The current Linear baseline, the decisions that unblock MVP1, and the path to a simple paid tutoring or low-ticket entry.
Linear owns issues, owners, estimates and sprint commitments. This roadmap explains the release path and the product decisions around it. Editorial revision: 12 October 2026 (UTC).
Reading the current board… Progress unavailable until the read completes.
A burnup compares total scope with completed work, so added scope is visible instead of looking like slower delivery. Use the existing Linear project graph for historical progress. This read-only view has today's baseline; historical project membership, scope added/removed and acceptance history are unavailable. No past trend or finish forecast is inferred from issue creation dates.
Counts show tickets, not estimated effort. Project membership is not yet a confirmed release exit scope. Deployed and accepted work needs environment, release evidence and a named reviewer in Linear; Done alone does not prove it.
Editorial context, separate from the live execution state. Source cutoff: #app-dev through 9 October; Linear is read at page load.
SUP-33, SUP-34 and SUP-35 cover admin cancellation/refund control, rescheduling and overlap locking. The last verified Linear read marked them Done and assigned to Wenbo, with a staging-to-production PR linked. Production deployment and end-user acceptance remain unverified.
BAMFAM (SUP-19) has a production announcement from 26 September; admin first-booking was reported 29 September. Its ticket remains partly open. Reconcile the remaining criteria rather than treating the entire feature as unbuilt.
Scheduling issue in LinearSUP-26/27/28/29/31 are component builds marked Done in the last verified read. Jay was still testing uploads and preparing the walkthrough, Wenbo review and team feedback. SUP-30 (admin uploads, tutors verify), SUP-32 (known gaps) and SUP-13 (integration) remain open in that read.
Keep MVP2 workflow, timers/accommodations, notifications and release QA visible. Built components do not establish a released assessment journey.
Assessment workflow in LinearA product recommendation for the week, pending scope and engineering acceptance in Linear.
Resolve parent/family rules (SUP-9), one student channel and its membership (SUP-10), and self-registration (SUP-21, currently outside the project). Decide whether SUP-18 invitation tracking and the supplementary booking issues are required for this release.
Joe + John decide the product boundaries; Wenbo confirms decomposition, estimates, owners and the cycle. These are decision roles, not new ticket assignments.
Verify SUP-33–35 and the remaining SUP-19 criteria against the named staging/production environment. Test overlapping requests, allowed roles, credits/refunds, rescheduling and all calendar confirmations. Record evidence and the acceptance reviewer before calling the release shipped.
For direct tutoring: create account → pay → book a qualified tutor → attend → receive the summary. Resolve trial/payment policy and family credit ownership first. For low ticket, choose one concrete paid deliverable and its first-value event; a simplified AP game plan is a candidate, not a settled offer.
Capture discoveries in Linear Backlog. Before adding one to a sprint, agree the owner, acceptance criteria, estimate and what moves out or which target changes. Keep MVP2 work separate unless its inclusion is accepted. There is no second editable task tracker here.
Proposed operating definitions. Actuals are not measured here, and targets are not yet accepted team commitments.
| Weekly KPI | Definition / proposed bar |
|---|---|
| Release scope ready | 100% of included release items have owner, criteria, estimate and agreed target; include required work outside the project. |
| Accepted commitments | Deployed + accepted items / the frozen original weekly commitments. ≥85% provisional; calibrate after two weeks. Scope removed is shown separately. |
| Critical defects | 0 open release blockers; newer blockers count too. |
| Critical journey tests | 100% of a finite agreed required suite pass with evidence; untested required cases remain in the denominator. |
| Booking/payment integrity | 0 wrong credits, duplicate charges/grants, overlaps or orphan paid purchases in a nonzero audited population. No audit means unknown. |
| Bookable subject coverage | 100% of advertised subjects have a qualified slot within a proposed 7-day promise; confirm that promise before launch. |
Pilot later, separately by offer: checkout completion within 24 hours; paid-buyer first value within 7 days; unplanned manual repair within 7 days. Use mature cohorts and server-confirmed payment. Low-ticket first value must be defined before scoring it.
Research basis: DORA delivery metrics for team-level baselines, and Stripe fulfillment for payment truth and idempotent delivery. Release frequency alone is not the launch gate.
Generated on demand. The recording and transcript reflect their own date, not the current board. No automatic weekly run is confirmed.
Loading the dated archive. Current progress is above.
Five stages. Each one is defined by a capability the business did not have before it, not by a list of features.
Click a stage to open it. Live issue status is distinct from release evidence.
Mar – Jun 2026 · the thesis that started the build
Immediate priority · confirm release scope and acceptance in Linear
Historical reports are not fresh acceptance tests. See the issue evidence and validate the current journey.
Next release layer · built components still need journey acceptance
Jan – Apr 2027 · revenue that does not need a sales call
2027+ · the data starts teaching
D-1: account creation and alerts are the agreed direction. Pay-first versus trial was Joe's leaning, not a final prohibition on free trials. Joe + John must settle the offer before SUP-21 is accepted.
SUP-9 needs two-parent/two-child linking and permissions; SUP-10 needs the channel creation and membership rules. Do not ask families to share account passwords.
The existing tutoring unit is 30 minutes per credit. Test-prep balances, overtime with insufficient credits and no-show/refund rules still need explicit decisions. SUP-30 must implement admin upload and tutor verification; timers/accommodations, notifications, content approval and integration remain acceptance gates.
Nov 15 core release and Dec 15 readiness remain product targets, not engineering commitments. Confirm estimates, scope and capacity in Linear before promising a finish date. The prior 60–70% migration estimate from September is not a current allocation or forecast.
Use the existing MVP project graphs in Linear for scope history. New release and acceptance evidence belongs on the issue; this view should read it rather than create another operating system.
Reading the current Linear SUP board… Cache age and source time will be shown.
Loading the board…
Who decides what gets built, who builds it, and how what the team hears from real students gets back into the plan. This is the structure Joe agreed, drawn.
Frontline issues generally enter through Rikki and/or Jay and get classified into one of four things. Only the fourth is a roadmap question.
John runs a first principles and jobs to be done review to work out what problem actually needs solving, who is affected, whether it is a bug or a missing capability, whether the requested feature is even the right solution, whether it belongs in the current release, and what adding it displaces.
Linear is the execution system of record: roadmap work, issues, priorities, sprint commitments, owners, status, blockers and dependencies. Slack is a communication and escalation channel, not the roadmap.
Ideas are welcome. Unpriced ideas are not. Every request runs the same route: grilled against the job it serves, vetted against what already exists, staged, and only then allowed near a sprint. Nothing enters a running sprint without naming what leaves.
Slack, a Loom, a call, or the assistant below.
Who is it for, what job, what they do today, how often, and what happens if we never build it.
No → parked, with the reason recorded.↳ Rarely. Vague goes to more-vetting, not rejected.
Duplicate of something open? Already staged? Then MVP1–4, or unranked.
No is the default. It lands in Backlog, staged and unestimated.↳ Yes only if you name what leaves to make room.
Existing today: Linear is the execution record; this page reads it. The existing intake assistant files only to Backlog when a human presses its explicit file button. It never assigns, estimates or puts a request into a cycle. Proposed human flow:
Urgent production bugs: alert the existing technical triage owner directly through the established escalation route, then document the reproduction, impact, issue and mitigation evidence in Linear. Urgency bypasses the normal batch, not documentation or release checks. No new meeting or automatic enrollment in MVP1 is implied.
Problem / affected role: Expected behavior → actual behavior: Example or reproduction (test data only): Impact / urgency / affected users: Observable acceptance criteria: Existing Linear issue or duplicate candidate: Defect / agreed MVP1 blocker / later enhancement: Decision needed / scope displaced or date effect: Wenbo review: feasibility, estimate, dependencies, owner, cycle Release evidence: environment, test result, reviewer
Fictional example: In a test family with student A and student B, a parent opens B's session and sees A's summary. Expected: only the selected student's permitted summary. Reproduce with synthetic accounts. Draft acceptance: B opens B's summary; an unrelated parent is denied; switching children never retains the prior summary. Orchestration checks existing SUP-9 first. John confirms whether this is an agreed MVP1 blocker; Wenbo sizes the fix. No live incident, new ticket, estimate or commitment is asserted.
Ideas are welcome. Untraded ideas are not. Every addition after a sprint is agreed costs something, and the cost gets named before the yes.