Should Onfido prioritize faster verification times or higher accuracy rates for our identity verification process?

Introduction

The trade-off between faster verification times and higher accuracy rates for Onfido's identity verification process is a critical decision that impacts user experience, security, and business outcomes. This scenario requires careful consideration of technical capabilities, user needs, and strategic priorities. I'll analyze this trade-off by examining the product context, identifying key metrics, designing experiments, and providing a data-driven recommendation.

Step 1

Clarifying Questions (3 minutes)

• Context: I'm thinking about the current market landscape for identity verification. Could you share any recent changes in competitive offerings or regulatory requirements that might be driving this trade-off consideration?

Why it matters: Helps understand external pressures and opportunities Expected answer: Increased competition or new regulations pushing for faster or more accurate verifications Impact on approach: Would influence prioritization of speed vs. accuracy based on market demands

• Business Context: Based on our revenue model, I assume faster verifications could lead to higher transaction volume. How does our pricing structure relate to verification speed and accuracy?

Why it matters: Aligns solution with revenue generation and customer value proposition Expected answer: Tiered pricing based on speed/accuracy, or flat rate regardless of performance Impact on approach: Would inform whether to offer different service levels or focus on optimizing a single offering

• User Impact: Considering our user segments, I'm curious about the distribution of use cases requiring immediate verification versus those that can tolerate longer wait times. Can you provide insights into this breakdown?

Why it matters: Ensures solution addresses needs of key user segments Expected answer: Majority requiring quick verifications for time-sensitive applications Impact on approach: Would lean towards prioritizing speed if most users need immediate results

• Technical: Given our current architecture, I'm wondering about the scalability of our machine learning models. How flexible is our system to handle potential increases in processing demands for higher accuracy?

Why it matters: Assesses feasibility of improving accuracy without significant infrastructure changes Expected answer: Some flexibility, but major accuracy improvements would require substantial updates Impact on approach: Would influence timeline and resource allocation for accuracy improvements

• Resource: Considering our team structure, I'm thinking about the balance of ML engineers versus front-end developers. Could you share our current team composition and any planned hiring?

Why it matters: Determines capacity for different types of optimizations (algorithmic vs. UX) Expected answer: Balanced team with plans to expand ML capabilities Impact on approach: Would inform whether to focus on algorithmic improvements or user experience enhancements