The First 100 Customers Course #6: How Skimmer Turned One Cold Call Into 76 Customers in 12 Months

TL;DR

Skimmer did not get 76 customers from one cold call. The call did something more useful: it turned a friend's complaint into evidence that an unknown buyer had the same urgent problem. Founder Ron Hash found a pool-service company on YellowPages.com and heard that running field operations on paper was killing the business. He kept his day job, built at night, and launched a mobile-first product for pool technicians.

The first paying customer was in Phoenix, although the source does not name the account or say how it arrived. Predictable acquisition came later from SEO for phrases such as “pool service software” and “pool service app.” Hash then called new users, watched their setup progress, answered support quickly, and guided them to run the next day's route. He reports 76 customers after 12 months, more than 370 after 24 months, about 906 after 36 months, and more than 1,500 roughly six months later, without paid marketing.

The mechanism was problem interview → field-level wedge → high-intent search → founder-assisted activation → fast support → word of mouth. The warning is that each step is narrower than it sounds. One interview is permission to keep learning, not product-market fit. SEO only compounds when signups activate and stay. And the founder-reported churn improvement from roughly 6% to 2% is useful directional evidence, not a defined monthly logo-churn metric.

What you will build

You will build a 30-day vertical-intent loop for one narrow workflow. The outputs are a ten-call problem ledger, a field-user wedge, a search-intent map, an activation event, a founder welcome-call script, a support promise, a value-metric pricing test, and a weekly scorecard. This is not a content-SEO marathon. It is a controlled path from an urgent search to a completed job and a retained account.

Who this is for

Use this when you serve a clear vertical or role, the buyer already describes the problem in stable language, and the first useful workflow can be completed in a day. It fits vertical SaaS, operational AI agents, field tools, and specialised B2B software where the founder can still speak to every new account.

Do not use it when the category is too new to have search intent, when the value appears only after months of implementation, or when one founder cannot safely support the workflow. In those cases, direct outreach, a design-partner motion, or practitioner partnerships should create the language before SEO tries to capture it.

Case snapshot

Stage	What happened	Evidence limit
Problem signal	A pool-service friend complained that available tools were poor	Founder-reported
Validation call	One Texas operator found through Yellow Pages confirmed paper was painful	Not identified as the first customer
2015	Named customer recalls seeing an early iPad prototype	Retrospective customer story
2017	Company says the first version launched in Phoenix	Launch date differs from prototype date
First customer	Founder says the account was in Phoenix	Name, date, channel, and invoice undisclosed
12 months	76 customers	Founder-reported
24 months	More than 370 customers	Founder-reported
36 months	About 906 customers	Founder-reported estimate
~42 months	More than 1,500 customers and over $1m ARR	Founder-reported; no audited financials
2020 onward	Unbundled invested; later leadership scaled the company	Later context, not early-channel attribution

The model: capture intent, then manufacture activation

Search traffic is not a distribution system. It is a stream of people expressing intent. The system begins when a founder chooses a phrase attached to an expensive workflow, builds the smallest product that completes that workflow, and makes the first successful use observable. Every early signup then becomes a supervised experiment.

Skimmer's loop joined acquisition and retention. SEO brought an operator already looking for software. The product replaced tomorrow's paper route. A welcome call removed the next block. Fast support protected the buyer's field operation. A good experience produced recommendations inside Facebook groups where Hash was not marketing. Those recommendations created more demand and likely more searches. Pricing by serviced pool reinforced the loop because a growing customer could add employees without buying seats.

Step 1: use one call as a gate, not a verdict

Hash's friend supplied the idea. The stranger in Texas supplied a second data point from outside his network. The useful part was not the number of interviews. It was the specificity of the workflow: multiple technicians, field records, central paper collection, chemical usage, and billing. That gave the product a job to replace.

Run ten calls with this worksheet:

• Trigger: what happened immediately before the buyer looked for help?
• Current system: paper, spreadsheet, generic tool, or human coordinator?
• Field user: who performs the job when the manager is absent?
• Failure cost: money missed, time lost, customer risk, or work repeated?
• Existing phrase: what would the buyer type into search?
• Tomorrow test: which real job could a first version complete tomorrow?

Pass condition: five buyers describe the same trigger, field job, and cost in similar language. Stop condition: every call produces a different workflow or buyers agree the problem exists but will not change their current system.

Step 2: build for the person doing the work

Existing tools were web based. Hash's wedge was the technician carrying the device beside a pool. He reduced taps for chemical readings and made the app work offline. The first iPad-only choice was premature: customers did not want to buy and protect tablets for field staff. A later Skimmer customer recalls rejecting that early prototype for exactly this reason. Expanding to iPhone and Android removed the hardware tax.

Map the manager's purchase promise against the field user's action. If the manager wants better billing, ask which field record creates the invoice. If the buyer wants an AI report, ask who supplies the source data and when. Build the shortest reliable path for that person first.

Pass condition: a new user completes one live job without founder data entry. Stop condition: adoption requires new hardware, double entry, or a manager standing beside every operator.

Step 3: own one high-intent search

Hash taught himself enough SEO to rank for “pool service software” and “pool service app.” He does not publish traffic, ranking dates, or conversion rates, so the lesson is not “SEO always works.” It is to capture language that already sits close to a purchase. A founder with no audience does not need thousands of educational visits if a small number of vertical buyers are searching for the exact replacement.

Intent layer	Page job	Proof	Decision
Problem	Name the costly workflow	Before/after process	Keep only if buyers use this language
Category	Explain the specialised solution	Live workflow	Primary early target
Alternative	Compare paper or generic tools	Migration path	Use when switching is common
Brand	Convert referred demand	Support and customer proof	Measure separately from non-brand SEO

Publish one category page, one workflow page, and one migration page. Give each a single live activation goal. Pass condition: five qualified organic signups or booked calls in 30 days and at least two complete the first job. Stop condition: traffic rises but buyer fit and activation remain flat.

Step 4: supervise the first activation

Hash built a private dashboard showing what a new account had done. During lunch or after work, he called new users and adapted the conversation to their progress. The onboarding flow then focused on five steps: sign up, enter customers needed tomorrow, build tomorrow's route, and run it in the field. The first step was already marked complete, making progress visible.

Use this welcome-call script:

1. “What real job must be completed next?”
2. “I can see you reached [event]. What stopped the next step?”
3. Complete the blocked step with the user, without taking control away.
4. Ask what they expected to happen and which word or screen misled them.
5. Schedule the next real use and record whether it happened.

Metric: percentage of qualified signups completing one live job within 24 hours. Pass condition: at least 60% after founder assistance. Stop condition: users complete setup but the real workflow still needs spreadsheets or founder labour.

Step 5: make support part of acquisition

Skimmer placed a phone number prominently, Hash answered calls, and he says email replies came within about 30 minutes. This was not a generic “delight customers” tactic. Field software can block revenue-producing work. Responsive support reduced the risk of adopting an unknown tool. Buyers then recommended Skimmer inside pool-professional Facebook groups even though Hash was not active there.

Write a support promise around the workflow's risk. Define the hours, response target, issue class, and fallback. Tag each conversation as activation, product defect, missing workflow, training, or bad fit. Every week, fix the highest-frequency block and tell affected customers. Hash alternated larger marketable releases with batches of small quality-of-life improvements; both created word of mouth.

Pass condition: at least 20% of new qualified accounts mention a person, group, or customer referral by month three. Stop condition: support volume grows faster than activated accounts or recommendations depend on heroic exceptions.

Step 6: price on a value unit buyers can calculate

Skimmer initially charged $0.50 per serviced pool with a $29 minimum. The buyer could state the number of pools, divide by two, and know the bill. Unlimited users let the software spread through the company without making every technician a purchasing decision. One customer told Hash that a rising Skimmer bill meant his own business was growing.

Test three candidate metrics: the unit that creates revenue, the unit that reflects workload, and the unit the customer already counts. Reject a metric when usage is unpredictable, value is delayed, or buyers need a calculator and a contract expert to understand the bill.

Pass condition: eight of ten buyers predict their monthly price within 20% and agree that paying more corresponds to receiving more value. Stop condition: the metric encourages customers to hide usage or prevents broad adoption.

What failed, and what not to copy

• Premature platform assumption: iPad-only software imposed new hardware on field teams. Phone support improved adoption and churn.
• Slow channel: SEO did not create next-week traction. Hash describes the first year as the long, slow SaaS ramp despite reaching 76 customers.
• Undefined metric: the reported churn fall from 6% to 2% lacks a period and type. Use it as a direction, not a benchmark.
• Founder memory: the first-ten channel mix was not remembered precisely. Do not invent attribution that the founder could not supply.
• Later-stage halo: Skimmer's later funding and current scale prove the market was durable. They do not prove that a new founder can repeat the result or that SEO caused every later customer.

Your 30-day implementation plan

1. Days 1–3: interview ten vertical buyers and complete the workflow ledger.
2. Days 4–6: choose one field action and define the 24-hour activation event.
3. Days 7–10: ship the smallest real workflow and instrument every step.
4. Days 11–14: publish category, workflow, and migration pages.
5. Days 15–21: call every qualified signup and supervise one live job.
6. Days 22–25: remove the most frequent activation block.
7. Days 26–28: test the value metric with ten buyers.
8. Days 29–30: decide to continue, change channel, or change wedge.

Continue only when at least five qualified accounts complete the live job, three return for a second use, and two arrive from search or referral without personal outreach. Do not buy traffic to rescue weak activation.

Lead Scorer implementation: run the loop with approval gates

Lead Scorer fits the research and founder-assistance parts of this motion. It cannot create search demand, guarantee 76 customers, or replace the founder's activation calls. Use it to build a small vertical account universe, preserve interview evidence, qualify before spending enrichment credits, draft outreach, and turn objections into better pages and onboarding.

Phase 1: define and discover

Start with the icp-offer-context skill. Record the vertical, company size, field role, current system, trigger, and disqualifiers. Then use registry or company-search tools to create a discovery list. Keep interview prospects separate from customers and content engagers. Do not enrich the full market.

Copyable prompt: Define an ICP for [workflow]. Exclude companies without [trigger]. Find 50 accounts in [region], but do not enrich contacts yet. Return evidence for every inclusion.

Score the list with icp-scoring-rubric. Set a pre-enrichment threshold such as 7/10 and inspect borderline accounts manually. Output: 15–25 companies with a visible workflow signal. Stop condition: fewer than ten accounts carry evidence of the same problem.

Phase 2: research, enrich, and draft

Run signal-research-dossier on the keepers. Require two sourced, dated signals when possible. Use contact-discovery only after ICP confirmation because email and phone discovery consume the most expensive credits. Target the operator closest to the workflow and a manager who feels its cost.

Copyable prompt: For each approved company, find the person responsible for [workflow]. Enrich only leads scoring 7 or higher. Stop and request confirmation before any bulk credit spend.

Draft the first interview ask with cold-email-first-touch. The ask is a 15-minute workflow interview, not a demo. Keep one verified signal, one problem hypothesis, and one question. Run outreach-qa-audit, then review every message. The agent may prepare a campaign draft; a human decides whether and when anything sends.

Phase 3: close the evidence loop

Tag replies as problem confirmed, wrong workflow, bad timing, active evaluation, or no fit. Feed repeated objections into Content Studio as source notes. Turn the buyer's language into the category page, migration page, onboarding copy, and call script. When a useful post attracts visible engagement, capture that audience separately, score it, and draft follow-up only for qualified people.

Approval points: approve the scoring rubric, approve enrichment credit spend, approve every outreach draft, approve campaign activation, and approve any reuse of a named customer insight. Pass condition: the system produces interviews and better activation evidence. Stop condition: it merely produces a larger list.

Checklist

• One repeated vertical workflow, described by at least five buyers.
• One field-level action completed within 24 hours.
• One category term carrying purchase intent.
• Every early signup observable through activation.
• Founder call and support promise attached to real workflow risk.
• Referral source captured at signup and during calls.
• Value metric understood without a pricing workshop.
• No paid acquisition before activation and return use pass.

Sources and limits

The early chronology, customer milestones, pricing, acquisition channel, support practice, and churn movement come from Ron Hash's July 2026 interview on The SaaS Podcast and its pricing playbook. These figures are founder-reported. The transcript does not name the first customer, define the churn period, publish SEO analytics, or audit the customer counts.

Skimmer's company history and Dazzle Pools customer story support the prototype-to-launch distinction and early iPad resistance. The company says the first version launched in 2017; the customer remembers seeing a prototype in 2015.

Mainsail Partners independently confirms the 2020 Unbundled investment context and announced $74m in growth capital in 2024, when it reported more than 29,000 pool professionals. An Inc. profile lists a 2017 founding year and later three-year growth, while AQUA Magazine reports Skimmer's later Deloitte Fast 500 result. Those sources validate later scale, not the causal attribution of the first 76 customers.