Faith Essentials 2026 — Operator's Brief

§1The situation, in 30 seconds

Faith Essentials has been approximately flat at $22–25k MRR (monthly recurring revenue) for three years. In 2025 that pattern broke downward: new subscribers fell from 1,264 (2024) to 696 (2025), a 44.9% year-over-year collapse. Annual revenue dropped 17% YoY, from ~$317k to ~$264k. Nine of twelve months in 2025 saw net-negative subscriber change.

Under the surface the picture is less uniform than the headline suggests:

The portfolio context that changes the frame

FE is one of several AlMaghrib Institute products. Seen inside the portfolio, its size relative to adjacent revenue lines matters for strategic options:

This portfolio context matters because it opens a strategic option (Path D, below) that wouldn't exist for a standalone company: FE could be repositioned as a retention layer for the seminar business rather than a standalone DTC product. Whether that's the right move depends on data we don't yet have — but the option should stay on the table.

§2The three root causes

Team 1's Five Whys analysis reached three root causes, all with independent evidence. Any 2026 plan has to address at least one primarily; the recommendation below addresses all three in different ways.

§3The recommendation, in one sentence

Install Path A's Week 1 mechanism (a 2×2 landing-page/creative A/B test + restored ad spend if it wins) as the Day-1 default action UNDER Path C's measurement scaffold (five pre-committed Decision Rules + biweekly Decision Review + a 15-instrument pipeline over 90 days), with Path B (product thesis work) and Path D (AlMaghrib ecosystem integration) held as conditional Month-4 and Month-9 successors whose commit decisions will be data-informed rather than today-guessed.

In plainer language: we fix the acquisition leak immediately using Path A's test, we build the measurement infrastructure that should have existed for years, and we keep two alternative strategies on standby for when we know more. Nothing is committed today that we can't reverse by Month 4. The rest of this brief walks through what that looks like week by week.

How this recommendation was decided

Four strategic paths were developed in parallel, each by a different dedicated advocate, with mandatory adversarial cross-challenge. Each was scored against a six-criterion decision framework. Final weighted scores:

Path C won because it's the path that scales with whatever the measurement reveals. But the Path C scoring already incorporates Path A's Week 1 action as the default move — that's the hybrid the recommendation formalizes. Path A alone would beat Path C on near-term speed; Path C alone would beat Path A on long-term learning. Running them together, with one cheap Path D optionality test in parallel, is what this brief actually describes.

§4Week 1–2: three parallel tracks

Three independent tracks, each with a distinct owner and a distinct assumption being tested. All three launch in the same 2-week window because they use different people and resolve different questions — running them in series would add a month of lost time with no correspondence benefit.

Total Week 1–2 out-of-pocket: approximately $6,200. Resolves four load-bearing assumptions (A21 Nov-5 landing-page mechanism; A16 decision-rule compliance; A39 seminar demand; A38 seminar/FE overlap).

§5Three free ops queries (Week 1)

Each of these is one email or one internal conversation. Collectively they tighten the Month-4 decision inputs dramatically at zero marginal cost.

§6The CEO conversation

This recommendation requires three CEO-level decisions. The conversation should happen by Friday, April 24, so Track 1 and Track 3 can launch the following week. Track 2's Decision Rules need to be drafted before this meeting, not after, so the conversation has something to sign off on.

Three decisions to ask Noor for

1. Scorecard re-sequencing. Defer Q3's daily-habit product rocks (widgets, streaks, daily-content pipeline, live/interactive programming) to Q1 2027 pending data. Keep Q2 app-revamp requirements work on schedule but delay the dev kickoff by 60 days (new target: dev kickoff August 1). Rationale: the daily-habit thesis that those rocks assume has not been user-validated, the app rebuild will cost the same to ship later with a validated thesis, and three months of measurement data will tell us whether to build for daily-habit users or library-completers.
2. External-contractor authorization. Pre-authorize up to ~$6,000 in external contractor engagement for (a) landing-page duplication for the 2×2 A/B, (b) measurement instrumentation build (Mixpanel / Metabase / dashboard setup), and (c) Eventbrite → FE provisioning scoping if Track 3 validates demand. This is already a pre-approved pattern (the Hamza precedent from Mar 31). We're asking Noor to sign off on the specific spend envelope so Samia and Kamran don't have to return for each decision.
3. Content-lead scope. The April-30 content-lead hire proceeds as planned, but the job spec is updated to include (a) measurement-aware content production — every piece tagged with an experiment ID and tracked for activation/completion outcomes; and (b) optional seminar-integration curriculum curation if Path D's Gate 1 test validates. This is a one-paragraph addition to the existing role description, not a redirect of the search.

Script outline (15-minute meeting)

1. Frame (2 min). "Team 2's analysis changed the diagnosis. Three things the prior exercise treated as problems turn out not to be problems (churn is improving, ARPU is rising, scholarship is a non-issue). One thing is the active crisis: acquisition volume dropped 45% in 2025, driven by two specific decisions. Unit economics are healthy; we have a volume problem, not an economics problem."
2. The structural issue (2 min). "There's one problem we didn't know we had: about half of every annual cohort, for the last four years, doesn't renew at month 12. That's been true under every product configuration we've run. We don't know why because we've never asked a churned subscriber. That's what the exit-survey piece of this plan is about."
3. The plan (4 min). Walk through the three tracks. Emphasize the parallel structure and the $6k total cost.
4. The three asks (4 min). Walk through each decision. Explain the fallback for each if denied.
5. Timeline commitment (2 min). Month 4 gate (July) will report back with acquisition results, measurement infrastructure status, exit-survey findings, and the Gate 1 demand result. Month 9 gate (November 15) is the 2027 roadmap decision point.
6. Decision (1 min). Ask for sign-off.

Fallbacks if any decision is denied

§7The five Decision Rules

These are pre-committed: if a metric trips its threshold, the named action happens, and the named owner executes. They are signed by Noor before any dashboard work begins. The point is not the dashboards; the point is the pre-commitment to act when the data says something is wrong. Without that, measurement replicates Rayyan's pattern (data exists, nothing changes).

Three rules are operational on Day 1 (DR-1, DR-2, DR-5). Two are placeholder-gated on instrument build (DR-3 and DR-4) — activated once the underlying measurement lands.

§8The biweekly Decision Review

This is the mechanism that separates "we have dashboards" from "we make decisions from dashboards." It's a 30-minute standing meeting, chaired by Kamran, with fixed agenda and public minutes.

Fixed agenda (30 min)

1. DR check (20 min). Walk through DR-1 through DR-5. For each: current number, threshold, status (green / watching / tripped).
2. Tripped-rule action (up to 5 min per trip). Named owner states what action was taken (if already tripped in a prior cycle) or will be taken this cycle.
3. Standing questions (5 min). One open instrumentation-research item per review (see §13). One open assumption being tested (see §17).

Attendance and accountability

• Required attendees: Kamran, Samia. Content lead joins from hire onward.
• Noor attends monthly (every other Review) for first-order signal.
• Minutes posted to shared Slack channel within 1 hour, every cycle.
• Falsification trigger: if attendance drops below 80% across any 3 consecutive cycles (i.e. two people missing or one person out for 2+), Path C's scaffold is flagged as inoperative and we revert to pure Path A. This threshold is the self-falsifying mechanism that makes the assumption "dashboards will be used" testable rather than aspirational.

Schedule target: first Review May 15 (Track 1 A/B result due); then every two weeks. Cycles 1–6 run from May 15 through August 7 (first three months of the scaffold).

§9Timeline & decision gates

Key dates and gates through Q2 2026 – Q1 2027. Month 4 and Month 9 are the two formal decision gates.

What happens between gates

§10Month 4 decision tree

By July 15, 2026 we have: Week 3 A/B result; Gate 1 demand-test result; overlap query result; ~60–90 exit-survey responses; 5–6 biweekly Decision Review cycles of operational data; WAU/MAU baseline from instrument I-7. The Month 4 Gate decides where to deploy Phase-3 effort (August onward).

§11Reversal conditions

Specific triggers that would cause this recommendation to be abandoned and a different approach chosen. Each one has a trip wire that makes the reversal visible rather than implicit.

1. Both A/B pages produce CPA ≥$40 at Week 3 result. Assumption A21 (Nov 5 landing-page is primary CPA mechanism) is refuted. Path A's core bet fails. Re-open Path B or Path D as primary candidates rather than Path C — because Path C's scaffold assumed Path A's mechanism would be the Day-1 default action.
2. Noor rejects all three Week-1 asks (Scorecard re-sequencing, contractor authorization, content-lead scope). Revert to pure Path A at 0.7 FTE without measurement infrastructure. Document the constraint. Not a failure of the recommendation — an operating constraint we now know we have.
3. Across any 3 consecutive biweekly Decision Reviews attendance drops below 80%, OR a tripped DR goes without action for 1 full cycle. Self-falsifying trigger in the scaffold mechanism. Abandon measurement-discipline layer; continue Path A alone. (Per §8, the attendance threshold spans 3 cycles so one missed meeting doesn't falsify the scaffold prematurely.)
4. Overlap query returns ≥60% AND seminar attendance ≥8k AND Gate 1 demand >5%. Reopen Path D for reconsideration at a higher weight — the data would show that "new-audience acquisition via seminars" is false BUT "existing-subscriber retention via ecosystem touchpoints" is a real and measurable mechanism. That's a Path D variant worth committing to earlier than the current Month-4 default.
5. Month 4–6 data shows A/B validated AND Mo3 activation <30% AND exit-survey shows strong single thesis across segments. Authorize Path B at Month 5 rather than waiting for Month 9. The thesis signal is strong enough to commit earlier.

§12Projected outcomes vs. do-nothing

Three scenarios, all measured against the same starting point (2,128 subscribers, $22,666 MRR, January 2026) and the same 12-month horizon.

12-month subscriber count. Do-nothing per DATA-BRIEF §6 Scenario A. Success case assumes A/B validates + volume restored to 100/mo + mix-shift compounds. Fallback assumes A/B mixed + Path C scaffold still protects downside.

§13What we are deliberately NOT doing

Each of these is a thing the Scorecard, prior strategy exercise, or common instinct would do. Each is deferred or declined for a specific reason.

§14Implementation research TODOs

This recommendation depends on getting measurements that FE has never instrumented. Most of them are "one conversation" or "one afternoon of research" away from being solved — but that work hasn't been done. The list below is what the team (or an agent with the right access) needs to investigate, in rough priority order.

These are not blockers for Week 1. They are things to figure out in parallel so that by Week 4 the instruments can actually ship.

The instructor-brand paradox — a content constraint worth naming

A separate structural issue that affects creative rebuilding under Path A and content decisions under Path B. It's not a TODO but a constraint to carry:

The paradox: the instructors most likely to attract new subscribers in ad creative have the least content available; the instructors with deep content are less recognizable to non-AlMaghrib audiences. Path A assumes existing content is enough creative raw material for 12 months of acquisition; Path B would have to decide whether repositioning respects this constraint. The content lead's scope should include "cross-instructor clip assembly" as a creative workaround — short-form ads that cut between multiple instructors rather than single-instructor features.

§15Budget allocation

Approved 2026 FE allocation: $82,000 total ($37k Q2 + $23k Q3 + $22k Q4). Operating stance (per Kamran's March 31 comment): overspend and ask for more if CAC economics justify — so this is an opening budget, not a hard ceiling.

Budget elasticity — what the CEO conversation should clarify

The "overspend and ask for more" stance means that paid acquisition can scale beyond $24k if the A/B test validates a recoverable CPA and volume economics support it. The Week 1 CEO query about whether the $82k is guaranteed or Q2-contingent is material: if contingent, Path A's mandate is stronger because it shows revenue earliest; if guaranteed, Path C's 90-day measurement ramp is lower-risk.

§16Team roles & allocation

§17Key assumptions being tested

Every plan rests on assumptions. These are the five that most directly determine whether this recommendation works, ordered by how much a refutation would reroute execution.

One assumption was refuted mid-process (context for everything above)

§18Appendix: ground-truth numbers

Core quantitative picture as of January 2026. All figures cite primary source (DATA-BRIEF section references use Team 1's numbering; underlying data in reference-docs/ CSVs).

§19Appendix: how this recommendation was derived

This brief is the operational surface of a two-team, four-phase strategy exercise conducted April 18, 2026. The full audit trail is in 2026-strategy/agent-teams-v2/.

1. Team 1: data & root cause. Two analysts processed all three primary-data CSVs (Subscription_metrics, Customer_MRR_changes, MRR_per_Subscriber) — two of which had never been opened by the prior strategy exercise. Produced DATA-BRIEF.md (11 sections) and ROOT-CAUSE.md (Five Whys, three branching root causes). Mid-process discovery: the "scholarship destroys ARPU" premise of the prior exercise was refuted; only 9 subscribers, 0.2% of MRR.
2. Team 2 Phase A: path proposal. A dedicated Skeptic read DATA-BRIEF + ROOT-CAUSE + the Strategic Review + Team Scorecard and proposed four MECE strategic paths. Included the explicit Team Scorecard pre-commitment tension (Q2/Q3 daily-habit rocks) as a design constraint every path must address.
3. Team 2 Phase B: path development. Four separate Advocates each developed one path fully: 6–9 month plans, data-cited arguments, pre-mortems, cheapest-validation tests, assumption registers. Each advocate submitted a plan for approval before writing the full argument.
4. Team 2 Phase B2: adversarial cross-challenge. Each advocate issued data-backed challenges to the other three paths. User-Voice challenged every path's user-behavior claims using transcripts of team meetings plus competitive/App Store evidence. Skeptic maintained the assumption register and tracked refutations. Material mid-process refutations: A22 (CPA linearity) partially validated but reframed; A23 (intrinsic churn improvement) REFUTED via cohort analysis. Two user corrections (ad budget elastic, external contractor pre-approved) loosened constraints all four advocates had assumed.
5. Phase C: synthesis. Six-criterion weighted scoring. Path C won on weighted score (4.25/5). The recommendation integrates Path A's Week 1 mechanism (via Path C's own concessions) and Path D's Gate 1 test as a parallel optionality — then defers Paths B and D to Month 4/Month 9 gates as conditional successors.

Why the recommendation is a hybrid (not a single path)

During Phase B2, two of the four advocates (Path A and Path C) converged on the observation that their recommendations weren't actually competing. Path A's advocate conceded Path A is "a 9-month bridge, not a destination" and instrumented an explicit Month 9 re-evaluation falsification trigger. Path C's advocate conceded that Path C's Week 1 action must match Path A's default (because DR-2's CPA threshold is tripped on Day 1). The hybrid in this brief formalizes that convergence: Path C's scaffold runs, Path A's mechanism is what the scaffold decides to do first, and Paths B and D are conditional successors that activate at Month 4 based on what the scaffold reveals.

Confidence levels

• High confidence on diagnosis (three root causes; A23 Mo12 cliff; two specific 2025 acquisition events).
• High confidence on Week-1 action design (parallel tracks with different owners + different assumptions being tested).
• Medium confidence on Path A's CPA-recovery range ($18–25 if A/B validates). Dependent on the test result.
• Medium-low confidence on Path C's scaffold actually changing team behavior. This is A16 — the biggest non-test assumption in the plan. The biweekly-Review attendance threshold is the falsification mechanism.
• Low confidence on the Mo12 cliff being solvable in 2026. The recommendation buys time and information; it does not commit to a solution.

What would change this recommendation most

Three data inputs, any one of which would materially alter the plan:

1. Actual seminar attendance & overlap. If ops confirms seminar attendance is below 4,000 unique/year or overlap ≥60%, Path D collapses to a specific retention-only variant. If attendance is ≥10,000 and overlap is ≤30%, Path D deserves more Week-1 investment.
2. Cash-runway status. If the $82k allocation is Q2-performance-contingent rather than guaranteed, Path A's mandate is stronger (revenue earlier) and Path C's 90-day ramp is riskier.
3. Any subscriber exit-survey pattern we don't yet see. If Mo12 churners mostly say "I finished what I came for" (library completion), the strategic direction is very different from "I didn't know what to do" (onboarding failure). The Week-1 exit survey is the first real window on this.