The First 100 Customers Course #13: How Judi Health Turned Union Trust Into $10M ARR With Zero Marketing

TL;DR

Judi Health began as Capital Rx in a market where a failed vendor could interrupt access to medicine. It had no customer references, licensed much of its first software stack, and showed a visible administrative fee while large competitors could make that line look like zero. Co-founder AJ Loiacono says the company still reached roughly $10–12 million in annual recurring revenue after about two years, with zero marketing spend.

The mechanism was not hidden growth hacking. Small union plans had fixed budgets and a strong reason to challenge rising costs. Founder-led data analysis earned the first trust. High-touch service made customers willing to act as references. Those references reportedly moved the company from accounts with roughly 500–1,000 covered lives to 4,000, 8,000, and then a 30,000-person Fortune 500 inbound.

This course teaches you to build the same kind of reference ladder. It does not claim Judi Health acquired exactly 100 customers through this motion. The exact first buyer and early customer count are not public.

What you will build

You will leave with one narrow account list, one evidence-review offer, one service scorecard, one reference-permission ladder, and one threshold for moving upmarket. The system has five stages:

1. Choose the smallest buyer with urgent pain.
2. Prove one expensive fact in the buyer's own data.
3. Make service measurable before asking for advocacy.
4. Turn a successful account into permissioned proof.
5. Increase account size only when the prior proof passes.

Who this is for, and who should not copy it

Use this system if your product replaces a critical workflow, requires security or legal review, handles sensitive data, or creates material switching risk. In those markets, a buyer may like the product and still be unable to approve an unproven vendor.

Do not copy the “zero marketing” headline for a low-cost self-serve tool that buyers can test safely. If trust is not the constraint, refusing to create reach may only hide weak demand. The useful question is not “Should I spend nothing?” It is “Which proof is blocking the next sale?”

Case snapshot

Stage	Buyer or result	Evidence	Limit
Start	Small self-insured plans	Roughly 500–1,000 covered lives	Founder-reported account sizes
Early wedge	Union plans	Fixed budgets plus high service need	Exact first buyer is unnamed
Reference ladder	Larger plans	4,000, then 8,000 lives	No conversion rate disclosed
Pull signal	Fortune 500 inbound	About 30,000 people	Arrived after the early reference work
Milestone	End of year two	About $10–12M ARR	Founder-reported, not audited

Capital Rx was founded in 2017. The company later expanded beyond pharmacy benefits and adopted the Judi Health brand. Its company history documents that transition. This article uses the current name for recognition and the Capital Rx name when describing the early company.

The model: evidence → service → reference → larger account

A cautious buyer does not accept one large claim. The buyer crosses a sequence of smaller risks. Capital Rx first had to prove that its view of pharmacy costs was real. Then it had to administer a plan without creating member pain. Only after delivery could a customer become proof for another account.

That gives you four linked jobs. Evidence makes the problem credible. Service proves that a young vendor can deliver. A reference lets a peer validate the claim. A larger account becomes reachable because its risk committee now has relevant proof.

This is not a viral loop. It is a proof chain. The output of each account is not merely revenue. It is one stronger reason for the next buyer to say yes.

Step 1: choose the buyer whose pain is stronger than your reputation gap

Large employers looked attractive, but they had more to lose from a new vendor. Even a founder's family connection could not bypass the request for a reference from a company of similar size. The early reachable segment was smaller self-insured plans, particularly unions.

Loiacono explains that union plans operate with fixed budgets and care deeply about member service. A five-percent increase cannot always be absorbed. Benefits are also a meaningful part of member compensation. That made cost visibility and service urgent enough to consider a new provider.

Your narrow-buyer worksheet

• Costly event: This buyer loses ______ when ______ happens.
• Why now: The buyer must act before ______.
• Proof access: The buyer can give us ______ records to test the claim.
• Switching risk: The buyer fears ______.
• Reference value: A win here is relevant to ______ next-stage accounts.

Pass condition: five named accounts share the same costly event and can provide the same kind of evidence. Stop condition: you chose the segment only for famous logos, or your only reason for urgency is “AI is changing everything.”

Step 2: sell a small true statement before a large promise

Capital Rx charged a visible flat administrative fee. Incumbents could show zero in that line while earning through other parts of drug spend. A buyer scanning a bid could therefore see the challenger as more expensive even when the challenger argued that net cost was lower.

The founder's wedge was not a slogan about transparency. He used claims data to point out costs the buyer had not seen and connect them to the contract. Independent reporting later verified the business model. Forbes reported that Capital Rx charged flat administrative and clinical fees. A U.S. Senate hearing record also describes its transparent benchmark, pass-through model, and flat fees.

Make your first offer an evidence review:

In your last ______ records, we will identify ______, quantify the current cost, and show the contract or workflow causing it. If the evidence is absent, we stop.

For an AI SaaS, the input might be 100 support tickets, 50 sales calls, or one month of failed workflows. Run the first reviews manually. The objective is to learn whether the same expensive condition repeats.

Metric: evidence-positive reviews divided by completed reviews. Decision gate: if fewer than three of the first five reveal the same problem, narrow the ICP or change the wedge before adding product or outreach volume.

Step 3: make service a distribution asset

The early product relied on licensed claims, authorization, reporting, and printing systems. Product novelty alone could not overcome vendor risk. Capital Rx invested in service instead. Loiacono says the company built a domestic call center and targeted more than 97 percent first-call resolution. Representatives could stay on difficult calls rather than optimize for a short average handle time.

The team also traveled to customer reviews. The founder says he remains the account manager for two old accounts because direct contact keeps feedback raw. This is expensive work. It matters because the plan sponsor hears when members cannot get help. Reducing that complaint noise gives the buyer a concrete experience to describe to a peer.

Your service proof scorecard

Measure	Target	Why it creates trust
Time to first response	Define before launch	The buyer knows the issue was seen
First-contact resolution	Improve every week	Fewer reopened problems
Critical open issues	Zero before advocacy ask	No testimonial over unresolved risk
Verified outcome	One before-and-after result	The reference can be specific
Reference permission	Explicit scope and date	Sales knows which proof is usable

Pass condition: the customer can explain what changed and why the team trusts you. Stop condition: any critical issue is unresolved, or the outcome exists only in your internal dashboard.

Step 4: build a reference ladder, not a logo wall

The founder reports that the first small cases created references. Those references made 4,000-life and 8,000-life accounts possible. A Fortune 500 company with roughly 30,000 people later arrived after seeing a post and hearing positive market feedback.

A logo on a website is weak proof. A relevant operator willing to answer a cautious buyer's question is stronger. Ask for permission in levels:

1. A qualified buyer may speak with you privately.
2. We may publish an anonymized result.
3. We may name the organization and approved result.
4. You may join a panel, customer session, or recorded conversation.

Record which level the customer approved and when. Never turn a private reference into a public case study without new permission.

Decision gate: move one risk level at a time. A 500-life reference may unlock 4,000. It does not automatically remove the concerns of a 100,000-life buyer.

Step 5: put cash into the constraint, not the fashionable channel

Capital Rx spent zero on marketing, sponsorships, paid search, and conferences for two years. It still paid for founder selling, data work, a service team, travel, customer reviews, compliance, and licensed software. Zero marketing did not mean zero distribution cost.

Use this allocation rule:

• If buyers doubt the problem, spend on evidence.
• If buyers doubt delivery, spend on service.
• If buyers trust both but never discover you, spend on reach.

This distinction also prevents a false causal claim. The founder dates the major regulatory transparency tailwind to 2024, years after the first milestone. It may explain later acceleration, not the first two years.

What failed, stayed slow, or cannot be copied

• Price presentation: a visible fee looked worse than a competitor's zero until the team proved net economics.
• Reference gap: large accounts could like the thesis and still fail vendor risk checks.
• Product constraint: the first stack relied on third-party systems and was not the final platform.
• Manual cost: founder sales, long calls, travel, and on-site reviews do not scale cheaply.
• Missing channel ledger: the host infers that the first $10 million was outbound. The founder confirms zero marketing and a reference motion, but does not enumerate every initial lead source.

The original business plan also targeted 70 percent greater operating efficiency than the large incumbents. Treat that as a founder-reported planning target, not audited proof.

Your 7-day implementation plan

1. Day 1, choose: define one buyer, one costly event, one disqualifier, and one reason the buyer can tolerate a young vendor. Output: a one-page ICP note.
2. Day 2, list: find 25 accounts matching that buyer. Keep them separate from every broad prospect list. Output: 25 named accounts.
3. Day 3, design: write the evidence review, required data, result format, and stop condition. Output: a one-page review template.
4. Day 4, contact: approach five operators with one ask: permission to test the claim on a small data sample. Output: five reviewed messages.
5. Day 5, deliver: run one review manually. Record baseline, finding, uncertainty, and missing evidence. Output: one evidence memo.
6. Day 6, operationalize: create the service scorecard and reference-permission levels before onboarding. Output: two checklists.
7. Day 7, decide: repeat, narrow, or stop. Do not increase outreach volume when the evidence is unclear. Output: one written decision.

Lead Scorer lab: run the reference ladder with approval gates

This implementation fits narrow cold outreach with heavy qualification. It does not fit a scraped list sent at scale. Lead Scorer can organize the manual research and drafting loop. It cannot guarantee customers, manufacture evidence, or send without your review.

Phase 1: store the claim boundary

Use the ICP & Offer Context skill to save the painful event, buyer, disqualifiers, proof you may cite, and claims you may not make. Required inputs are the evidence review, its data requirements, and the account-size ceiling for the first cohort.

Define one ICP for [product]. The costly event is [event]. Disqualify [conditions]. We may claim [verified proof] and must not claim [unsupported outcome]. The first cohort must stay below [risk ceiling].

Pass condition: a reviewer can reject an account without guessing. Stop condition: “any B2B company” remains in the ICP.

Phase 2: build and score before enrichment

Create a dedicated list of 25 accounts for one buyer type. Do not mix enterprise names with a broad founder list. Apply the ICP Scoring Rubric and keep only leads at 8/10 or above. The score should reward verified pain, evidence access, reachable operator, and reference relevance. It should penalize switching risk you cannot yet support.

Run Signal Research Dossier only on the keepers. Require two dated signals such as a contract renewal, public cost review, leadership change, implementation complaint, or disclosed operating constraint. A company name without a dated reason is not ready.

Credit gate: use Contact Discovery only after the lead passes 8/10 and has two verified signals. This keeps expensive enrichment focused on people you are prepared to contact.

Phase 3: draft one evidence ask

Use Cold Email First Touch to create one short draft per lead. The ask is not a demo. It is permission to run the small evidence review. Every message must cite a verified signal and contain one ask.

For each approved lead, draft one first touch under 150 words. Use the dated signal, name the costly event, and offer a manual [evidence review]. Do not promise savings or a result. Leave every message in draft and stop before activation.

Review every message with Outreach QA Audit. Reject generic personalization and any claim not present in the evidence ledger. A human approves the final cohort and decides whether a campaign may be activated.

Phase 4: turn objections into proof

Use Reply Triage to classify interested, timing, wrong-person, objection, and no responses. Put recurring objections into Content Studio with their source evidence. Publish useful answers manually, then use Signal Audiences to capture and qualify the people who visibly engage.

The resulting loop is controlled: researched list → reviewed draft → human approval → reply classification → sourced content → new signal audience. The customer-success scorecard stays outside the campaign until delivery produces a real result and the buyer grants reference permission.

Final output: 25 accounts, at least five researched keepers, one draft per keeper, one objection ledger, and zero sends without approval.

Saveable checklist

• One buyer with urgent pain, not only a large budget
• One claim testable in the buyer's own data
• Five evidence reviews before more product or volume
• A service scorecard with a resolution measure
• Explicit, scoped reference permission
• One account-risk step upward at a time
• Human review before outreach activation
• No claim that this produced Judi Health's first 100 customers

Sources and evidence limits

• A Product Market Fit Show interview with AJ Loiacono, the primary source for the early chronology.
• Judi Health company history, for the 2017 founding and later brand transition.
• Forbes' 2022 Capital Rx profile, for independent reporting on the fee model and later scale. Forbes reported more than 150 customers representing 1.2 million people by 2022.
• U.S. Senate hearing record, for the transparent benchmark, pass-through structure, and flat fees.
• Forbes' 2025 Judi Health profile, for the later $3.25 billion valuation and funding structure.
• Capital Rx product and client evidence, for later service testimonials, including from a union health and welfare fund.

The $10–12 million ARR, zero-marketing period, early account sizes, 97 percent service target, 70 percent efficiency target, and union-first chronology are founder-reported. No public source names the literal first customer or documents the first 100, early churn, conversion rate, payback period, or complete lead-source mix. Later scale confirms that the company grew. It does not prove that the early reference motion alone caused every later result.