Poki Yoki — The Next 90 Days, Down Two Paths

Path B Classic DTC sales   vs   Path A Game-first free model. Week-by-week projected numbers, a reality-checked verdict, and the one decision to make on Day 1.

v1 · 2026-06-13 · budget anchor: $15,000 media / 90 days (~$5K/mo) · figures reality-checked by an adversarial CFO pass · for marketing/ad review

How to read this. §1 is the verdict + the 5 numbers for the wall. §2 shows the two 90-day curves side by side (the shapes are the whole story). §3 is the reality-checked scorecard. §4 is the unit-economics fix that changes everything. §5 is cash & survival. §6 is the budget-sensitivity (it matters a lot). §7 is the recommended plan. §8 is the three things you must measure. §9 is the precedent that should give everyone pause. The team's first projections were optimistic; every number here is the corrected, conservative version.

Contents

1. The verdict & 5 numbers for the wall
2. The two 90-day curves (week by week)
3. Reality-checked 90-day scorecard
4. The unit-economics fix: lead with Family
5. Cash flow & survival
6. Budget sensitivity ($7.5K / $15K / $30K)
7. The recommended 90-day plan
8. The 3 things you must measure (kill-switches)
9. The precedent that should give everyone pause
10. Sources

1 The verdict & 5 numbers for the wall

Path B makes more cash. Path A builds a more valuable asset. Neither is profitable in 90 days.

At a $15K media budget, neither path produces positive contribution-after-CAC in 90 days — that's the honest headline. Both are investments in proving channel economics and building an audience. Path B (classic DTC) wins on cash in every scenario (it generates revenue from Week 1); Path A (game-first) builds a bigger long-term list asset but burns far more cash to do it and converts almost nothing inside 90 days. The right move is not "pick one" — it's Path B as the revenue engine, restructured around the Family SKU, with a small, stage-gated Path A test.

Five numbers to put on the wall

#The numberWhy it matters
1Path B ≈ 270 orders vs Path A ≈ 44 orders in 90 days (at $15K)A gap. Path A's rosier "234-order" draft contained a 7.5× math error between list size and per-week funnel; 44 is the reconciled figure.
2Path B ≈ $10,600 cash vs Path A ≈ $5,500 cash at Day 90The $5,100 gap = 1–2 months of extra runway for a company starting at ~$19.7K cash with $3–5K/mo burn.
3Starter ($45) is first-order unprofitable above ~$22 CAC; Family ($85) is profitable below ~$43 CACThe single most important unit-economic fact. Every Day-1 decision should be judged through it.
4Corrected base CAC ≈ $55 vs breakeven $22 (Starter) / $43 (Family)You need a ~30% CAC cut or a ~30% AOV lift. The AOV lift is available on Day 1: lead with Family.
5Neither path is profitable at Day 90The real question isn't "which makes money in 90 days" (neither) — it's "which best positions Month 4–12 given the cash constraint." Answer: Path B + a small Path A slice.

2 The two 90-day curves (week by week)

The endpoints matter less than the shapes. Path B earns from Week 1 and ramps as the pixel learns and retargeting kicks in. Path A earns almost nothing for 4–6 weeks (it's building a list), then accelerates — but never catches up inside the window.

Path B Classic DTC — cumulative, corrected base ($15K)

By end of…Spend (cum)Orders (cum)Revenue (cum)Contribution CM2Cash on handPhase
Week 2$2,100~28$1,290$620$18,185Learning phase, CPM penalty
Week 4$4,200~70$3,220$1,540$17,005Exits learning ~now
Week 6$6,300~120$5,520$2,640$16,005Creative iteration
Week 8$8,400~170$7,820$3,740$15,005Retargeting warms
Week 10$10,500~220$10,120$4,840$14,005Family-SKU weeks turn CM3+
Week 12 (Day 90)$15,000~270$12,420$5,940~$10,6000.83× ROAS · $55 CAC

Path A Game-first — cumulative, corrected base ($15K)

By end of…Spend (cum)Cup orders (cum)Revenue (cum)Verified moms on listCash on handPhase
Week 2$1,4000–1~$50~140$18,300Infra + first plays; $0 sales
Week 4$3,800~3~$150~280$15,950First trickle (welcome flow)
Week 6$6,600~9~$450~520$13,500List building, retargeting on
Week 8$9,400~18~$900~780$11,150"Your Poki needs a real cup"
Week 10$12,000~30~$1,500~1,040$8,400Hard push + promo code
Week 12 (Day 90)$15,000~44$2,200~1,270~$5,5000.15× ROAS · $341 CAC

The crossover: Path A never catches Path B inside 90 days

Path A's contribution accelerates in Weeks 9–12 as the list matures, but it can't close the deficit built in Weeks 1–8 when Path B was producing 2–6× more buyers per dollar. The two only cross beyond the window — and only if Path A's email list is genuinely large and converting (the optimistic case). With the corrected ~1,270-mom list, Path A's 12-month follow-on revenue is ~$8K/yr, not the $180K the first draft implied.

3 Reality-checked 90-day scorecard

Low / base / high bands at $15K, after the adversarial pass (CVR ceilings lowered, repeat rate halved to ~6%, AOV adjusted for launch skew, Path A list-size reconciled).

Path B Classic DTC

MetricWorseBaseBetter
Effective AOV (after discounts)$42$46$52
Peak CVR (Week 12)1.4%1.8%2.5%
Total orders (incl. ~6% repeat)175270413
Gross revenue$7,350$12,420$21,476
Blended ROAS0.49×0.83×1.43×
Blended CAC$86$55$36
CM3 (contribution − media)−$11,500−$9,060−$5,088
Cash at Day 90~$8,200~$10,600~$14,600
Email list built~350~540~825

Path A Game-first

MetricWorseBaseBetter
Viral coefficient (k)0.100.300.45
Kid-contamination multiplier2.5×2.0×1.7×
Verified moms on list (Day 90)6001,2702,100
90-day list→buyer conversion1.5%3.5%6.0%
Cup orders (cohort model)944126
Gross revenue (≈$50 AOV)$450$2,200$6,300
Blended ROAS0.03×0.15×0.42×
Blended CAC$1,667$341$119
CM3 (contribution − media)−$14,820−$14,120−$12,480
Cash at Day 90~$4,800~$5,500~$7,200

4 The unit-economics fix that changes everything: lead with Family

Starter $45 → CM2 ≈ $22 → unprofitable above ~$22 CAC Family $85 → CM2 ≈ $43 → profitable below ~$43 CAC ← lead with this

The path to profitability is not better CVR, lower CPM, or faster learning — those take weeks. It's higher AOV through Family-SKU emphasis, which is in your control on Day 1. If Family mix rises from ~30% to ~45–50% of orders, blended CM2 climbs from ~$22.57 to ~$31–33, CM3 per order improves by ~$10, and the path to first-order profitability shortens from Week 9–10 to Week 5–6.

⚠ Operational tension to resolve first

The "lead with Family" play is the single highest-leverage lever — but Family inventory is reportedly down to ~5–6 units (May 2026 data). Leading with Family requires a Family restock first, or the strategy stalls on availability. Confirm Family stock before building the campaign around it. (Also worth noting: actual paid ROAS to date is ~0.2× on a small sample — which is why the corrected base uses 0.83×, not the 2.17× category median.)

5 Cash flow & survival — the real constraint

~$19.7K
starting cash (Feb 2026) · ~4–6 months runway at $3–5K/mo burn
−$1.5K
Path B max contribution trough (Week 9), then recovers
−$15–21K
Path A trough at Day 90 — never recovers in-window

For a founder-funded company ($25K deferred salary owed, a second production run to finance), cash timing is decisive. Path B generates revenue from Day 1 and consumes ~14% of the 3,443-unit inventory (489 units in the higher-spend cash model) — healthy sell-through that funds Amazon/retail next. Path A converts almost no inventory (≈38 units / 1.1%) while spending the same media plus game/COPPA overhead — it depletes cash with little to show inside 90 days. The ability to earn in Week 1 vs. Week 6 is the difference between surviving to Day 180 and not.

6 Budget sensitivity — it changes the answer

Budget / 90 daysPath BPath ARead
$7,500110–140 orders; stuck in perpetual learning phase (under the ~$1,050/wk/ad-set minimum to exit)Don't run it — never builds list mass; ~20 orders, ~$1KBelow min viable. Redirect to influencer seeding + Amazon.
$15,000270 orders, $12.4K, ~$10.6K cash44 orders, $2.2K, ~$5.5K cashA 60/40 split here is the worst of both — neither gets enough.
$30,000550–700 orders, $25–32K rev, ~2.0× ROAS; weekly CM3-positive ~Week 7–8~85–100 orders, ~$4.4–5.2K; list worth ~$18–22K/yrThe threshold where Path B is genuinely viable and a hybrid makes real sense.

Implication: at $15K, don't dilute — concentrate on the proven path and run Path A as a small bounded test. The hybrid logic gets stronger as budget rises toward $30K.

7 The recommended 90-day plan

★ Run Path B restructured around Family; run Path A as a $2,500 stage-gated test

Follow the proven 70/20/10 split: ~70% to the proven channel, ~20% to the emerging game channel, ~10% to creative/influencer tests. Path B is the revenue engine and control arm (continuous). Path A is a bounded experiment that must prove verified-mom→buyer conversion before earning more budget — it cannot be the majority allocation with ~$19.7K cash and 3,443 units to protect.

$15,000 / 90-day allocation

Bucket$%Purpose
B Meta paid social (Mawn)$9,00060%Revenue engine · 2 ad sets max · Family as hero
B Creative production (UGC/static/video)$1,50010%10–15 assets/mo to feed the algorithm
B Influencer seeding (mom micro-creators)$1,50010%30–50 gifted units; amplify winners as dark posts
A Game traffic (TikTok/Meta) + promo$2,25015%Build ~400–500 verified moms + a ~1,500 retargeting pool
Buffer / holdout-test cost$7505%CPA rescue, clean holdout, pivots

Stage-gates: run to the thresholds, not the calendar

8 The 3 things you must measure (set kill-switches before Day 1)

9 The precedent that should give everyone pause

Every proven "game sells product" case is PAY-FIRST, play-second — not play-free, pay-later

Webkinz (the canonical case, ~$100M/yr at peak) sold the plush first; the game unlocked after purchase and drove repeat plush buys. When Ganz tried the inverse in 2009 (memberships without a toy purchase), growth stopped. McDonald's Monopoly (>$2B incremental/decade) is the same shape: buy → receive game pieces → return to buy more — it accelerates existing buyers, it doesn't manufacture cold ones. KFC's "Shrimp Attack" (+106% sales) worked because the game dispensed an expiring discount redeemable at point of sale. The historical record is clear: branded games reliably drive repeat and deepening purchases; there is no proven case of a free game generating cold top-of-funnel buyers for a physical product at a profit. Poki Yoki's Path A would be attempting something the precedent does not support — which is exactly why it must be a measured test, not a launch strategy.

This doesn't kill the game funnel — it repositions it. The Arcade's proven role is retention, repeat purchase, brand love, and a warm retargeting pool (the Webkinz/Monopoly job), not cold acquisition. That's a valuable asset worth keeping and feeding — just not the channel to bet the launch budget on.

10 Sources

Benchmarks: Triple Whale 2025 (baby Meta CPA $30.04, CPM $10.92, CTR 1.91%, CVR 1.85%, ROAS 2.17×) · MHI Growth Engine 2026 (DTC CVR, Meta-vs-TikTok, learning phase) · Varos 2025 (baby Google CPC/CPA) · Lebesgue / AdStellar 2026 (Meta learning-phase 50-event threshold, post-learning CAC) · Trendtrack 2025 (creative-iteration CAC reduction) · Klaviyo 2025 (welcome/abandoned-cart, sweepstakes-list conversion) · GoBolt 2025 / Top Growth Marketing (shipping + 3PL costs) · Rijoy 2026 (baby/toddler DTC bounce/loyalty) · Sensor Tower "Parents Wrapped" 2025 (parent-impression share). Precedent: Webkinz (Wikipedia / VirtualPetList / David Kushner) · McDonald's Monopoly (Restaurant Business Online / eventXgames / Chief Marketer) · KFC "Shrimp Attack" (Bazaarvoice). Repo: company/financials.md · marketing/key-numbers.md · brand/positioning.md · pokiyoki-dashboards-overstate-dtc (actual paid ROAS ~0.2×) · GTM cockpit spec (Mawn $3K/mo from June 15). Method: 7-analyst research swarm (2023–2026 sources) + an adversarial CFO reality-check that corrected the team's first projections (Path A list-size reconciled, CVR/repeat/AOV adjusted down). Figures are decision-grade estimates; the load-bearing unknown — the 90-day verified-mom→buyer rate — is unmeasured until the §7/§8 sprint runs.

Poki Yoki · The Next 90 Days, Two Paths · for marketing/ad-professional review · the honest read: neither path is profitable in 90 days — Path B preserves cash and builds social proof; Path A is a measured bet on a long-term asset. Run B around Family; test A to a kill-switch.