How to Build a Customer Academy in Two Weeks Without an Instructional Designer

A customer academy that takes six months to launch usually never launches. The team hires a designer, scopes ten tracks, negotiates with legal, and by month five the CS org has moved on to whatever crisis replaced adoption in the quarterly plan.

Two weeks is the right timeline for the first version. Not because two weeks is a limit, but because two weeks is a forcing function that keeps scope honest.

What should you build in the first two weeks?

One course. One segment. One enrollment trigger.

Pick the single customer role that generates the most inbound onboarding requests. Usually that is the customer admin at a new mid-market account. Everything else waits.

The course you build for that role should cover the ten to fifteen tasks that admin has to complete in the first 30 days. Not everything they will ever do. Not the advanced workflow they might touch in month six. The critical path from signed contract to product value.

If you cannot describe that path in a paragraph, stop and write it first. The scoping mistake most teams make is building content before they have written down the outcome.

What content sources should you actually use?

You do not need new content. You need to structure what you already have.

• The best onboarding call recording of the last quarter. One CSM, one great customer, one 45-minute session that ends with an activated account. Chop it into five to seven lessons.
• Your top ten support tickets. These are the questions your admins ask in their first month. Each is a lesson or a FAQ inside a lesson.
• The product documentation. Not linked from the course, embedded into it, so the learner does not tab out and lose context.
• Two or three slides from a recent QBR deck. For the segments where you can show what good looks like using real customer outcomes.
• A five-minute welcome from a leader. Whoever the customer admin already trusts. Usually the head of CS, sometimes a founder.

The bar for the first course is not production quality. The bar is that a new admin can watch it, take the quiz, and pass. Polish comes after evidence of demand.

How should you structure the first course?

Five to seven lessons. Eight to twelve minutes each. One quiz per lesson with three to five questions. One certification exam at the end with fifteen to twenty questions.

Each lesson opens with a one-line outcome ("by the end, you can configure user roles") and ends with a two-line recap. Learners drop off when they cannot tell whether they are supposed to be taking notes or just watching.

How do you write quizzes that actually teach?

Three rules keep quiz writing from consuming a week.

• Test the workflow, not the trivia. Ask "what would you do next if a user says they cannot log in" instead of "which menu contains SSO settings". Trivia checks memory. Workflow questions build confidence.
• Explain every wrong answer. When a learner picks the wrong option, show the reason. Half the learning value of a quiz is in the explanation.
• Keep it to three to five questions per lesson. Longer quizzes get skipped. Shorter quizzes get taken but do not teach. Three to five is the range where completion stays above 70 percent.

If you cannot write ten to fifteen good workflow questions per lesson, you probably do not understand the lesson well enough to teach it yet.

How do you enroll the right customers?

Auto-enroll, from your CRM, based on lifecycle stage. Do not build a public catalog on day one.

• Trigger. New customer contract signed, or a new user added to an existing account with the admin role.
• Enrollment email. From the assigned CSM's address, not from noreply@. Two sentences and a button. The button opens the academy on your subdomain, already logged in.
• Reminder cadence. Day 3, day 7, day 14. Then stop. More reminders convert nobody and annoy everyone.
• Segment exclusion. Anyone already flagged as a power user in your product analytics does not need Admin Fundamentals. Do not enroll them.

The enrollment layer is where academies quietly fail. If a customer has to remember the training exists, most will not.

What does the launch checklist look like?

Ten items. Print it. Cross them off.

1. Course scoped to one segment, outcome written in one sentence.
2. Five to seven lessons drafted from existing content.
3. Quizzes written, three to five questions per lesson.
4. Certification exam with fifteen to twenty questions.
5. Academy hosted on your own subdomain, not a third-party URL.
6. SSO configured for at least Okta or Google Workspace.
7. Enrollment rule connected to CRM lifecycle stage.
8. Welcome email drafted, sent from a real person.
9. Completion data flowing back to CRM on an account field.
10. One CSM briefed to pilot the course with five customers before wide launch.

Miss any of items 5 through 9 and the academy is a video library, not a learning system.

How do you know it is working after two weeks?

Two numbers, one week apart.

• Week 1 after launch. Percent of newly enrolled admins who started lesson one. Target above 70 percent. If lower, the enrollment email is broken.
• Week 2 after launch. Percent of started learners who reached the certification exam. Target above 40 percent. If lower, lesson length or quiz difficulty is off.

Retention and renewal signal takes longer, usually a quarter. Do not wait a quarter to iterate. Fix the funnel numbers weekly.

The mistake to avoid

The biggest mistake is scoping the first academy like a corporate LMS project. Ten tracks, hired designer, six-month rollout, launch tied to a keynote. That version never ships, and the CS team quietly returns to onboarding customers over Zoom, one at a time, forever. Ship one course for one segment in two weeks, watch the completion numbers, and let evidence drive the second course. An academy that exists and is 60 percent as good as your imagined perfect academy is infinitely more valuable than the perfect one that lives in a Figma file.