🖥️

Desktop Experience Required

The Opportunity Tree is a complex data visualization best viewed on a larger screen.

Evidence Report

Outschool — Search Strategy

Cross-source evidence analysis mapping the full opportunity space for improving search — from user pain points to prioritized solutions.

PeriodNov 2025 — Jan 2026

SourcesZendesk + Amplitude + Hotjar

Tickets Analyzed3,842

Search Sessions2.4M

Root pain points identified

Sub-problems mapped

Solution hypotheses

87%

Cross-source corroboration

~$4.2M

Booking revenue at risk (?)

Opportunity Tree

Improve the Search Experience

🔍

Improve the Search Experience

Increase search-to-enrollment conversion from 3.2% → 5.5%

First impression — Outschool doesn't seem to have relevant content

284 tickets44% bounce▲ +22% MoM

Confidence 91%

Supply constraint — catalog doesn't cover user intent

142 ticketsTop 50 queries: 18% have <3 results

💡

Identify supply gaps from zero/low-result queriesExplore

Browse/landing page doesn't inspire discovery

44% bounce, avg 8s on page

💡

Personalized landing with curated collections

💡

Social proof: "Parents in your area also enrolled in…"

Can't narrow down to see available classes matching my use case

612 tickets31% refine search 3+ times▲ +34% MoM

Confidence 94%

User doesn't know how to query (keywords vs. filters)

238 tickets67% use only text, never filters

💡

Autocomplete guiding users to form better queriesPlanned

💡

Query rewriting — parse topics from natural language

💡

Suggestions: "Did you mean 'Elementary math'?"

Users shown irrelevant results to their query

187 tickets41% leave after page 1 with no click

💡

Synonyms + semantic searchPlanned

💡

Implicit topic filtering (auto-detect subject)

💡

Topic-based retrieval taxonomy

User sees 0 results when results DO exist

94 tickets12% of searches return 0 results

💡

Query rewriting + semantic search fallback

Overwhelmed with choice — too many similar results

389 ticketsAvg 47 resultsStable

Confidence 88%

Too many similar results, no obvious differentiation

198 tickets4.2 avg clicks before enrollment

💡

Contextual filters / faceted searchPlanned

💡

Personalized filters based on learner profile

💡

Remove/flatten duplicate classes

Can't tell which option is "best for me"

79 tickets

Can't tell if these classes are high quality/trustworthy

276 ticketsLow enroll rate on unreviewed

Confidence 82%

No preview of teaching style or class energy

134 tickets

💡

Short video previews on class cardsFailed Q3 '24

💡

AI-generated class summaries from reviews

Reviews are insufficient or hard to parse

87 tickets

💡

Highlight "verified parent" badges

Methodology

How This Report Was Built

💬

Qualitative Analysis

3,842 Zendesk conversations processed via AI theme extraction. Each ticket categorized by pain point, user segment (new parent vs. returning), and proximity to abandonment.

📊

Behavioral Analysis

2.4M search sessions in Amplitude analyzed for query patterns, result interactions, filter usage, drop-off points, and paths from search → enrollment.

🔗

Cross-Source Triangulation

14 sub-problems grouped into 6 root causes. Opportunities ranked by evidence across qualitative + behavioral sources. Counter-evidence surfaced to prevent confirmation bias.

Desktop Experience Required

Qualitative Analysis

Behavioral Analysis

Cross-Source Triangulation

Imagine This, Updated Every Week