Specific results from specific engagements.

From zero growth insight to a compounding system that predicts churn and drives experiments

$272K–$505K annual revenue impact identified

Full analytics rebuild, churn prediction model, and experiment pipeline. 90% analytics cost reduction. Churn flagged 30–60 days before cancellation. 118+ decision-ready dashboards.

Read case study →

$2.5M+ in annual revenue opportunity found inside 40+ missing analytics events

20% → 35% activation rate improvement

The highest-value feature had 13% discovery and zero tracking. 40+ critical events were missing entirely. Full analytics rebuild surfaced $2.5M+ in recoverable annual revenue.

Read case study →

HIPAA-compliant PostHog architecture, 33 dashboards, and a 55% analytics cost reduction

$2,176 → ~$975 monthly analytics bill

Rebuilt PostHog setup from the ground up: HIPAA compliance, clean event taxonomy, 33 decision-ready dashboards, and a customer win-back system built in parallel.

Read case study →

PLG product DNA audit: tracking plan, 13 competitors mapped, activation strategy delivered

62 events in the tracking plan · 19 product surfaces audited

Full product DNA classification across 10 dimensions. Exposed 4 cross-dimension strategy conflicts. 62-event tracking plan and activation roadmap handed to the team.

Read case study →

3 behavioural retention flows protecting $105–155K in annual MRR from cancellation

40–50% save rate target across 3 churn archetypes

3 distinct churn archetypes identified from 295+ verified event types. Cancellation intercepted at the moment of intent — before the account is lost, not after.

Read case study →

8 PHI exposure types found in a live PostHog instance — including PHQ-9 scores and suicide risk assessments

99.9% reduction in high-risk autocapture events after remediation

A billing spike triggered the audit. What we found was 6.6M clinical events per week capturing depression scores, patient demographics, and medical record IDs. Documented, classified, and remediated in 10 days.

Read case study →

Migrated 906K events from Mixpanel to PostHog. Cut analytics costs 90%. Zero data loss.

90% cost reduction — $20K–$50K/yr to effectively $0

4 years of Mixpanel data, 228 event types, and 4 events containing PHI that couldn't migrate as-is. Full audit, HIPAA risk assessment, SDK mapping, and 6-week implementation plan.

Read case study →

114 events. 13 dashboards. 37 UX issues caught before users hit them.

90% analytics cost reduction vs enterprise alternatives

New platform launched with zero analytics. 80% of users never sent their first packet. 60% document signing abandonment. Built the full instrumentation stack from scratch — events, dashboards, UX audit — in 3 weeks.

Read case study →

Full GTM stack for a Design SaaS: pricing, positioning, personas, and onboarding

$2.6M+ projected Year 5 ARR · 23–67% pricing advantage vs. incumbents

Entered a crowded market with a 23–67% pricing advantage by targeting the right buyer. 4 validated personas, inbound + outbound motion, demo and onboarding playbook delivered.

Read case study →

Pricing architecture and value metric redesign for a retail investor platform with 17,600+ users

172% price expansion potential identified

Existing pricing was leaving 172% of addressable revenue uncaptured. 3-tier structure redesigned around the right value metric, backed by 62-event analytics instrumentation.

Read case study →

MongoDB Atlas migration for an HR platform carrying 73,717 registered users

30–50% infrastructure cost savings

Monolithic database restructured to MongoDB Atlas Flex Tier on AWS. Sub-300ms API response target. Cost cut without touching the user experience.

Read case study →

Custom ML analytics infrastructure replacing manual dashboards for an Amazon PPC platform

1.9x retention lift for active users · ~70% less manual analysis

2,100+ lines of custom Python replaced a reporting workflow that produced no actionable signal. Decision-ready dashboards built. Analysts moved from data prep to interpretation.

Read case study →

45% of signups stalled before activating — 28 missing events, 4 user segments, $2.5M roadmap

1.9x retention lift confirmed for early activators · discovery was the bottleneck

A platform with a confirmed retention advantage could only reach 38% of signups with it. A structured activation audit segmented the user base, identified 28 missing analytics events, and produced a phased roadmap to unlock the other 62%.

JTBD

Read case study →

From 1M members to a $6.5M ARR model — TAM, churn, and conversion redesigned

$24.5B TAM · 0.064% penetrated · $250–400K incremental ARR identified

An HR learning platform with 1M+ members had no credible market sizing, unknown churn drivers, and a weak free-to-premium conversion. Growth model built from scratch: TAM/SAM/SOM, 6 churn archetypes with mitigation, and a Year 5 ARR projection at $6.5M.

TAM/SAM/SOM

Read case study →

$100K ARR boost from fixing activation on one retention-driving integration

$100K ARR boost · framework scaled across multiple integrations

Users were not activating the integration that drove retention. Analytics audit, integration onboarding redesign, activation milestone design, and tracking build produced a repeatable integration activation framework.

Activation Analytics Integration Onboarding

Read case study →

Analytics audit and tracking to prove whether a key integration activation fix worked

$100K ARR boost tied to one activation fix · framework scaled across integrations

The retention-driving integration needed a measurable activation path. Analytics audit, activation milestone design, onboarding tracking, and a repeatable framework helped prove and scale the fix.

Analytics Audit Activation Tracking

Read case study →

$2,400 LTV gap confirmed, drop-offs located, analytics cost cut from $100K/yr to $110/mo

3 wks from no data to live cohort retention dashboard

A fitness SaaS platform knew onboarding was breaking LTV. No instrumentation, no proof, no prioritised backlog. A scoped onboarding diagnostic — B2B layer only — built the data foundation their previous $100K/yr quote had blocked.

JTBD Cohort Analysis

Read case study →

JTBD validation across 60 real sales calls found 85 jobs the framework didn’t know existed

27→112+ jobs · #2 priority feature corrected 45% down

A 27-job JTBD framework built from team recall was coded against 60 recorded sales calls. Persona distribution, feature priorities, and competitive moats were validated — or corrected — against what the data actually showed.

JTBD Kano

Read case study →

1M+ members, $529K ARR, 95.8% never converted — TAM sized, 6 churn drivers found

$24.5B TAM validated · $250–400K conversion opportunity found

Strong community engagement, weak monetisation. Seven research streams across market sizing, churn analysis, ODI scoring, and AI architecture — built a roadmap from $529K to $6.5M ARR with every assumption sourced.

JTBD ODI Cohort Analysis

Read case study →

60 sales calls. 85+ jobs. Feature #2 priority corrected 45% down.

3× jobs found vs. framework · persona distribution off by 2–3×

A 27-job JTBD framework built from team recall was validated against 60 recorded sales calls. Manual data entry confirmed at 100% (claimed 88%). The #2 priority feature had only 43% frequency — half of what the team believed.

JTBD Sales Call Analysis

Read case study →

4 personas. 4 onboarding paths. Goal-based routing built from 8 seller interviews.

5-screen system branching modal · 5–30 min setup time by path

Generic onboarding was failing four radically different buyer types. Eight seller interviews across experience levels mapped motivation, ability, and anxiety per segment — producing guided vs. expert path routing from a single entry point.

Persona Research BJ Fogg Model Kano

Read case study →

9.4M NPI records, 6 segments independently sized — $1.41B TAM confirmed bottom-up

$180M TAM overstatement corrected · 6 segments sized independently

A single global market figure was rebuilt into a six-segment bottom-up model. NPPES full-dataset analysis corrected the multi-location practice count from 35,000 to 15,047 — with a source citation on every claim.

NPPES Census Bureau AMA Data

Read case study →

From 80% onboarding drop-off to measurable activation. 114 events, 37 UX blockers fixed.

90% analytics cost reduction · funding the activation system

A healthcare platform launched a new product with no activation measurement. 80% of users never sent their first packet. Built the full activation funnel from scratch — milestone definition, event taxonomy, funnel dashboards, and 37 UX blocker fixes.

Read case study →

7 critical gaps found in onboarding architecture. 40+ events, 3 churn archetype flows built.

40–50% target cancellation save rate across 3 archetypes

A healthcare forms platform redesigned onboarding for a new pricing model. Architecture review uncovered 7 critical gaps, including a 7-14 day SMS delay and missing persona routing. State-machine orchestration and cancellation prevention flows delivered.

Read case study →

Activation funnel rebuilt from 14% to 42% — $1.2M ARR recovered in 90 days

14% → 42% activation rate · 6 weeks to 8 days time-to-value

A B2B analytics SaaS was losing 86% of signups between registration and first value. 14 activation steps compressed to 5. Three persona-specific paths delivered. $1.2M ARR recovered from accounts that never activated.

Persona Analysis

Read case study →

Churn flagged 45 days before cancellation — $380K annual revenue protected with ML prediction model

18% → 9% monthly churn · 87% model accuracy · 3 intervention playbooks

A $4M ARR B2B SaaS was losing 18% of revenue to churn with no leading indicators. Built a Python/SageMaker churn prediction model using product usage signals. 45-day early warning gave the CS team time to intervene before cancellations happened.

Read case study →

Single pricing tier cost the company $840K/yr — Van Westendorp study found 3 distinct segments

$840K annual revenue lift · +22% average deal size · 3 tiers

A project management SaaS with $8M ARR was charging every customer $79/mo regardless of usage. Willingness-to-pay analysis across 200 users revealed 3 distinct price segments. Restructured into Basic/Pro/Enterprise without increasing churn.

Van Westendorp Feature Segmentation

Read case study →