Onboard your first 10 customers

An 11-step playbook. Open in Dock and you'll get four surfaces seeded:

- **Customers** (table) — one row per customer, with stage, success metric, last touchpoint, owner
- **Brief** (doc) — the canonical onboarding playbook, written down so the next CSM doesn't re-derive it
- **Issues** (table) — bug reports + feature requests from the cohort, prioritized
- **Pointers** (table) — kickoff templates, customer-success guides, retention benchmarks

Read `Brief` top-to-bottom on first open. Each row in `Customers` is one of the 10. Click into a row to see status, milestones, and weekly check-in notes.

Outcome

10 customers using the product weekly, hitting their first success metric within 30-60 days, willing to give a reference call, and giving you 80% of the bug reports / feature requests that shape the next 6 months of roadmap.

Estimated time: 60-90 days from kickoff to 'self-running'
Difficulty: intermediate
For: Founders + first CSMs at companies under $1M ARR.

What you'll need

Pre-register or install before you start.

• Slack Connect (Free tier supports Connect (rate-limited), $7.25/user/mo Pro removes limits) — Shared channel with each customer. Faster than email, lower friction than ticketing.
• Loom (Free tier (25 videos), $15/user/mo Business) — Async video walkthroughs. 5-min Loom replaces 30-min meetings often.
• Linear ($8/user/mo Standard, free for under 10 users) — Track bug reports + feature requests from the cohort. Tag by customer.
• Calendly ($10/mo Standard plan) — Self-serve booking for kickoffs + check-ins.
• PostHog (Free tier (1M events/mo), $0.0005/event after) — Product analytics: see what each customer is actually doing in-app.

---

The template · 11 steps

Step 1: Define what 'successful onboarding' means before kickoff

Estimated time: 2-3 hr per customer (or once for a generic ICP)

Onboarding without a defined success metric is a feel-good ritual. Pick the metric: weekly active users in the customer org, first-week-Aha-moment completion, time-to-first-value, integration live in production. Pick one number, write it down, share it with the customer at kickoff. If they're not on the success metric by day 30 or 60, you have a problem to fix, not a feeling to debate.

Tasks

• Pick the metric: WAU? Time-to-first-value? Integration deployed? Workflow automated?
• Define the threshold (e.g. '5 WAU in their org by day 30')
• Define the sub-milestones (kickoff done, integration tested, first workflow shipped, 5 WAU)
• Write the metric into the Customer row + the Brief
• Confirm the metric with the customer at kickoff — get their head-nod that it's the right outcome

Pointers

• [Guide] Time to value: defining it (Reforge)
• [Guide] Customer success metrics (Lenny's Newsletter)

[!CAUTION] Gotchas

• 'They use it' is not a success metric. Pick a frequency + a workflow they actually completed.
• If you can't measure the metric in your own product, instrument it BEFORE the customer kicks off. Asking them to self-report is unreliable.

Step 2: Run a 60-min kickoff call within 7 days of contract signing

Estimated time: 60 min call + 30 min prep + 30 min recap

The kickoff sets the cadence and the expectations. Schedule within 7 days of signing — momentum dies fast. Bring the team that will own the relationship (founder + lead engineer at this stage). Walk the customer through: what we'll do together, what they need to do, what success looks like, what the first 30 days look like.

Tasks

• Schedule a 60-min kickoff within 7 days of signing
• Prep: review their stated use case from sales notes, build a 1-page agenda
• Run the agenda: introductions, success metric confirmation, week-1 plan, week-by-week cadence, escalation path
• Capture customer-side champion(s) + technical contact + decision maker
• Send a recap within 4 hr: agreed metric, agreed milestones, agreed cadence, action items by name
• Add the customer to a shared Slack Connect channel

[!CAUTION] Gotchas

• Skipping the kickoff because 'they're a small team and don't need it' is the most common cause of onboarding rot. The kickoff is the contract.
• If you don't capture the technical contact + the decision maker as separate names, you'll spend month 2 wondering who to escalate to.

Agent prompt for this step

Draft the kickoff call agenda + recap email for {customer_name}.

Read the customer's row in the Customers table + any sales notes attached.

Output:
1. A 1-page kickoff agenda (5-8 bullet sections, 60-min total, time-boxed)
2. A 1-paragraph recap email template to send within 4 hours of the call (agreed metric + 30-day milestone + cadence + action items as a numbered list)
3. A weekly check-in template the customer can expect every Friday

Constraints:
- The agenda is a working document — leave 'CUSTOMER FILLS IN' placeholders where their voice belongs
- The recap email is under 200 words
- No 'looking forward to partnering' marketing voice — write like you'd write to a friend's company

Step 3: Open a shared Slack channel — be in their workspace, not yours

Estimated time: 15 min to set up per customer, ongoing

Email is too slow for early customers. Slack Connect (or Microsoft Teams shared channels) lets you respond in 5 min instead of 5 hours and surfaces problems while they're still small. The channel is a leading indicator: customers who go silent in Slack are usually 2-3 weeks from churn.

Tasks

• Send the customer a Slack Connect invite within 24 hr of kickoff
• Add their champion + technical contact + decision maker
• Pin a welcome message: 'reply here for anything, target 1-business-day response, urgent issues get tagged @here'
• Add the channel to your team's Slack-channel-watchers tool (don't rely on default notifications)
• Set the SLA: respond within 4 business hours, even if the answer is 'looking into it'

[!CAUTION] Gotchas

• Slack Connect channels under the free tier have a 90-day message history limit. Upgrade to Pro before you lose the kickoff thread.
• If your team can't commit to 4-hour response on the channel, don't open one. A dead Slack channel is worse than email.

Step 4: Walk the customer through their first real workflow personally

Estimated time: 60-90 min per customer

Self-serve onboarding documentation does not work at customer 1-10. You walk them through it, screen-share, in real time, with their data. That session is where you find the 5 sharp edges in your product nobody else flagged. Take notes ruthlessly — every 'oh this is confusing' is a roadmap item.

Tasks

• Schedule a 90-min screen-share for week 1 ('let's set up your first workflow together')
• Have them share THEIR screen — you watch, you don't drive
• Take notes on every confusion, every 'how do I' question, every back-button press
• After the call: log every issue you noticed to the Issues table tagged with the customer
• Within 24 hr: ship a doc / Loom / fix for the top 3 friction points

[!CAUTION] Gotchas

• Driving the screen-share yourself = you don't see what they don't see. Make THEM drive.
• If you find 0 issues in the screen-share, you weren't watching closely enough. There are always issues at customer 1-10.

Step 5: Set up a weekly 30-min check-in cadence

Estimated time: 30 min/customer/week for 60 days

Weekly 30-min check-ins for the first 60 days are non-negotiable. The cadence is what differentiates onboarding from 'we sold them and forgot'. Same time every week, recurring on the calendar, with a structure: what shipped, what's blocked, what they need from us, what we need from them.

Tasks

• Schedule a recurring 30-min weekly check-in within 14 days of kickoff
• Use a 4-section structure: their wins this week, blockers, our actions, their actions
• Send a 1-paragraph recap within 4 hr after each check-in
• After day 60, taper to bi-weekly if they're hitting the success metric
• If they cancel 2 check-ins in a row, that's a churn signal — escalate to the champion

[!CAUTION] Gotchas

• Skipping a weekly check-in 'because there's nothing new' is the wrong instinct. The check-in IS the relationship; cancelling it kills it.
• If the customer dominates the check-in with feature requests, you're missing the success-metric conversation. Steer it.

Step 6: Instrument product usage so you don't have to ask

Estimated time: 1-2 days to instrument, ongoing to monitor

Asking 'how's it going?' on the weekly check-in is fine. Asking 'what features are you using?' is unprofessional — you should already know. Instrument every action your product takes, pipe it to PostHog (or similar), and build a per-customer usage view. By week 3 you should be able to predict churn before they tell you.

Tasks

• Identify 3-5 events that map to 'using the product' for your category (logins, key actions, integrations triggered)
• Instrument them in PostHog or similar with customer_id as a property on every event
• Build a per-customer dashboard: WAU, key event count, last seen
• Set up a Slack alert when a customer's WAU drops 50%+ vs prior week
• Add 'usage trend' as a column in the Customers table

Pointers

• [Tool] PostHog product analytics
• [Guide] Choosing the right activation metric (Reforge)

[!CAUTION] Gotchas

• Tracking 'page views' as your activation signal is misleading. Pick events that prove the customer got value (workflow completed, integration triggered, output exported).
• Customers don't like being told 'we noticed you stopped using us'. The phrasing matters: 'wanted to check in, anything we should be aware of'.

Step 7: Capture every bug + feature request in one place

Estimated time: Ongoing, 30-60 min/week

Customer 1-10 generate the bug reports + feature requests that define your roadmap for the next 6 months. Capture them all in Linear (or your tracker) tagged with the customer + severity. Resist the temptation to 'remember' — by customer 6 you'll have 200 issues across 10 customers and you'll lose half of them.

Tasks

• Every issue raised in Slack / email / call gets a Linear ticket within 24 hr
• Tag with customer name + severity (P0 outage, P1 blocker, P2 fix soon, P3 nice-to-have)
• Add the customer's verbatim quote to the ticket — context dies fast otherwise
• Triage weekly: roll up by tag to see which feature requests have multiple customers (those become roadmap)
• Reply in Slack to acknowledge every ticket within 4 business hours, even if 'we're not fixing this now'

[!CAUTION] Gotchas

• Telling a customer 'we'll fix that' without a date and then not fixing it is the fastest way to torch trust. Always commit to a date or explicitly say 'not in the next quarter'.
• Bug reports without verbatim quotes lose context. Six weeks later you won't remember why P1.

Agent prompt for this step

Process the issue {customer_name} raised in Slack today.

I'll paste the message + context. Output:

1. A clean ticket title (action-oriented, "Fix X" / "Add Y" / "Investigate Z")
2. A 1-paragraph description with the customer's verbatim quote in italics
3. Severity (P0 outage / P1 blocker / P2 fix soon / P3 nice-to-have)
4. A draft Slack reply acknowledging it within 4 hours: 'logged as <ticket-id>, here's our current take, will follow up by <date>'
5. Add the row to the Issues table tagged with the customer

Constraints:
- Verbatim quotes only — don't paraphrase what they said
- If severity is unclear, default to P2 and flag for the user
- The Slack reply is honest about when we'll fix it, not vague

Step 8: Hit the 30-day success milestone — celebrate it explicitly

Estimated time: 1 hr per customer at the milestone

When a customer hits the agreed success metric, mark it: send a celebration email, share an internal Slack note, ask for the next thing (case study, reference call, expansion conversation). The 30-day milestone is the most important conversion point of the whole onboarding — that's when 'they bought' becomes 'they got value'.

Tasks

• Within 1 day of hitting the metric: send a personal email celebrating + reaffirming the next milestone
• Internal Slack note + add to the Brief log: customer X hit 30-day milestone
• Ask: 'would you be open to a 15-min reference call for a similar customer we're talking to'
• Ask: 'would you be willing to be part of a case study'
• If neither: ask for written feedback you can quote anonymously

[!CAUTION] Gotchas

• Asking for case study + reference + expansion all in one email reads as transactional. Stagger them across 2-3 weeks.
• If the customer hits the metric and you DON'T mark it, you've trained them to expect silence. Future asks land cold.

Step 9: Handle the first churn risk — fast and direct

Estimated time: 1-2 hr per at-risk account

Some of your first 10 will churn or threaten to. Don't pretend it isn't happening. The signals: usage drop 50%+ for 2 weeks, weekly check-in cancelled twice, slow Slack replies, the champion changes role internally. When you see them, escalate within 48 hr: founder calls champion, asks directly 'are we still working?'. Half the time the issue is fixable; the other half you save the relationship for the future.

Tasks

• Define the at-risk signals: WAU drop 50%+, missed check-ins, champion role change, slow Slack
• When 2+ signals fire: escalate to founder within 48 hr
• Founder runs a direct 1:1 call with the champion: 'how is this going for you, what should we do differently'
• Within 7 days: ship the fix or the alternative (lower price, paused contract, free month, scope reduction)
• If they churn: post-mortem in the Brief — what would have prevented it

Pointers

• [Guide] Saving accounts: First Round playbook

[!CAUTION] Gotchas

• Pretending churn signals aren't there because the conversation is uncomfortable is the most expensive mistake you'll make. Lean in.
• Discounting your way out of churn rarely works long-term. Fix the actual problem (UX, value, fit) or let them go cleanly.

Step 10: Capture the case study while the win is still warm

Estimated time: 3-5 hr per case study

By day 60-90, customers who hit the success metric should be willing to be a case study. Capture the case study while the memory is fresh — by month 6 they'll have moved on and the specifics blur. The best case studies are interview-driven: founder interviews champion for 30 min, draft the story, get sign-off, publish. Quotes + numbers + specifics > marketing voice.

Tasks

• Schedule a 30-min case-study interview with the champion at day 60-75
• Ask: what was the situation before, what changed, what's the measurable outcome, what's the quote
• Draft a 600-word case study with 2-3 verbatim quotes + 1-2 outcome numbers
• Get sign-off from champion + their legal / comms before publishing
• Publish on the website + send to the rest of your prospect pipeline

[!CAUTION] Gotchas

• Customers will sometimes redline the draft heavily — accept it. A real customer-edited case study reads more credible than a polished marketing draft.
• If they ask not to be named publicly, accept it gracefully. An anonymized 'a 200-person SaaS company in healthcare' case study is still useful.

Agent prompt for this step

Draft the case study for {customer_name}.

Read the Customers table row, the Brief log entries, the Issues table for this customer, and any interview transcript I attach.

Output a 600-word case study with this structure:
1. Headline (action-oriented, with the outcome metric)
2. Situation (3-4 sentences: customer profile + what was hard before)
3. Solution (3-4 sentences: what they did with us, no marketing voice)
4. Outcome (3-4 sentences: 1-2 numbers + 1 verbatim quote)
5. Quote pull-out (1-2 sentences from the champion, verbatim)

Constraints:
- No 'unleash', 'transform', 'revolutionize'
- Specific numbers > vague 'significantly'
- The quote is verbatim from the interview — do not invent

Step 11: Document the playbook so customer 11+ doesn't start from zero

Estimated time: 1-2 days

By customer 10 you've earned a real playbook. Write it down: kickoff agenda, week-by-week milestones, screen-share runbook, escalation paths, common bugs, common questions. Customer 11+ is where you start to leverage what you learned. Founders who don't write it down end up doing white-glove for 100 customers and burning out.

Tasks

• Write the canonical kickoff agenda + recap email template
• Write the week-1 / week-2 / week-4 / week-8 milestone checklist
• Write the screen-share runbook: what to demo, what to watch for
• Document the top 10 issues + their canonical resolutions
• Document escalation: who to ping, what to say, when to bring in founder
• Hand the playbook to the next CSM hire (or your future self)

[!CAUTION] Gotchas

• Documentation written 'when there's time' never happens. Block 1-2 days for it after customer 10 lands; treat it as critical-path work.
• Playbooks rot — customer 11-20 will surface 5 things customer 1-10 didn't. Update the playbook quarterly, not once.

---

Hand the template to your agent

Paste the prompt below into your agent's permanent system prompt so the agent reads, writes, and maintains this workspace as you work through the steps.

You are an agent on the "Onboard your first 10 customers" playbook workspace.

Your role: maintain the four surfaces (Customers, Brief, Issues, Pointers) as the user onboards each customer.

Cadence:
- Each Monday morning: surface every customer's status, days since last touchpoint, current milestone, and any unresolved issues from the Issues table.
- When the user logs a kickoff call, draft the kickoff recap email + the 30-day milestone the customer agreed to.
- When PostHog usage drops 50%+ for a week, flag the row red and draft an outreach.
- When the customer hits the success metric, draft the celebration email + the case-study request.

First MCP tool calls:
1. list_surfaces(workspace_slug="onboard-your-first-10-customers")
2. list_rows(workspace_slug="onboard-your-first-10-customers", surface_slug="customers")
3. get_doc(workspace_slug="onboard-your-first-10-customers", surface_slug="brief")

Hard constraint: never email a customer without explicit user approval. Drafting + flagging + tracking is your job; sending is theirs.

---

FAQ

Why white-glove and not self-serve at this stage?

Self-serve onboarding works when you have 10000 customers and a 5% activation rate — the 500 who succeed pay for the 9500 who don't. With 10 customers and a 50%+ activation expectation, the math inverts: every customer matters too much to risk on docs and onboarding flows you haven't validated. White-glove until customer 50+ is the right default.

How long does white-glove onboarding take per customer?

Realistic budget: 5-10 hours/week per customer for the first 30 days, tapering to 1-2 hours/week by day 60. With 10 active customers in onboarding at once, expect 1-2 founders or CSMs full-time on this work. That's why 'first 10 customers' is a deliberate ceiling.

Should I onboard 10 customers in parallel or 2-3 at a time?

2-3 in parallel until you've documented the playbook, then scale up. Onboarding 10 in parallel before you know what you're doing means each customer gets a worse experience and you can't learn fast enough between them.

Can my AI agents actually help with customer success?

Yes for the off-stage work: drafting kickoff agendas, drafting check-in recaps, monitoring usage data and flagging churn risks 2-3 weeks early, drafting case study first drafts. Not for the customer-facing call itself. Use agents to keep you informed and ready, not to replace the human relationship.

What's the most common reason early customers churn?

Three reasons in order: (1) they never hit the first success metric because onboarding stalled (the most common, and most preventable), (2) the champion left or changed role internally and the new owner doesn't see value, (3) the product doesn't actually solve the problem they thought it did (real fit issue, hardest to fix). Watch for (1) by week 3 — that's when you can still save it.