8th largest tyre brand — this is a distribution-complexity product problem at manufacturing scale. Product decisions connect factory output to a dealer network spread across 27 Indian states.

Key delivery:

• 500+ dealer registrations through Dealer App UI/UX improvements — reducing friction in the onboarding flow
• 30% reduction in order processing time through workflow optimisation
• 40+ distributor onboardings via Distributor Portal improvements with 100% warranty replacement adherence
• Led INR 15M strategic research initiative as single POC — 200+ data point competitor analysis across Michelin, MRF, CEAT, Bridgestone — informing win/loss feature strategy for the following year

The research initiative was significant: I was the sole PM owning a 15 million rupee strategic study with direct input to the product roadmap and sales feature prioritisation. That breadth of responsibility — from dealer UX to competitive intelligence to strategy — is what cross-functional product leadership actually looks like.

Situation: MAXXIS needed to understand its product-level competitive position against the top tyre brands in India — and translate that into a feature prioritisation framework for the product roadmap.

Task: Design and execute a comprehensive competitive analysis study as the single PM point of contact. Budget: INR 15M. Scope: Michelin, MRF, CEAT, Bridgestone across 200+ feature and market dimensions.

Action:

• Designed the analysis framework: product features, pricing, distribution reach, dealer satisfaction, warranty terms
• Managed the research execution across 200+ data points — dealer surveys, mystery shopping, product teardowns
• Synthesised into a win/loss feature matrix: where MAXXIS overinvests relative to competitors, where it underinvests
• Presented findings to product and commercial leadership; output directly shaped next-year roadmap priorities

Result: Research findings adopted into product strategy. Roadmap reprioritised based on gap analysis. Promoted from execution PM to strategic research lead as a result.