1000 Signups. Zero Revenue | SubSub
Diagnosis and 90-day plan for SubSub -- a YouTube creator monetization platform with 1,000 signups and zero paid conversions | Funnel analysis, activation gap, hypothesis prioritization, unit economics, experiment roadmap | By Yevhenii Holovei
Problem Framing
What Might Actually Be Happening
1,000 signups / 0 paid conversions is not a single problem -- it is a compound failure with multiple plausible root causes
Users sign up, trying to figure out what this product is about and how to use it, feel lost and go. No single JTBD is the clear hero. Users don't know what this product is for.
Users sign up, connect a channel, see some analytics -- and never reach a moment where the paid plan becomes obviously necessary. "Aha" is unclear and probably takes too many clicks.
Mass-market creators with 10K–500K subs are mostly hobbyists or semi-pro. $149/month is a meaningful expense. The value has to be clearly revenue-generating, not "insight-generating".
Freemium → $149 is a cliff, not a ladder. Wrong ICP signal from traffic: organic, LinkedIn etc. brings curious visitors, not necessarily creators with intent and budget.
How to Determine Which Hypothesis Is Real
Triangulate with small cheap signals rather than waiting for a perfect dataset
| Hypothesis | Cheapest Signal That Confirms / Denies | Time to Answer |
|---|---|---|
| Wrong ICP | Cohort-slice current 1,000 signups by channel size band. Look at the retention curve per band. If 10K–50K sub creators churn out and 100K+ retain, the ICP is narrower than marketing thinks. | 1 day |
| Weak value prop | 10 × 30-min user interviews (5 active free users, 5 churned). Single question: "What did you think this product would do for you?" If answers vary wildly, value prop is unclear. | 1 week |
| Onboarding issue | Session replays (Hotjar / Clarity) of 50 signups. Count drop-off between "account created" and "first meaningful action". If >60% drop before first action, onboarding is the bottleneck. | 2–3 days |
| Pricing / packaging | Look at pricing-page exit rate and "plan clicked" rate. Run a paywall A/B test with $79 and $49 trial entry points, weekly subscription. | 2 weeks (test) |
| Product quality | Support tickets + NPS from 1-month active users. If quality is the problem, negative signal will show up here. | 1 day |
Working with uncertainty.
Funnel & Metrics
Core Funnel: Signup → Payment
Seven discrete steps -- each with a specific "done" definition so the data team can instrument without ambiguity
Leading & Lagging Indicators
| Type | Metric | Why It Matters |
|---|---|---|
| Leading | Activation rate (7-day) | Predicts paid conversion 2–4 weeks out. If this moves, revenue will follow. |
| Leading | Time-to-first-value (median) | Onboarding health. Target: < 10 minutes end-to-end. |
| Leading | Qualified signup rate | % of signups matching ICP (10K–500K, active channel, EN/UA). Filters bad traffic before it hits the funnel. |
| Lagging | Free → Paid conversion | Core growth metric. |
| Lagging | Net new MRR | Ultimate business metric. Weekly view. |
| Lagging | Gross / net dollar retention (90d) | Tells us if the value is real after the novelty wears off. |
Weekly Dashboard
One page, five blocks -- shown to founders every Monday; PM owns the narrative next to the numbers
Signups, activations, paid, MRR, churn. WoW deltas.
Conversion rate between each of the 7 funnel steps, coloured by delta vs 4-week rolling average.
1/7/14/30-day retention for last 8 weekly cohorts.
Every live test -- hypothesis, metric, lift, ship/no-ship decision.
3 direct quotes from users this week (interview, support, churn survey). Counteracts decoration-bias in the numbers.
Hypotheses & Prioritization
Growth Hypotheses
Each hypothesis is framed as "if we do X, Y will change, because Z" -- activation and monetization levers mixed on purpose, as they feed each other
| # | Hypothesis | Why | Impact | Effort | Validation |
|---|---|---|---|---|---|
| H5 | Outbound to top 20 power users from free tier: "we'll set up Live / Fan Funding on a call -- free". | High-touch assisted activation. Reveals whether value is real when friction = 0. | High (learning) | Low | Manual, 2w |
| H1 | Narrow onboarding to ONE hero flow (e.g., Channel setup). Hide the rest until day 2. | Today: choice paralysis across Analytics/Live/Fan Funding. One clear path → faster TTV. | High | Low | A/B, 2 weeks |
| H7 | Qualify signups with a 3-question survey; route non-ICP users to a different flow. | Stops wasting activation on people who will never pay. Frees us to measure real conversion. | Med | Low | 1 week |
| H3 | Add a $49 Creator tier to break the $0→$149 cliff. | Price ladder lets undecided users commit. Lower-tier LTV is still profitable on a gross margin. | High | Med | Landing-page test → live test |
| H2 | Introduce a 14-day paid trial of Pro features triggered on activation events (not signup). | Product asks for commitment AFTER value is felt, not before. | High | Med | A/B, 4 weeks |
| H6 | Replace Explorer with a 14-day full trial, then downgrade to limited free. | Forces a payment decision at the moment of peak engagement. | High | Med | Live test, 3w |
| H4 | In-product "Next step" nudges from Marketing / Founder when a user hits an activation milestone. | Expert/Founder brand is already a traffic source -- extend it into product via 1:1 feel. | Med | Low | Cohort test, 2w |
Priority Order for Execution
Simplified ICE weighted by confidence -- with 0 paid conversions, bias toward tests that teach us something even if they don't win. Effort is engineering-capped.
90-Day Execution Plan
Each month builds on learnings from the previous one. Theme progression: Diagnose → Convert → Scale. Traffic is not scaled until the funnel produces repeatable paid conversion.
Week 1: Set up full funnel, pull cohort data, stand up weekly dashboard. Week 1–2: 10 user interviews (5 active, 5 churned) + session replays. Week 2–3: Concierge activation (H5) -- 20 signups, offer 30-min setup call. Week 2–4: Ship H1 (narrow onboarding) and H7 (qualify signups). Week 4: Diagnosis memo + ICP definition alignment with founders.
Week 5–6: Trial mechanic experiment (H2 or H6) -- trigger paid trial on activation event. Week 6–7: 2 copy/UX iterations on pricing page; A/B test tier order, headline, social proof. Week 7–8: Ship H3 ($49 Creator tier) as pricing-page test first. Weekly: 2 user interviews, 1 experiment shipped, 1 killed.
Week 9–10: Identify the single highest-converting tier + flow. Kill the rest. Week 9–11: Ship H4 (founder in-product nudges) -- retention lever. Week 11–12: Marketing scale-up brief with validated channels, creative angles, and segments. Week 12: Quarterly retro; build month 4–6 plan.
Success Criteria by Month
| Month | Success Criteria |
|---|---|
| Month 1 | Activation rate measured (even if low). ≥2 paying customers -- even if from concierge. Validated ICP definition + documented "why users don't pay". |
| Month 2 | Free→paid conversion ≥1%. ≥10 paying customers total. At least one converted without human touch -- proves the loop is repeatable. Trial→paid rate baseline measured. |
| Month 3 | ≥20 paying customers, free→paid ≥2%. At least one growth loop producing qualified signups for free. CAC per paying customer trending down; 30-day retention ≥80%. Monetization loop documented, not tribal. |
Sequencing logic
Monetization & Unit Economics
Minimum Retention for Viability
Starting from stated assumptions: price $129/mo, blended CAC $18, infra+support $6/mo per active user
| Metric | Value | Note |
|---|---|---|
| Blended CAC per signup | $18 | Given |
| Target free→paid conversion | 2% | Our target |
| Effective CAC per paying customer | $900 | $18 / 2% |
| Gross margin $/mo ($129 plan) | $123 | Price − infra |
| Payback month | 7.3 mo | 900 / 123 |
| Break-even churn | ≈13.7% | % from previous month average |
| Viable LTV/CAC >3 target | $2,700 | 3× SaaS benchmark |
| Required churn for viability | 4.5% | 123 / 2,700 |
LTV Scenario Analysis
| Scenario | Churn % | Lifetime (mo) | LTV ($) | Eff CAC ($) | LTV/CAC | Payback (mo) |
|---|---|---|---|---|---|---|
| Pessimistic | 18% | 5.5 | $677 | $1,200 | 0.56 | 9.8 |
| Break-even | 13.7% | 7.3 | $898 | $900 | 1.0 | 7.3 |
| Optimistic | 4.5% | 22 | $2,700 | $600 | 4.5 | 4.9 |
Honest answer
Alternative Packaging -- Would I Test It?
Yes. $149 plan is probably the biggest monetization leak after activation. Proposed four-tier model:
Creator tier breaks the $0→$149 cliff. Pro tier validates whether there's a product value for $79. First month discount at $79. Lowers entry barrier while preserving upside.
Weekly subscription A/B test. Lowers the entry barrier psychologically while keeping monthly equivalent roughly the same. Easier first commitment for a creator who is uncertain.
Process Thinking
As the Only PM: How I Organize Growth Work
Growth work fragments fast. One theme per month (Month 1: Diagnose; Month 2: Convert; Month 3: Optimize) -- say no to everything outside it.
Every experiment has an owner, a hypothesis, a single metric, a decision rule, and a ship date. Nothing runs longer than 2 weeks without a decision.
Every roadmap change must be backed by a data point OR a user interview. Founder intuition is a hypothesis, not a conclusion.
Balancing Discovery, Experiments, Delivery
Rough weekly time allocation for the PM -- deliberately discovery-heavy for the first 90 days. Rebalances toward delivery (~40%) once the funnel works.
| Work Type | My Time | Examples |
|---|---|---|
| Discovery (qual + quant) | 40% | User interviews, session replays, cohort digs, competitor teardown. |
| Experiments (design + analyze) | 35% | Writing test briefs, pricing experiments, A/B analysis, decision memos. |
| Delivery (ship + spec) | 20% | Specs for engineering, copy iterations, backlog grooming, QA. |
| Comms & alignment | 5% | Weekly dashboard narrative, founder update, marketing sync. |
Cadence
What I Would Propose on Day 1
Hypothesis + metric + decision rule written up front. No running tests indefinitely. If it doesn't have a decision rule, it doesn't run.
Founders set the goal; PM chooses the levers. Otherwise context-switching kills the team. Founder intuition enters the backlog as a hypothesis, not as an instruction.
Closing note