Darrin Fan · Growth operations work sample

Turn a stalled expert pipeline
into a measurable next step.

An expert network can grow while qualified experts still wait to start work. This sample shows how I would use an authorized pipeline export to find those delays, prioritize investigation, and test an onboarding improvement.

What the team could do with it

1. Review the delay. Open a prioritized queue, inspect contact → response → qualification dates, and confirm whether the missing activation reflects onboarding confusion, scheduling, lost interest, or an unrecorded event.
2. Capture the reason. Save a structured finding and prepare a clear next-step draft. Export the queue and notes for review; the sample sends no emails.
3. Check the economics and workload. Compare six channels using conditional conversion rates, completed-activation time, and recorded cost per activation. Use the planner to translate a target into contacts, hours, and modeled cost.
4. Test a fix. Rank one-stage conversion opportunities, estimate recovery scenarios, and define an onboarding pilot with a primary metric and guardrails.

A worked example

The fictional funnel is 1,799 contacted → 705 responded → 425 qualified → 234 activated. Using these rates, a target of 100 activations requires approximately 769 initial contacts and 143 labor hours. With 180 available hours, the model shows 37 hours of spare capacity.

At assumed effort of 2 / 8 / 20 / 10 minutes per contact / response / qualification / activation, $35 per labor hour, and $0.30 tools cost per contact, modeled acquisition cost is $5,237, or about $52 per activation. These are planning assumptions, not measured team performance.

Recovering half of the 153 stalled records would represent about 77 additional activations. This is an illustrative scenario, not a forecast or achieved result. Existing records may need no new sourcing, but investigation and onboarding still require time.

The experiment I would scope

First, confirm eligibility and reasons for delay. Compare current onboarding with a clear checklist plus one reminder, randomly assigning eligible experts and observing each for 14 days. Measure first-task completions / all randomized eligible experts; monitor task quality, opt-outs, and operator time.

A reference test of 37% → 55% conversion needs roughly 120 participants per arm at 80% power and two-sided 5% significance. A small stalled cohort is a feasibility pilot. Its actual baseline must be established before designing a powered test.

What this demonstrates—and what it does not

This is my week-one approach to funnel diagnosis, operational follow-through, and experiment design. It uses 1,803 invented rows across six channels and weekly cohorts. Validation excludes both copies of a duplicate and two invalid records, leaving 1,799 usable contacts.

The data does not establish causal channel differences, realized savings, or AfterQuery’s performance. The sample has no access to company systems. Uploads and calculations stay in the visitor’s browser; notes can be exported. Publication research is optional and does not verify practitioner qualifications.

How it could fit into a real workflow

Authorized ATS / Airtable export → daily validation → review newly stalled experts → human-approved follow-up. Confirm event definitions, ownership, data access, and a useful reporting cadence before connecting or automating a production workflow.