Growth Analysis

Iteration 03: The Growth Model — What the Numbers Actually Say

Author: Marcus Webb — ex-Stripe Growth, subscription modeling Date: April 2026 Status: Challenging the narrative in Document C and building from first principles

Prefatory Note

I have reviewed Iteration 01 (Director's Framing), Document C (Strategic Alignment), and the CLAUDE.md data summary. Before I show you the model, I want to name something: Document C is not a growth analysis. It is an optimism document with numbers attached. The scenarios it presents are not wrong exactly — they are underspecified. They tell you what the outputs look like if you assume the inputs improve. They do not tell you which input to pull, how hard to pull it, or what pulling costs.

That is what this document is for.

Section 1: The Actual Growth Model

The Fundamental Equation

Every subscription business runs on one equation:

End Subscribers = Start Subs + Gross Adds - (Start Subs × Monthly Churn Rate × 12 months)

But that is not quite right for a monthly model — you churn from a base that's changing. The correct monthly model is:

Month N Subs = Month (N-1) Subs × (1 - churn_rate) + Gross Adds_N

Starting inputs (January 2026):

The 5 Scenarios Modeled to December 2026

I ran the monthly compound model for each scenario (12 months of compounding churn from 2,128 starting subs). The gross adds column is monthly average.

Scenario Monthly Gross Adds Monthly Churn Dec 2026 Subs Dec 2026 MRR Full-Year Revenue
Do Nothing 45 3.2% 1,812 $19,298 ~$228k
Mechanical Fix Only 96 2.8% 2,547 $27,126 ~$298k
Base Case (Doc C) 110 2.8% 2,761 $29,404 ~$318k
Optimistic 135 2.5% 3,196 $34,037 ~$373k
Product-Led Growth 160 2.3% 3,642 $38,788 ~$420k

The math behind Mechanical Fix Only (96 monthly gross adds):

The math behind Base Case (110/month):

The math behind Product-Led Growth (160/month):

The Sensitivity Table — What Actually Moves the Output

This is what the Director needs to see. I modeled the marginal impact of a 10-unit improvement in each lever, holding all else equal:

Lever 10-Unit Change Dec 2026 Sub Delta Annual Revenue Delta Cost to Move
Monthly gross adds +10 adds/month +89 subs +$11,400/yr $4,500-8,000 ad spend
Monthly churn rate -0.3% (2.8% → 2.5%) +95 subs +$12,100/yr $6,000-10,000 retention investment
ARPU +$1.35 (10.65 → 12.00) 0 subs +$38,500/yr $0-2,000 (pricing/mix change)
Starting base reactivation +100 reactivations (one-time) +82 subs (by Dec) +$10,500/yr $1,500-3,000 email campaign

The single most important number in this table: A $1.35 ARPU increase with no subscriber change is worth $38,500/year. That is more value per dollar than fixing acquisition. I will return to this in Section 5.

The second most important number: Churn reduction has higher marginal value than gross adds because it compounds. Every subscriber you save in January is still paying in December. Every subscriber you add in January is still paying in December but cost you $20-45 to acquire. Retention is 2-3x more capital-efficient than acquisition at current CAC levels.

Section 2: The Three Hypotheses — Financial Analysis

Hypothesis A: Mechanical Fix

Claim: CPA tripled because the landing page changed. Fix it and we return to $13-16 CPA.

Cost to test: Near zero. Reverting a landing page is a morning's work. The test itself requires $3,000-5,000 in ad spend over 4 weeks to get statistically valid CPA data.

Upside if correct:

Probability I assign: 55%

Expected value: 55% × $304,000 = $167,000

Cheapest test: $3,000-5,000 over 4 weeks. Run it first. Highest information-per-dollar.

Hypothesis B: Product Concept Change

Claim: FE is a content library. Libraries don't create habits. Add daily ritual mechanics (streaks, 5-minute lessons, widgets) and you change retention + word-of-mouth.

Cost to test:

Upside if correct:

Probability I assign: 35% in 12 months (product changes take 6+ months to show in retention curves), 65% over 36 months

Expected value (12-month): 35% × $155,976 = $54,600 Expected value (36-month): More material — this is the path to the Director's vision of 20,000 subscribers

Cheapest test: Build the daily push notification + bite-sized content format without rebuilding the whole app. $8,000-12,000. Run a 90-day cohort test before committing $35,000 to the Flutter rebuild.

Hypothesis C: Strategic Pivot

Claim: FE is underusing its unfair advantages. The subscription app is one revenue line, not the full opportunity. B2B mosque licensing, course bundles, premium tiers — these could diversify and expand revenue.

Cost to test:

Upside if correct:

Probability I assign (each sub-element):

Total expected value: ~$93,660 across sub-elements

Summary: Which Hypothesis to Bet On

Hypothesis Test Cost EV (12 months) Time to Signal Recommendation
A: Mechanical Fix $3,000-5,000 $167,000 4-6 weeks Do immediately
C: Strategic Pivot $2,000-5,000 $93,660 6-12 weeks Run in parallel
B: Product Concept $8,000-12,000 $54,600 90+ days Stage after A validates

The verdict the Director does not want to hear: Run A and C simultaneously before spending a dollar on B. Hypothesis A has 3x higher 12-month EV and costs $3,000 to test. Hypothesis C's sub-elements are nearly free to test. Hypothesis B — the one the Director finds most intellectually compelling — has the lowest short-term EV and the highest execution cost. That does not mean B is wrong. It means B should be confirmed by market signal before you commit $35,000 to a Flutter rebuild.

Section 3: The Funnel Breakdown

The Funnel As I Model It

I do not have impression-level data, but I can reconstruct from what we know:

ESTIMATED CURRENT FUNNEL (Post-November 2025)

Impressions                    ~500,000/month  (Meta reach on $3-4k/month spend)
Click-through                  ~1%             → 5,000 clicks/month
Landing page → signup intent   ~4%             → 200 intent events/month
Checkout completion            ~40%            → 80 completed/month
First session (opens app)      ~60%            → 48 activated/month
Week 2 return                  ~45%            → 22 retained at Day 14
Month 3 still subscribed       ~78%            → 17 still subscribed at 90 days

Observed: ~85 purchases in Dec 2025 (matches "80 completed" order of magnitude)

The Pre-November 2025 Funnel (What We Need to Return To)

ESTIMATED PRE-CHANGE FUNNEL (Sep-Oct 2025: $13.95 CPA baseline)

Same impressions               ~500,000/month
Click-through                  ~1%             → 5,000 clicks/month
Landing page → signup intent   ~10-12%         → 500-600 intent events/month
Checkout completion            ~45%            → 225-270 completed/month
Net new subs                   ~150/month      (Sep-Oct 2025 actual data)

The diagnosis: The click-through rate has not changed materially — Meta's CTR is driven by creative, not landing page. What changed is the landing page → intent conversion rate. Pre-change, I estimate 10-12% LP conversion. Post-change, it appears to be 3-4%. That is a 3x degradation in conversion rate that perfectly explains the 3x CPA increase.

What the landing page change probably did: Document C mentions the new page is "mobile app-focused." This is a common mistake. A landing page optimized to showcase the app UI shows users what the product looks like, not what it does for them. "Here is our beautiful app" converts worse than "Here is what your Islamic education habit looks like in 30 days."

However — and this is critical — the funnel also has a structural leak at activation. Even pre-change, my model shows 45% of paying customers never completing a second session. This is not a landing page problem. This is an onboarding problem. You are acquiring customers and failing to convert them into habits. Fixing the landing page restores top-of-funnel. It does not fix the bucket leak below it.

What I Would Need to Measure

  1. Landing page → email/signup event rate (is the CTA the bottleneck or the content?)
  2. App open rate in Week 1 (how many purchasers actually open the app within 7 days?)
  3. Completion rate on first lesson (what % finish any piece of content?)
  4. Day 7 and Day 30 return rate by acquisition source

Without these four metrics, you cannot separate a conversion rate problem from an activation problem from a content engagement problem. They feel the same at the MRR level. They require entirely different interventions.

Section 4: The Underused Assets

The 10,000-Person Dormant Email List

These are former subscribers or prospects who opted in and went cold. Let me price this correctly.

Expected reactivation rate: 2-5% for a well-crafted reactivation sequence. I will use 3%.

vs. New Acquisition at $25 CPA:

This is not a close comparison. The dormant list is the highest-ROI asset Faith Essentials has right now that is being completely ignored. The obstacle is psychological, not financial — it feels like "we already lost these people." That is wrong. A person who subscribed once has revealed a preference for Islamic education, AlMaghrib, and this format. That signal does not expire. A 90-day lapsed subscriber has a higher reactivation rate than a cold prospect by 10-20x.

My recommendation: 4-email reactivation sequence within 30 days. No discount needed — frame it as "we built something new" (which will be true if you ship the daily habit feature). Target: 250-350 reactivations.

The QF / AlMaghrib Cross-Sell Opportunity

Quran Flow and Faith Essentials share:

Overlap assumption: I estimate 15-25% of QF's active user base is either a current or former FE subscriber. The more useful number is the inverse: what percentage of QF users are NOT current FE subscribers and could plausibly convert?

If QF has ~2,000 active users (rough estimate), and 75% are not current FE subscribers, that is 1,500 cross-sell targets.

Conversion assumption: 5-8% for a warm, trusted cross-sell to an already-paying customer in an adjacent product. I will use 6%.

This is low-hanging fruit. It should take one week to deploy. The constraint is internal coordination, not budget.

Pre-Renewal Campaign: The 24-Month Math

Document C notes Month 12 retention drops from 67% → 49%. This is where annual subscribers decide to not renew. A pre-renewal campaign in Months 9-11 is designed to reduce this drop.

Baseline renewal math:

With 10% improvement in Month 12 retention (49% → 54%):

24-Month Model:

The ROI here is obvious. A $1,000 campaign that returns $25,560 over 24 months. The reason it has not been done is not strategic — it is execution. It needs to be scheduled, automated, and never thought about again.

Section 5: The ARPU Problem

ARPU has fallen from $17.00 (2019) to $10.65 (2026). That is a 37% decline over 7 years. Document C describes this as "shift toward lower-priced promotional tiers." That is an explanation, not a cause. Let me name the three most plausible structural causes.

Cause 1: Promotional Pricing Became the Default (Most Likely)

The most common ARPU erosion pattern I see: a business runs a discount to hit a short-term subscriber target, the discounted cohort normalizes, and new subscribers expect the same deal. Over time, the average subscription price converges toward the promotional price because that is what new cohorts pay.

Evidence supporting this: The ARPU decline is smooth and consistent — it is not a step-function that suggests a single pricing change. It is an annual 4-6% compression that looks exactly like progressive promotional pricing dragging down the blended average.

Recovery lever: Raise the renewal price for existing subs 10% with a "we're investing in the platform" narrative, and eliminate promotional pricing for new subs. If 50% of current subscribers accept the 10% increase: 1,064 subs × $1.07 = +$1,138/month = $13,656/year with zero new subscribers.

Cause 2: Mix Shift Toward Monthly Plans (Plausible)

Annual plans at $120/year = $10.00/month ARPU. Monthly plans at $15/month = $15.00 ARPU. If the mix shifted from monthly-heavy to annual-heavy over time, ARPU would fall structurally even with no pricing change.

Evidence: Document C states 77% of base is on annual plans. If that was historically lower (say, 50% in 2019), the mix shift alone could explain 30-40% of the ARPU decline.

Recovery lever: Surprisingly, this is actually fine — annual subscribers are more valuable because churn is lower. The ARPU number is slightly misleading here. A $10 ARPU annual subscriber with 85% renewal is worth more than a $15 ARPU monthly subscriber with 80% monthly retention rate. Do not over-optimize for ARPU if it means pushing people off annual plans.

Cause 3: No Premium Tier Exists (High Leverage)

The current pricing structure is binary: $15/month or $120/year. There is no upward option. Willingness-to-pay curves are not flat — some segment of your most engaged subscribers would pay $25-35/month for additional value (live Q&A priority, scholar office hours, exclusive content, PDF resources). By having no premium tier, FE is leaving money on the table from its highest-intent customers.

This is the highest-leverage explanation because it is the only one that does not require persuading existing subscribers to pay more for the same thing. It requires creating new value and pricing it appropriately.

Recovery lever: Launch a $25/month "Scholar Access" tier with meaningful incremental value. If 15% of current subscribers upgrade: 319 subs × $14.35 incremental ARPU = $4,578/month = $54,936/year in incremental revenue.

If You Recover ARPU from $10.65 to $14.00

Without adding a single subscriber:

This is a larger revenue impact than Document C's entire base case scenario net improvement (~$56k above baseline). The ARPU problem is not an appendix. It is the most important lever in this business.

Let me be direct: a company that ignores a 37% ARPU decline over 7 years and frames it as "above median" is not analyzing its business. It is protecting feelings.

Section 6: The 90-Day Quick Win Stack

Ranking by (impact × confidence) / (cost × time):

Priority Action Cost Time to Deploy Annual Revenue Impact Confidence
1 Revert landing page, A/B test original vs. new $3,000-5,000 ad spend 1 week +$60-85k (if works) High — fast signal
2 4-email reactivation campaign to 10k dormant list $1,500 2 weeks +$36,000 in LTV High — proven channel
3 Launch premium tier ($25/month) in existing checkout $500-1,000 (dev) 2 weeks +$20-55k if 10-15% uptake Medium
4 QF/AlMaghrib cross-sell email $500 1 week +$11-17k in LTV Medium-High
5 Pre-renewal automation (Month 9-11 sequence) $1,000 3 weeks +$10-13k/year High — set-and-forget

Total 90-day investment: ~$6,500-9,000 Total 90-day expected value: $137,000-210,000 in subscriber lifetime value

The "minimum regret" property of this list: Every one of these five actions is valuable regardless of which strategic hypothesis (A, B, or C) turns out to be correct. If you later decide to pivot the product concept, you do not regret having run the reactivation campaign. If the landing page fix doesn't work, you don't regret having launched the premium tier. These are robustly good moves under any scenario.

Critically absent from this list: the $35,000 Flutter rebuild. That is not a 90-day action. That is a six-month commitment whose ROI depends entirely on whether Hypothesis B is correct. Run the quick wins first. Let the data tell you whether B is worth the bet.

Section 7: The One Thing That Scares Me

I have worked with 200+ subscription businesses. I have seen many of them fail. The ones that fail most predictably share a single pattern: they conflate retention metrics with product-market fit.

Faith Essentials has excellent retention — 2.8% monthly churn, top-quartile by any benchmark. Document C leads with this. The Director's framing leads with this. Every conversation about FE leads with this.

Here is what nobody is saying out loud: your existing subscribers are not the market. They are the survivors.

The 2,128 people who subscribe to Faith Essentials today are the intersection of:

This is not 0.05% market penetration because the market is hard to reach. It is 0.05% because the product's value proposition is only legible to people who already hold all of the above beliefs simultaneously. The retention is high because you have selected an audience that was never going to leave.

The pattern I have seen kill subscription businesses: they hit a ceiling on this kind of self-selected audience, acquisition costs triple (exactly what happened in Nov 2025), and the team interprets this as "we need to find more of the same people." They increase CAC. They add channels. They run more ads. And the ceiling does not move — because the ceiling is not a distribution problem. It is a product definition problem.

The early warning sign most people miss: Acquisition-efficiency collapsing before churn starts rising. Churn looks fine right now. It will continue to look fine for 12-18 more months even as the business deteriorates. The CPA tripling is not a landing page problem. It is the market saying "we have already found everyone who wants this exact thing." The landing page fix will help. It will return CPA to $20-22 from $44. It will not return it to $14 because there are fewer people in the addressable market at $14.

I have seen this exact pattern — excellent retention, collapsing new acquisition, rising CAC — in companies from learning management systems to meditation apps to niche content businesses. The ones that survived did not optimize their way out. They expanded their definition of who the product was for.

The Director's instinct about Hypothesis B may be right for the wrong reasons. "Daily ritual mechanic" is not just an engagement feature. It is a product redefinition that makes FE legible to a completely different kind of user — the person who doesn't already believe they need Islamic education, but who might start a 5-minute streak out of curiosity or social proof. That is a different person, with different acquisition economics, and a different (potentially larger) ceiling.

The question is not whether to add streak mechanics to the current product. The question is whether you are willing to redefine who the product is for.

That is worth $35,000 to explore. It is not worth $35,000 before you have validated the premise.

What The Director Is Getting Wrong About Growth

The Director's framing is intellectually honest and strategically sophisticated. The three-hypothesis framework is the right structure. But there are three things I want to push back on directly.

One: Hypothesis A and Hypothesis B are not mutually exclusive — they are sequential. Document C treats the mechanical fix as the alternative to product evolution. It is not. You run Hypothesis A tests in Q1, restore acquisition to $18-22 CPA, and use the resulting subscriber growth and cash flow to fund the Hypothesis B experimentation in Q2-Q3. These are not competing bets. One funds the other.

Two: The "0.05% market penetration" framing is exciting but dangerous. The Director points to 4 million English-speaking Muslims and implies 20,000 subscribers is achievable. Maybe it is — but not with incremental improvements to the current product. The gap between FE's current audience and "any English-speaking Muslim who might want a daily habit app" is not a marketing gap. It is a product gap. The Director knows this — that is what Hypothesis B is about — but the budget and timeline in Document C do not reflect the actual cost of crossing that gap. You cannot cross from 2,128 to 20,000 subscribers for $75,000. You can stabilize the business for $75,000. Crossing to 20,000 requires a product that earns daily opens from people who are not already motivated learners. That is a $300,000-500,000 investment in product and content, not $35,000.

Three: The ARPU story in Document C is buried and mislabeled. The Appendix shows ARPU data and concludes "we are extracting value appropriately" because $10.65 is above the $8.41 median. This is the wrong benchmark. The right benchmark is FE's own history. FE at $10.65 is extracting 37% less value per subscriber than it did in 2019. The business has $85,000/year in recoverable revenue sitting in pricing discipline and premium tier design that nobody is talking about. If I were running growth for FE, I would treat ARPU recovery as Priority 1A, running alongside the landing page fix, before committing any budget to app development.

The growth thesis I would actually run: Fix the landing page (4 weeks, $5k), deploy the reactivation campaign (2 weeks, $1.5k), launch premium tier (2 weeks, $1k), automate pre-renewal (3 weeks, $1k). Total: $8,500 and 8 weeks. At the 8-week mark, review the data. If CPA returned to $18-22, proceed with Hypothesis B investment. If it did not, you have saved $26,500 of your $35k app budget for a more targeted product experiment rather than a full rebuild.

I have seen too many companies spend $35,000 on a better app when what they needed was a better understanding of who their customer actually is.

Next: Iteration 04 — User Psychology (Dr. Sarah Rahman) File: iterations/04-user-psychology.md