Taikan — Cost Analysis & Pricing Strategy

Last Updated: 2026-08-05 (sections 0–0.2 are current; sections 1–8 date from 2026-03-12 and are superseded wherever they conflict)

0. Current cost model (2026-08-05)

What changed since March

Actual spend today

Railway ~$7/month, everything else on free tiers. Two of those are load-bearing and should not be:

Vercel Hobby prohibits commercial use. Taikan is a commercial SaaS running on it. That is a terms violation rather than a capacity limit, and the failure mode is being switched off, not degrading. Pro is $20/mo.

Resend free is 3,000/month AND 100/day. One 400-member gym sending monthly receipts is 400 emails; if they land on one billing day the daily cap breaks in a single batch. These carry payment receipts and dunning. $20/mo buys 50K.

Spotter (AI) cost model

Sonnet 4.5 for the main loop, Haiku 4.5 for quick routing, with ephemeral prompt caching at 1h TTL already implemented — that halves the bill, since cached input reads at $0.30/M instead of $3/M.

Per user question: ~$0.07 cold, ~$0.03 warm; multi-step tool use means 2–3 model calls per question.

Budget $4–6 per paying gym per month. The ceilings in libs/shared/src/lib/constants/platform-tiers.ts (Lite $3, Pro $20, Elite $50) sit ~5× above moderate usage: high enough that nobody normal hits them, low enough that one runaway org cannot eat a tier’s revenue.

AI is roughly half of variable cost at scale, and the only line that scales linearly with customers.

Scenarios (₪1 ≈ $0.27)

Fixed business costs are the easily-forgotten, unavoidable ones: Bituach Leumi ~₪170, accountant ₪300–500, Google Workspace $7, domains $4, Claude Code $20. They dominate at 5 customers and are noise by 100.

Break-even is ~4 paying gyms. Fixed runs ~$273/mo once Resend and Vercel are paid; variable is ~$6/gym; a mixed Pro/Elite customer brings ~$81.

Not modelled: EAS/Expo build minutes. Free-tier builds are limited and store releases consume them; likely $19–99/mo depending on release cadence.

Spend priority

1. Resend $20 — before onboarding a gym of any size. Transactional email carrying payment receipts must not silently fail.
2. Vercel Pro $20 — compliance, not capacity.
3. GitHub Actions — already caused false CI failures once.
4. Railway headroom as load grows.
5. Sentry, PostHog, Clerk — genuinely distant.

Sentry $70 with Seer: not yet. A 10× multiple on current infrastructure spend for an AI debugging aid, while transactional email is one busy day from failing. Better Stack already covers the uptime gap Sentry free leaves. Revisit at ~20–25 paying gyms.

0.1 Payment provider costs (2026-08-05)

Two separate things, easily conflated.

Taikan’s own merchant account — a real fixed cost

Taikan bills gyms through Cardcom, so it holds its own terminal:

• ₪140/month platform terminal
• ~1% clearing on everything Taikan collects from gyms

At 25 gyms averaging ₪300 that is ₪7,500 collected, so ₪75/month of clearing on top of the ₪140. This line was absent from the March model entirely.

Per-gym terminals — pass-through, not margin

Cardcom’s platform deal, confirmed in writing 2026-08-05:

Realistic ≈ ₪85/gym/month, passed through rather than absorbed — at Pro (₪249) absorbing it would be 34% of that customer’s revenue.

Cardcom performs all KYC and regulatory onboarding. Taikan is never a payment facilitator: no מאגד status, no held funds, no chargeback liability. For a solo operator that removes a whole category of risk and is worth more than the price.

Open: clearing to gyms is “around 1%, volume-tiered” with no tier table supplied. Whether tiering is per-merchant or across Taikan’s aggregate book is the single most valuable unresolved term — ten gyms at ₪67k each is ₪675k/mo flowing through Taikan and should price well under 1% for everyone.

Also open: does Cardcom bill Taikan per gym terminal (requiring Taikan to re-bill), or bill the gym directly? That decides whether ₪85 × N appears in Taikan’s P&L at all.

Where the crossover sits

At ₪450 average membership, Taikan + Cardcom at 1% beats Arbox + Grow below ~150 members and loses above it: the Pro→Elite step (+₪250) collides with a clearing gap that has grown to ~₪170/month by then, while Arbox stays flat at ₪568 for unlimited members.

Consequence: at high volume, a gym on its own provider account via external payment links is cheaper than a Taikan-managed terminal. Erez at ~400 subs pays ₪2,058/mo on Taikan + his own Grow at 0.75%, versus ₪2,384 on Cardcom at 1%. See ADR-0017.

0.2 Pricing recommendation (2026-08-05)

Cost is not the constraint. At ₪249 against ~₪22 of variable cost per gym (Spotter ~$5 plus ~1% clearing on the collection), gross margin is ~91% and break-even is four customers. Pricing is a positioning decision, not a margin one. The cost model does expose three real defects.

Leave Pro at ₪249

A strong acquisition price, and break-even sits at four gyms. Customers are worth more than margin right now; raising the entry price to fix problems that live at the top of the range would be backwards.

Defect 1 — one tier spans 151 members to infinity

A 160-member studio and a 600-member multi-location box both pay ₪499. The small one hits a 2× cliff at 151; the large one pays ₪0.83/member/month. The ICP (Israeli boutique gyms, 150–400 members) sits entirely inside that gap.

Studio turns the cliff into a step and covers where most real customers land (Erez is ~400). Elite at ₪699 for 600 members is ₪1.17/member, still cheap against Arbox or Wodify, while capturing value currently given away.

Defect 2 — unlimited locations is a direct margin leak

Elite includes unlimited locations flat, but every branch needs its own Cardcom terminal at ₪85/mo. A three-location gym costs ₪255/mo in terminals against ₪499 of revenue. Charge one location included, +₪149 per additional: it covers the terminal and bills honestly for something that genuinely costs money.

Defect 3 — Lite is ~40× more expensive to serve than §3 assumes

Section 3 puts Lite cost-to-serve at $0.08/org. With Spotter a Lite org can burn its full $3/month AI budget plus infrastructure share. Fifty free orgs is $150–250/mo of pure loss, most of the fixed cost base. Cut Lite’s monthly AI budget from $3 to ~$1 (≈10 Spotter questions): still a real taste, capped downside.

The lever that outweighs all of the above

Moving Cardcom from 1% to 0.85% saves a ₪180k/month gym ₪270/month, more than the entire gap between Studio and Elite. For larger customers, winning the clearing-rate negotiation matters more than any repricing, and costs nothing.

Calibration caveat: ₪429 and ₪699 derive from the cost curve and from Arbox’s ₪568, not from willingness-to-pay data, which does not exist with one real customer. Treat them as the shape of the answer and let the next three or four deals set the level.

1. Platform Tiers — Revised (aligned with code)

Implementation: See GitHub Issue #58  — “Align platform tier features, pricing, and gating” (blocks #53).

Key Changes from Previous Distribution

Full Feature Set by Tier (from roadmap)

2. Vendor Costs — Complete Breakdown

Infrastructure Vendors

Application Vendors

Israeli Business Vendors

Development & Operations

3. Cost-to-Serve per Tier

Usage Assumptions

Per-Customer Infrastructure Cost

Note: Most vendors stay within free tiers at launch. Costs shown are marginal per-customer allocation when splitting shared resources. At <50 customers, most services remain free. Railway is the only guaranteed cost.

Fixed Monthly Costs (Regardless of Customer Count)

Cost-to-Serve by Feature Set

Each feature has a different infrastructure cost profile. Understanding which features are “free to serve” vs “expensive to serve” informs tier gating decisions.

Tier 1 — Zero/Negligible Cost Features (Lite)

These features cost almost nothing to serve per customer. They run on shared DB + compute. Minisite moved to Pro to reduce Lite cost-to-serve and create a clear upgrade incentive.

Tier 2 — Low-to-Medium Cost Features (Pro adds)

These features add moderate per-customer costs, mostly from email and DB volume.

Note: WhatsApp AI bot and automations moved to Elite tier (highest per-org cost at $2-5/mo). Leaderboards and analytics dashboard moved down to Pro — low infrastructure cost ($0.05 and $0.50/org respectively) and table-stakes for gym owners. See docs/product/whatsapp-ai-bot-research.md for the full WhatsApp cost analysis.

Tier 3 — Higher Cost Features (Elite adds)

Cost-to-Serve Summary by Tier

SaaS gross margins of 90%+ are exceptional. Even with conservative estimates, Taikan’s infrastructure costs are a tiny fraction of revenue. The real cost is your time.

Feature Cost Waterfall

Green = negligible cost, Blue = low cost, Orange = medium cost, Red = primary cost driver

Key insight: WhatsApp AI bot ($2-2.50/org/mo) and WhatsApp automations ($2-5/org/mo) are the largest per-customer costs, now gated to Elite tier. Email ($1.60/org/mo) is the primary cost driver at Pro tier. Everything else — DB, compute, storage, Redis — is nearly free thanks to generous free tiers and efficient serverless pricing.

4. Three-Scenario Cost vs Revenue Analysis

Scenario 1: LAUNCH (Month 0-3)

2 customers: CrossFit gym (338 members, Elite) + Online trainer (100 clients, Pro)

Scenario 2: GROW (Month 6-12)

15 customers: 5 Lite + 7 Pro + 3 Elite

Scenario 3: SCALE (Month 12-18)

80 customers: 20 Lite + 40 Pro + 20 Elite

With Transaction Fee (1% on member payments)

Transaction fees nearly double revenue at scale. Consider adding in Phase 2 when payment processing is mature.

Revenue vs Costs Growth Trajectory

Bars = revenue, Line = total costs. Revenue scales linearly with customers; costs scale sub-linearly thanks to shared infrastructure and free tiers.

Break-Even Analysis

Without accountant (not required for Osek Patur) and with ₪0 Bituach Leumi (if employed above cap — see Section 8), launch costs are ~₪471/mo. Infrastructure (Railway) is the dominant cost at ~55%. Two paying customers (₪748/mo) nearly cover costs at alpha pricing.

5. Competitor Pricing Comparison

Full Management Platforms

Workout-Only Platforms

Competitive Price Positioning

The “Stack Tax” — What Gyms Actually Pay

Taikan vs Competition

6. Pricing Rationale

Why ₪249/₪499

Your reference CrossFit gym pays Arbox Professional (~$239-299/mo). If they also use SugarWOD ($179-219 for 338 members), their combined spend is $420-520/mo.

Lite Limits (free tier guard rails)

Competitive Positioning

7. Transaction Fee Analysis

Model Options

Revenue Projection with 1% Transaction Fee

Assumes average gym processes ₪150-200/member/month in payments:

At scale, transaction fees nearly equal subscription revenue. Recommendation: Launch with subscription-only. Add transaction fees in Phase 2 when B2B2C payment processing is mature and value is proven. Frame as “payment processing service fee.”

8. Business Registration & Tax Impact

Recommended Path: Start as Osek Patur

When You Hit the Threshold

You’ll likely need to convert to Osek Murshe at ~10-12 paying customers (Grow phase). This is a good problem to have.

VAT Impact on Pricing

Once Osek Murshe, prices are listed + מע”מ (standard Israeli B2B practice):

• Pro: ₪249 + 18% VAT = ₪294 to customer (you keep ₪249)
• Elite: ₪499 + 18% VAT = ₪589 to customer (you keep ₪499)
• You can deduct VAT on all business expenses (infrastructure, tools, accountant)

Pre-Launch Costs

Note: An accountant is not legally required for Osek Patur (source ). Self-manage with GreenInvoice for invoicing + file one annual tax declaration. Hire a CPA when approaching the ₪122,833/yr Osek Murshe threshold.

Bituach Leumi for employed founders: If already employed with salary above the ₪51,910/mo cap (2026), Bituach Leumi on side business income is ₪0 — the employer already covers contributions up to the maximum insurable income.

9. GTM Positioning Summary

Value Proposition by Customer Type

For your CrossFit gym (338 members, Arbox Professional + SugarWOD):

• Currently paying: ~$420-520/mo for two disconnected platforms
• Taikan Elite: ₪499/mo (~$136) — saves 74%, single platform
• Added value: Israeli payment gateways, GreenInvoice, minisites, community feed
• Migration path: Arbox data import tool (Issue #38)

For the online trainer (100 clients, SugarWOD):

• Currently paying: ~$69-99/mo for workout tracking only (no payments, no scheduling)
• Taikan Pro: ₪249/mo (~$68) — comparable price but includes full platform
• Added value: Member management, billing, scheduling — replaces the need for separate tools
• Migration path: SugarWOD doesn’t expose data easily, but Taikan can import via CSV

Competitive Position

Price (monthly, USD)
$500 ─ Mindbody Ultimate ──────────────────────────────────
      │ Arbox Pro + SugarWOD ─────────────────────────────
$400 ─ Wodify + SugarWOD ────────────────────────────────
      │ Arbox Standard + SugarWOD ────────────────────────
$300 ─ Arbox Professional ────────────────────────────────
      │
$200 ─ Wodify Essentials ────────── No workouts ─────────
      │ WodBoard ─────────────────── No Israeli payments ─
      │ Taikan Elite (₪499) ★ ──── Full platform ──────
$100 ─ Arbox Standard ───────────── No workouts ─────────
      │ SugarWOD ─────────────────── Workouts only ───────
      │ Taikan Pro (₪249) ★ ────── Full platform ──────
  $0 ─ Taikan Lite ★ ─────────────── Limited free tier ───

Taikan is the only platform that offers full management + workouts + community + Israeli compliance at ANY price point. Competitors either lack features (Arbox: no workouts) or lack Israeli focus (SugarWOD, Wodify, Mindbody).

Feature Coverage by Platform

Green = included, Red = missing. Taikan is the only platform with full coverage across all categories.

10. Costs You Might Not Have Considered

11. Solo Entrepreneur Notes

1. Your CrossFit gym is your unfair advantage. A real 338-member reference customer is worth more than subscription revenue. Consider a significant launch discount in exchange for being a public case study + feedback partner. Their word-of-mouth in Israeli CrossFit is your GTM.

2. SugarWOD replacement is the wedge. The online trainer pays $69-119/mo for workout tracking alone. If Taikan matches SugarWOD’s workout UX and adds scheduling/payments, you’re replacing a limited tool with a full platform.

3. Arbox data migration is the moat. If switching from Arbox is painless, you remove the biggest barrier. If painful, even a better/cheaper product won’t win. Prioritize Issue #38.

4. Don’t absorb VAT. Israeli B2B customers expect ”+ מע”מ” pricing. All competitors do the same. Absorbing it cuts your margin 18% for zero competitive benefit.

5. Infrastructure costs are exceptional. At $70/mo for launch, your COGS is basically zero. This is the power of serverless. Protect this by staying on free tiers as long as possible.

6. Offer annual pricing. 2 months free on annual plans (17% discount). Pro annual: ₪2,490/yr (vs ₪2,988 monthly). Improves cash flow and reduces churn.

7. Support load is real. A 338-member gym will generate questions. Build self-serve help (knowledge base via minisite), in-app tooltips, and set expectations on response times from day 1.

12. Cost Efficiency Strategies

Starter Tier — Minimize Cost-to-Serve

The Starter tier is a loss leader. Every shekel saved per free org extends runway.

Result: Starter cost-to-serve dropped from ~$0.58 to ~$0.08/org/mo — a 7x reduction. At 20 free orgs, that’s $10/mo vs $11.60/mo (negligible either way, but the tighter limits drive faster upgrades).

Feature Gating as Cost Control

The expanded PLATFORM_FEATURES enum (6→20 features) enables fine-grained cost control. Expensive features are gated to paying tiers:

Infrastructure Scaling Strategy

Implementation Reference

The feature gating alignment is tracked in GitHub Issue #58  — this must be completed before billing checkout (#53) can go live. See libs/shared/src/lib/constants/platform-tiers.ts for the expanded PLATFORM_FEATURES enum and tier mappings.

Planned Feature Additions to PLATFORM_FEATURES

13. Competitor & Tool Deep Dives

Research notes from evaluating tools and competitors mentioned by prospects.

Fittr (fittr.com)

B2C fitness coaching marketplace (India, 5M+ users). NOT a competitor.

Different market entirely — connects individual clients with personal trainers/dietitians. No scheduling, billing, or studio management tools. No B2B offering. Irrelevant to Taikan’s studio management positioning.

UpGrade360 (upgrade-360.com)

Arbox marketplace add-on for WhatsApp engagement automation. NOT a standalone platform.

Pulls member data from Arbox CRM and sends automated WhatsApp messages: birthdays, inactive member alerts, trial class reminders, membership expiry warnings. Requires an existing Arbox subscription.

Key takeaway: Validates that WhatsApp automation is a high-value feature for Israeli gym owners — they’re willing to pay extra (on top of Arbox) for it. Taikan plans to build this natively (see FIT-54), eliminating the need for a separate add-on. Taikan goes further with a conversational AI bot (FIT-17) that handles natural language booking via WhatsApp — a capability no competitor offers.

Paddle (paddle.com)

Merchant of Record (MoR) payment platform. Not suitable for Taikan’s current model:

Verdict: Revisit for international expansion when Taikan needs multi-country tax compliance. Not suitable for Israeli launch.

14. Global Expansion Readiness

Assessment of Taikan’s readiness for markets beyond Israel.

Blockers

• Platform tier pricing defined in ILS only
• No Stripe/Stripe Connect integration (#39 is Phase 4)
• ILS defaults hardcoded in payment service layer
• GreenInvoice is Israel-specific invoicing

Path to Global

1. Complete Israeli market launch and prove product-market fit
2. Refactor currency handling (remove ILS hardcoding, add currency per org)
3. Add Stripe Connect (#39) for international payment processing
4. Implement regional pricing (per-country tier pricing)
5. Consider Paddle for automated tax compliance at scale