The First 100 Customers Course #2: What Connect The Dots Learned From a Churned First $5K Deal

TL;DR

Connect The Dots did not begin its commercial motion with a database of strangers. Founder Drew Sechrist says the company first gave the product to several companies for free and asked them to use it. He then looked across relationships earned during his Salesforce years for a company with the right problem at the right moment.

Skyflow provided that moment. Sechrist knew its CEO, and the company had hired a relationship-led sales executive who wanted to activate the network around the business. Skyflow became the first paid customer for approximately $5,000. It later churned.

The repeatable lesson is not “ask friends for introductions.” It is earned relationship → timely buyer job → connector-approved introduction → bounded paid bet → activation → retention gate. A relationship creates access. Payment proves a buyer will take a bet. Only repeated product value earns permission to scale.

What you will build

This course produces a first-buyer system for narrow B2B SaaS. You will define one buyer job and trigger, recruit behavior-based design partners, separate warm paths from signal-led cold accounts, write a connector-controlled introduction, sell a bounded paid bet, and install activation and retention gates before increasing outreach.

The output is not a large contact list. It is a small evidence ledger showing why each account entered the pipeline, who can honestly connect you, what the buyer must accomplish, what the buyer paid to test, whether the product activated, and whether the job repeated after the founder stopped carrying it.

Evidence and limits

The first-customer sequence, approximate deal size, and churn come from Sechrist's July 2026 interview on The First Customer. They are founder-reported. No public contract or churn record independently confirms them, and the interview does not disclose a first-10 or first-100 customer count, CAC, payback, or retention rate.

Independent reporting gives context rather than proof of the first deal. TechCrunch reported in December 2021 that Connect The Dots had raised a $5 million seed and a $15 million Series A, had 55 employees, and distributed the product through a beta-by-request model. That means capital, team, and free access should not be mistaken for a proven repeatable acquisition system.

Stage	What happened	Evidence status
Pre-revenue learning	Several companies used the product free and supplied feedback	Founder-reported; no count disclosed
Buyer trigger	Skyflow hired a sales leader who wanted to use the company network	Founder-reported
First paid deal	Approximately $5,000	Founder recollection; not independently audited
Retention	Skyflow later churned	Founder-reported; reason not publicly documented
December 2021	$20M reported funding, beta distribution, 55 employees	Independent TechCrunch profile
Later product	Team workflows, CRM, API, and MCP connections	Current company-described functionality

The model: access, payment, activation, retention

Four gates are often collapsed into one. A connector can create access because they trust both sides. A paid pilot can show that a buyer values the proposed result enough to spend money. Activation can show that users reached the promised outcome. Retention can show that the job remains important after novelty and founder attention disappear.

Connect The Dots crossed the first two gates with its first paid customer. The later churn is evidence that it had not yet secured the fourth. That is useful, not embarrassing: it prevents a founder from scaling a channel that is better at transferring trust than the product is at keeping it.

Step 1: define the buyer job before searching the network

Sechrist's relationship with Skyflow's CEO mattered because a new sales leader had a concrete job: make the company's existing relationships useful for revenue. Without that trigger, the same connection would have been social proximity rather than purchase intent.

• Buyer: the person accountable for the outcome.
• Trigger: the change that makes the job urgent now.
• Hidden asset: data, relationships, workflow, or capacity they cannot use.
• First outcome: an observable result within 14 days.
• Failure: the condition that pauses selling and returns work to product.

Write the job in the buyer's language, then test it with five people who did not help invent the category. Pass: they recognize the job and can name the trigger. Stop: your only reason to contact them is that you know someone nearby.

Step 2: make free design partners produce behavior

Free access is useful only when it buys evidence. Connect The Dots asked early companies to use the product and provide feedback. A design partner who attends calls but never performs the job is not a user; they are an interested observer.

Give each partner one weekly job, one named user, one activation event by day 14, two observed sessions, and a fixed continue-pay-stop review. Log the user's actions, founder interventions, missing inputs, failed states, and workarounds. Avoid “Would you use this?” surveys. Measure whether the behavior occurred twice.

Pass: three unaffiliated teams perform the job twice. Stop: the founder performs the result while users only react to a demo. This gate matters because a warm first buyer can otherwise conceal the absence of independent usage.

Step 3: keep three acquisition lists

A warm path is not a contact record. It is a credible relationship between a connector and a buyer, plus permission to use it. Keep that evidence separate from cold accounts so you do not attribute the connector's trust to your message or product.

List	Admission rule	Next action
Warm path	A connector genuinely knows the target and the buyer job is active	Ask permission to introduce
Signal-led cold	The account shows the trigger, but no credible relationship exists	Send a direct hypothesis without false familiarity
Watch	Potential fit, no timing evidence	Monitor the trigger and publish useful material

Report meetings, activation, and retention by list. If warm accounts meet but do not activate, the network is doing its job and the offer or product is not. If signal-led cold accounts reply but lack the job, the signal definition is too broad.

Step 4: prewrite the introduction and return control

Sechrist describes a manual tactic from Salesforce: identify a person with a real path to a target, then make the introduction easy by drafting the email. The connector must retain full control. Their relationship is not inventory you own.

I think [buyer] may be dealing with [observed job] after [trigger]. We built [specific outcome] for [narrow user]. If you believe the fit is real, would you introduce us? I drafted a short note below. Please edit or ignore it. I will not use your name without your approval.

The forwarded note needs one buyer fact, one outcome, and one small ask. Do not insert a biography, feature list, or urgency the connector cannot endorse. Pass: well-matched connectors approve or improve the note. Stop: several say the fit would spend trust without helping the buyer.

Step 5: sell a bounded paid bet

The approximately $5,000 first deal was small enough to be described as a bet and large enough to test willingness to pay. Your equivalent should cover one team, one workflow, one first outcome, and one decision date. State what the founder will do manually and what the buyer must supply.

• One accountable buyer and one named daily user.
• One activation event observable within 14 days.
• One price tied to the test, not an invented future contract value.
• One list of excluded requests and unsafe edge cases.
• One expand, continue, change, or stop review date.

Pass: an unaffiliated buyer pays and reaches the activation event. Stop: payment arrives because of the relationship, while users cannot reach the promised result. Keep “won” and “activated” as separate fields in the CRM.

Step 6: instrument activation before adding outreach

Define activation as a buyer result, not a login. For relationship intelligence, that might be finding a credible path to a named account, obtaining connector approval, and completing an introduction. For another SaaS, it may be a reconciled invoice, an approved report, or a workflow completed without founder intervention.

For every paid bet, record time to first outcome, user actions, founder minutes, failed inputs, recovery work, second use, and the date the buyer would notice if the product disappeared. Review after customer one, three, and five. Do not average away a repeated failure.

Step 7: let churn revoke permission to scale

Sechrist says the first customer later churned and that the engagement exposed product problems. The interview does not publish a formal reason, so the responsible move is to avoid inventing one. A churn review should separate what is known from the founder's preferred story.

1. Why did the buyer agree to meet and pay?
2. Which activation event happened, failed, or required founder rescue?
3. Did the job repeat after the first success?
4. Did urgency, ownership, or budget change?
5. Does the evidence demand a product fix, onboarding fix, ICP change, or no action?

Use a conservative scale gate: three unaffiliated paid customers activate, and at least two retain through the proof period. Two repetitions of the same product failure pause new outreach for one product sprint. The exact numbers are an operating rule for your experiment, not a claimed Connect The Dots conversion benchmark.

A 30-day execution plan

• Days 1–3: choose one buyer, trigger, hidden asset, and 14-day outcome.
• Days 4–7: recruit three behavior-based design partners.
• Days 8–10: build warm-path, signal-led cold, and watch lists.
• Days 11–14: ask ten credible connectors for permission.
• Days 15–18: offer a bounded paid bet to qualified buyers.
• Days 19–24: observe activation and log every manual intervention.
• Days 25–27: run the retention review and classify failures.
• Days 28–30: scale only if activation and retention gates pass.

The first-customer ledger

Keep one row per account with the buyer job, trigger date, source list, connector and approval date, introduction copy, meeting outcome, paid-bet terms, activation event, time to activation, founder minutes, repeated use, renewal decision, churn facts, and next action. Link every public signal to its source and date.

Review the ledger as a funnel with distinct gates: qualified trigger, connector approval, meeting, payment, activation, repeat use, and retention. Warm and cold rows may share a product funnel, but they must not share an acquisition denominator. This is how you learn whether a network is revealing demand or merely manufacturing meetings.

Do not turn the first $5K into a pricing rule

The approximate amount is evidence of a paid experiment, not a universal starting price. The interview does not disclose scope, delivery cost, margin, procurement terms, or the value Skyflow expected. Copying the number without those variables would turn a useful case into theatre.

Price your first bet from the cost and value of the bounded outcome. Write down the manual work you will perform, the buyer time required, third-party costs, the downside if the test fails, and the value of a successful first result. The amount should be high enough that payment is a meaningful commitment and low enough that both parties can stop at the review date without pretending they signed a permanent platform decision.

After the test, separate three questions. Did the buyer pay? Did the user activate? Did the account retain? A larger initial price cannot repair a failed activation, and a discounted pilot does not explain churn by itself. Update pricing only after the ledger shows repeated value and a stable delivery cost.

Turn objections into trigger evidence

A declined introduction can improve the system if you record the reason. Use a small taxonomy: no active job, wrong buyer, weak relationship, unclear outcome, poor timing, product risk, or connector discomfort. Do not treat every “not now” as a copywriting problem.

Aggregate reasons after every ten connector requests. If most targets lack an active job, tighten the trigger and move them to watch. If connectors understand the job but cannot endorse the outcome, strengthen the proof or narrow the promise. If buyers meet and pay but activation fails, stop researching more contacts and repair the product path.

Publish useful material only when it answers a repeated, sourced objection: a checklist for a new sales leader, a migration worksheet, a risk review, or a teardown of the manual workflow. Engagement with that material is a signal to reassess, not permission to claim a relationship. The account returns to the warm or cold list only when the original buyer job becomes observable.

When this motion fits—and when it does not

The method fits narrow, high-consideration B2B products with identifiable accounts, a job that becomes urgent after a visible trigger, and founders or teammates who have earned relevant relationships. It also requires a small result a buyer can evaluate before a broad rollout.

It does not fit anonymous consumer acquisition, commodity low-ticket software, invented LinkedIn familiarity, or a product that still fails its core job. It also fails when the connector cannot explain why the introduction helps the buyer. Do not automate approval, scrape private relationships, or treat a shared employer as permission.

Sources and evidence limits

• Drew Sechrist on The First Customer, 22 July 2026. The stored 33,729-character transcript was audited in full for the original research pack.
• TechCrunch's December 2021 funding and beta profile for independent historical context.
• Connect The Dots relationship intelligence overview and product guides for current company-described workflows.
• The Connect The Dots Sales Playbook, 15 February 2022, for the company's network-curation and multi-threading process.

The first customer, approximate price, free-user sequence, and churn remain founder-reported. No first-10, first-100, CAC, payback, or retention figures were disclosed. The independently verified historical funding total is $20 million as of December 2021; later funding claims from the interview are not used here. Copy the evidence gates, not an invented growth curve.