Pricing strategy Draft

How Kashi prices, and why

Pricing is the single-biggest unforced-error surface in B2B SaaS. Undercharging burns capital. Overcharging kills conversion. Picking the wrong pricing metric makes both happen at the same time. This document explains the four frameworks B2B SaaS companies actually use to set prices, applies them to Kashi, and lays out a 90-day plan to validate the numbers before locking them.

Status. This is a pricing hypothesis and operating doctrine, not a pricing truth. The current structure and provisional numbers are commercially coherent enough to test, but not yet commercially proven. Willingness-to-pay, discount pressure, and procurement behavior from real deals are the evidence that would validate any of this. None of that exists yet.

Drafted 2026-04-21 · Companion to /business.html §4 + /jp-launch-runbook.html

4 frameworks
value-based, cost-plus, competitive, hybrid — most SaaS uses hybrid
3 tiers
Good / Better / Best + Enterprise — the industry standard shape
~80%
share of SaaS revenue the middle tier typically captures
10 conversations
customer WTP interviews before locking any provisional price

00Why pricing is strategy, not math

There are three pricing failure modes that kill B2B SaaS companies before they reach scale. In descending order of how often they actually happen:

  1. Wrong pricing metric. The company charges per-seat but the value scales with something else (meetings, data volume, governance depth). Every customer feels the price is misaligned with what they get. Churn is high for reasons the team can't name.
  2. Undercharging. The founders are self-conscious about asking for money, they anchor to their own salary-adjacent intuition, they set the price at 1/5 of what the buyer would gladly pay. Every deal feels fine until the runway runs out.
  3. Freemium-without-conversion-path. A generous free tier draws users; none of them ever upgrade because the free tier covers their real need. Free tier becomes an expensive charity program.

Each of these mistakes is avoidable with a deliberate process. "Pick a number that feels right" is the process almost every first-time founder uses, and it is the process that produces these three failures. This doc is the alternative.

The single sentence that defines good B2B pricing. The price should be a fraction of the economic value the customer captures from using the product, measured along a metric that scales with that value.

01The four canonical B2B SaaS pricing frameworks

Every B2B SaaS pricing decision uses one of these four, or a hybrid of them. Understanding which one applies to Kashi is the prerequisite to picking actual numbers.

Expand framework deep-dive — four pricing frameworks explained, with Kashi application ~1,400 words
Framework 1 Value-based pricing Use for Kashi

Price as a fraction of the economic value the customer captures from the product. The gold standard for B2B software because marginal costs are near-zero and the only defensible price ceiling is customer willingness-to-pay, not cost.

Typical capture rate (working heuristic, not a law): B2B SaaS companies commonly charge somewhere in a 10-25% band of the value they deliver. Higher tends to trip buyer reluctance; lower tends to leave money on the table. Treat 10-25% / 18-37% / "resist thresholds" as working commercial heuristics that help stress-test a number, not as stable universal rules the market will enforce.

For Kashi: value = averted cases of quiet resignation and mental-health leave, plus productivity recovery from earlier intervention. The rebuilt-ROI memo puts a 500-person company at ¥131M/year of productivity drag from team-dynamics issues. A 10-20% reduction via earlier intervention is ¥13-26M/year recovered. Charging 10-20% of that value gives ¥1.3-5.2M/year per 500-person company. Current Provisional Professional-tier pricing at ¥800 per covered employee per month lands at ¥4.8M/year for a 500-person company — inside a value-capture heuristic band (10-25% is a working rule of thumb, not a universal law). The math is directionally plausible, not yet willingness-to-pay validated. Real discount pressure, procurement objections, and signed deals are the only evidence that would actually confirm it.

Framework 2 Cost-plus pricing Avoid for SaaS

Price = COGS + target gross margin. How restaurants, retailers, and most physical-goods companies price. Almost always wrong for software.

Why it fails for SaaS: the marginal cost of serving one more user is near-zero. A cost-plus calculation produces a floor, not a ceiling, and following it strictly leads to race-to-the-bottom pricing. Cost-plus still matters as a check (your gross margin must be healthy — for SaaS, 70-85% is the target band), but not as the primary method.

For Kashi: treat cost-plus as a gross-margin sanity check. Starter tier gross margin is N/A (free). Pro tier gross margin targets 85% (compute + email support). Enterprise tier gross margin targets 65% because the CSM + consultation hours are meaningful cost. These are healthy bands. If any tier drops under 60% gross margin, the price is too low or the tier is overscoped.

Framework 3 Competitive anchoring Sanity check only

Price as a function of the closest competitor's price ±20-50%. Useful when the market has established reference prices the buyer already knows. Dangerous as a primary method because it implicitly concedes that your product is a substitute for the competitor, which undercuts differentiation.

For Kashi: the closest category (engagement survey: Wevox, Geppo, Culture Amp) prices at ¥300-1,200/user/mo. The deeper category (harassment-scanning: Archaic, FRONTEO) prices higher but doesn't publish. Kashi's provisional Pro tier (¥800 per covered employee/mo) sits at the top of the engagement-survey band and below the harassment-scanning band. The positioning is defensible as "governance infrastructure, priced above surveys because it does more than surveys, priced below content-scanners because we refuse to read content." Use this framing in sales, but do not let it set the number. The number comes from value-based math.

Framework 4 Hybrid (industry standard) Kashi's actual model

Most B2B SaaS companies combine all three. Different parts of the product get priced differently.

  • Bottom of funnel (freemium / Starter): priced near zero to drive adoption. Cost-plus is relevant because you don't want the free tier to bankrupt you.
  • Middle (Professional / mid-market): competitive-anchored, because mid-market buyers do compare shop. Priced as "better than the obvious alternative" with a clear differentiation story.
  • Top (Enterprise): value-based, custom-negotiated, often not published. The value of Enterprise features (governance-compliance, SLAs, SOC 2, residency) is asymmetric across buyers, so pricing has to flex.

For Kashi: this is already the model. Starter free (bottom), Professional per covered employee (middle, anchored against surveys + content-scanners), Enterprise contact-led with base + per-covered-employee as the internal hypothesis (top, value-based, negotiated). The decision tree that matters is not whether to hybrid, but what goes in each tier.

02Choosing a pricing metric (what you charge FOR)

The pricing metric matters more than the price. A good metric makes every customer feel the cost matches their use. A bad metric produces arguments about every invoice.

Kashi's unit is "per covered employee in scope," not "per seat." Value does not scale with how many people log in to a dashboard — it scales with workforce population whose meetings flow through the detectors. "Seat" implies named dashboard users (an admin, an HR business partner, a manager); "covered employee" is the correct unit.

What "covered employee in scope" means, precisely

A single, unambiguous definition prevents downstream invoice arguments. An employee is in scope if any of their meetings during the billing period were ingested by Kashi. The table below defines who counts:

Population Counts toward the bill? Why
Employees whose meetings were ingested Yes This is the thing Kashi actually analyzes; value scales with it.
Contractors / agency staff whose meetings were ingested Yes Same ingestion footprint, same marginal cost, same governance concern.
Executives, admins, HRBPs who only view dashboards No Named dashboard users are included in the tier — they do not add to the per-covered-employee count.
Dormant accounts / employees with zero ingested meetings No No ingestion, no detection, no bill.
Employees who opted out of participation No No meetings enter the detectors for them; they're excluded from the count.

The count is computed monthly from the ingestion log: distinct employee identifiers whose transcripts entered at least one detector run. It is visible to the customer; there is no black-box billing surprise.

Metric Best when… Breaks when… Fit for Kashi
Per covered employee in scope Value scales with the workforce population whose meetings flow through the detectors. Buyer plans governance budget annually against headcount. If the billed unit is misdefined (named dashboard users vs actual ingested population), invoice arguments follow. See the who-counts definition above. Pro tier — the correct unit for Kashi. Not "per seat" (named dashboard users); per employee whose meetings are ingested.
Per-meeting analyzed Usage is highly variable per customer. Cost tracks closely with usage. Governance and compliance-adjacent buyers plan against annual budgets and expect low invoice volatility. Consumption billing feels category-wrong — it produces monthly surprises that procurement dislikes even when the totals are fine. Poor fit — rejected for Kashi. Annual budget predictability matters more than usage-accuracy for this buyer persona.
Per-team (bundle) Natural usage unit is the team, not the individual. Large companies with many small teams pay too much; flat-team pricing caps upside. Weak fit — team boundaries are fluid; seat-based pricing captures the same value more flexibly.
Flat-rate per company You want simplicity. SMBs love it. Leaves enterprise money on the table; a 50-person co pays the same as a 5,000-person co. Starter only — fine for the ≤20-employee free tier. Wrong for revenue.
Base + per-covered-employee (hybrid) Enterprise value has fixed + variable components. Deployment / security review / SSO+SCIM / governance setup / consultation / dedicated success motion drive fixed; per-covered-employee drives variable. Small buyers find the base intimidating; you need a clear story for why the base exists. Enterprise tier — internal working hypothesis: ¥10M base + ¥800 per covered employee/mo. The base is a fixed-value service fee, not a software fee. Not published; negotiated per engagement. Defensible only to the extent the fixed-service layer is real and well-scoped.
Kashi's metric recommendation (directional). Per covered employee for Professional, base + per-covered-employee for Enterprise, flat (free) for Starter. This is directionally coherent for governance-category SaaS; the numbers are still subject to willingness-to-pay validation and first-deal pressure.

03Tier strategy — Good / Better / Best + Enterprise

The classic B2B SaaS tier pattern exists for behavioral reasons, not mathematical ones.

The trick is feature-gating — deciding which capabilities belong to which tier. Kashi's current feature gates (proposed):

Entry

Starter

Free
≤20 employees
  • 3 core structural detectors
  • Self-visibility only
  • No executive view
  • 90-day hard-delete retention
  • Community support
Workhorse (landing) · Published

Professional

¥800 per covered employee/mo
Provisional · ≤500 covered employees
  • All 7 detectors (incl. keigo asymmetry)
  • Manager Mirror + Executive Brief
  • Feedforward commitments
  • 24-month retention
  • JP data residency (Tokyo primary)
  • Downloadable labor-consultation packet
  • Email support (24h response)
Deployment depth · Contact-led

Enterprise

Contact us
scoped to deployment complexity
  • Everything in Professional
  • Evidence vault (E2E encrypted)
  • SSO / SAML / SCIM
  • Dedicated CSM
  • Hands-on labor consultation (就業規則 review + rep-process sit-in)
  • SOC 2 Type II + ISO 27001 report access
  • 24/7 support + SLA
  • NAQ-R validation study access

Internal working hypothesis: ¥10M base + ¥800 per covered employee/mo. The base is a fixed-value service fee covering deployment effort, security/compliance review, SSO+SCIM setup, governance process setup, consultation, and dedicated success motion — not software. Not published until the fixed-service package is more stable and more defensible in procurement.

What happened to "Enterprise+" / ROI-share pricing? Earlier drafts floated a fourth tier priced as a share of measured productivity recovery. That idea is demoted from active doctrine. It is premature, introduces attribution and gaming disputes (who measures the recovery? over what window? against what counterfactual?), and reaching that far beyond current product maturity weakens the seriousness of the whole pricing model. Revisit only when Enterprise is stable and there is a recovery-measurement methodology that customer, vendor, and a neutral third party can all defend.
Tier-design trap to avoid. Do not put the keigo-asymmetry detector behind the Enterprise tier. It is Kashi's category-defining differentiator and must be visible from the Professional tier for the sales pitch to land. Keep it in Pro; use governance depth (consultation, residency, certifications, SLAs) as the Enterprise gate.

04Competitor benchmarks

These are the reference prices in the Japanese workplace-analytics market. Competitors are anchors and sanity checks, not proof of the right price. The source column calls out confidence level — Published means disclosed list pricing, Inferred band means triangulated from public disclosures, Estimated means analyst consensus without vendor disclosure, and Weak-confidence estimate means the number should be treated skeptically.

Expand competitor benchmark table 9 vendors, with source confidence
Product Category Price Source / note
Geppo (Recruit × CyberAgent) Engagement survey (monthly 3-question) ¥298/user/mo Published geppo.jp pricing page (2025)
Wevox (Atrae) Engagement survey + ONA analytics ¥400-700/user/mo Published + inferred Published list price; volume-discount bands inferred from Atrae IR disclosures.
Culture Amp Engagement + performance surveys ¥800-1,200/user/mo Estimated Industry analyst estimates (G2, TrustRadius) for JP SMB tier. Weak-confidence comparator.
MS Viva Insights + Glint M365 telemetry + engagement survey ~¥5,000/user/mo (bundled in M365 E5) or ¥1,000-1,500/user/mo standalone Published MS pricing; Viva Suite SKU list.
Workday Peakon Engagement survey, sentiment AI ~¥1,000-1,600/user/mo Inferred band Typical Workday enterprise quote range; not published.
Archaic ハラスメントチェックAI Content-scanning harassment triage ~¥1,000-2,000/user/mo (est.) Weak-confidence estimate No published price; estimated from similar JP enterprise-SaaS bands.
FRONTEO KIBIT Eye Content-scanning email/chat triage Enterprise only · custom Not public Historically deployed at MUFG / Aeon; pricing never disclosed. Category activity has gone quiet post-2022.
15Five + Kona Manager-effectiveness coaching ~$14-20/user/mo (~¥2,100-3,000) Published 15Five pricing; Kona is an add-on.
Kashi (Pro) Governance infrastructure (structural + keigo) ¥800 per covered employee/mo (provisional) Provisional This doc is testing whether the number holds under real commercial pressure.
Kashi (Enterprise) Governance + consultation + certifications Contact-led (internal hypothesis: ¥10M base + ¥800 per covered employee/mo) Unpublished Priced against deployment complexity; base is a fixed-service fee, not software.

Positioning — the claim, not the settled fact

The positioning Kashi is arguing for: sit above engagement surveys (we do more than ask people how they feel) and below content-scanners (we refuse to read message bodies). This is a claim the market has not yet accepted as settled. Competitor prices in the table above are anchors and sanity checks, not proof that the positioning is right or that the number is correct.

The sales one-liner that follows from this table: "We price in a similar band to Culture Amp or Wevox at the top of their range, because we do something those products do not do. We price below Archaic or Viva because we refuse to do something those products do — specifically, read content." The buyer decides whether the framing holds.

05Applying the frameworks to Kashi — the recommendation

Putting frameworks §01-§04 together, the current pricing structure is directionally plausible and coherent enough to test, not validated. The table below traces each tier back to its framework rationale, with confidence marked honestly. ROI math in the "value-based rationale" column is a support for plausibility, not a proof: pattern detection alone does not cleanly demonstrate that a given company has recovered a stated yen amount.

Tier Price Value-based rationale Competitive rationale Confidence
Starter Free · ≤20 employees Value capture < ¥5M/yr at this scale; cost of serving one more free user is near-zero. Free tier standard in the category (Geppo offers free trial, Wevox offers partial free). High — keep as-is.
Professional ¥800 per covered employee/mo · ≤500 (published, provisional) A 500-person company modeled at ~¥131M/yr of team-dynamics drag with assumption-sensitive math. 10-20% modeled recovery = ¥13-26M. Charging ¥4.8M/yr = 18-37% against that modeled band. Useful as a plausibility check, not as a proof. The productivity-drag math is highly dependent on assumptions still being tested. Anchors against published engagement-survey pricing (Wevox, Geppo, Culture Amp JP band) as a sanity check, not as proof of the number. Directional — requires willingness-to-pay signal from 10 real pilot conversations before locking.
Enterprise Contact-led (internal hypothesis: ¥10M base + ¥800 per covered employee/mo) Value-based reasoning applies, but the base is primarily a fixed-service fee (deployment effort, security/compliance review, SSO+SCIM, governance setup, consultation, dedicated success) — not a software fee. The ¥10M figure is an internal negotiation hypothesis, not a published commitment. No direct competitor bundles this fixed-service layer at this tier; competitor pricing is not a useful anchor here. Low-to-medium — the base is only defensible to the extent the fixed-service scope is real and well-scoped. Keep unpublished while that package stabilizes.

Price-insensitivity bands

What happens if Pro goes to ¥1,200 per covered employee/mo (50% higher)?

What happens if Pro goes to ¥500 per covered employee/mo (38% lower)?

Final recommendation for this pass. Publish Professional at ¥800 per covered employee/mo as a provisional number. Present Enterprise as contact-led, with the ¥10M base held as an internal negotiation hypothesis until the fixed-service package is defensible in procurement terms. Collect willingness-to-pay and discount-pressure signal through the first 10 real conversations. Revisit pricing after customer #10 or at Month 12, whichever comes first. Do not change the published price during that window — instability damages credibility more than minor margin optimization recovers.
What would actually validate this pricing. Realized commercial behavior, not model neatness. Specifically: (1) signed customer prices relative to provisional list; (2) first real discount pressure and its reason — was the pushback about the number, the metric, the scope, trust, deployment burden, or category confusion? That diagnostic signal matters more than any Van Westendorp curve or value-capture band this document describes.

06Bootstrap-phase pricing — the first 0→5 customers

The frameworks above assume you have willing-to-pay data and market leverage. In Phase 1 (the pre-revenue bootstrap) you have neither. This is the section the rest of the industry treats as an afterthought.

The "no data, no leverage" problem

When you have zero paying customers, you face a chicken-and-egg: you can't price intelligently without WTP data, and you can't collect WTP data without running pricing conversations. Most founders guess a number. Better approach:

  1. Run 10 structured WTP conversations before putting any price anywhere. Use Van Westendorp (see §07). Output: a band of acceptable prices for your core target segment.
  2. Pick a provisional price in the upper half of that band. Founder psychology biases low; consciously correct upward.
  3. Put that price on the site as "Professional" and leave Starter free. Do not publish Enterprise — say "Enterprise available; contact us."
  4. Offer the first 3-5 pilot customers a founding-customer discount (see below).
  5. Charge customer #6 full list price. If you don't have 5 paying customers by Month 6, the problem is product or positioning, not pricing.

Founding-customer pricing — cheaper than pilot concessions

The JP-launch runbook §05 includes a ¥2-4M "pilot concession" line. For pre-revenue bootstrap, that line is zero because you have no revenue to concede. But you still want to compensate the first customer for the risk they take on an unproven vendor. The right instrument is founding-customer pricing:

Specific bootstrap numbers (recommended). First paid customer: ¥400 per covered employee/mo for a 2-year founding-customer window, capped at ¥500k/year for ≤100-person companies. That's roughly ¥40k/month of revenue from a 100-person pilot — meaningful psychologically (first real MRR) and easy on their budget (under the approval authority of most JP SMB CEOs).

Founding-customer concession policy (formal)

One-off commercial concessions harden into unwritten precedent fast. Fix the policy in writing before the first deal, not after.

Term Policy
Who qualifies A JP SMB or mid-market company agreeing to be the first 3-5 named customers in public references; willing to take 1-2 reference calls per quarter; willing to sign within 60 days of first scoped conversation.
Maximum customers under this policy Five. The sixth paying customer pays list. Non-negotiable.
Discount depth ~50% off provisional Pro pricing (¥400 per covered employee/mo) for the founding-customer window.
Duration 24 months from contract start. Auto-steps to list at Month 25. Explicit in contract. Never "forever."
What the customer gives in return Named-reference rights, attributable quote, 1-2 reference calls per quarter, logo usage in investor decks, right to cite as deployment case study.
Who can approve exceptions CEO only. No sales discretion below that line — the point of the policy is to prevent a sales motion from drifting into permanent ad-hoc discounting.

Why not free?

Free pilots look attractive — zero friction, zero commercial negotiation. But there are three reasons to charge something:

Starter remains free as a top-of-funnel surface. That's different from giving the first Professional customer a free pilot.

07Willingness-to-pay research — Van Westendorp

What this method is, and is not. Van Westendorp is a useful directional instrument for detecting obvious mispricing — is ¥800 obviously too low, obviously too high, or merely plausible? Ten conversations do not statistically prove a final price, and no small-sample WTP study should be treated as validation. Signed customer behavior, discount pressure, and procurement objections sit above any WTP curve in the evidence hierarchy.

Before locking any price, run the Van Westendorp Price Sensitivity Meter informally. It's a 4-question survey that produces a price-acceptance band without asking customers to directly state what they'd pay (which is unreliable). Its job at Kashi's stage is to flag a bad number, not to prove a good one.

Expand Van Westendorp method — four questions, four price points, practical plan the mechanics

The four questions

  1. Too cheap: "At what monthly per-user price would you think the product is so cheap that its quality must be suspect?"
  2. Cheap (a bargain): "At what price would you consider the product a bargain — a great deal for the money?"
  3. Expensive (but still considering): "At what price would the product start to feel expensive, but you'd still consider buying it?"
  4. Too expensive: "At what price is it too expensive to consider?"

Ask each in JPY per user per month. Collect from 10+ respondents. Plot the four curves; the intersections give you four prices:

  • Point of Marginal Cheapness (PMC) — where "too cheap" and "expensive" cross. Below this, buyers doubt quality.
  • Point of Marginal Expensiveness (PME) — where "too expensive" and "cheap" cross. Above this, buyers refuse on principle.
  • Indifference Price Point (IPP) — where "cheap" and "expensive" cross. The median buyer's sweet spot.
  • Optimal Price Point (OPP) — where "too cheap" and "too expensive" cross. The price at which the fewest buyers reject on price.

Defensible price range = [PMC, PME]. Pick within that range based on positioning goals.

Practical plan for Kashi

Van Westendorp formally requires a structured survey with 50+ respondents. Kashi doesn't have that yet. Run it informally:

  1. Build the 4 questions into your Phase-1 pilot-conversation script. Ask them toward the end of the conversation once the prospect understands what the product does.
  2. Collect responses from the first 10 pilot conversations. Do not share the current ¥800 price beforehand — it anchors their answers.
  3. At conversation #10, plot the curves in a spreadsheet. If ¥800 per covered employee/mo is inside [PMC, PME], treat that as "no obvious mispricing detected" and hold. If it's clearly below PMC or above PME, revise before customer #6. The band is directional; do not treat ten responses as statistical validation either way.

Don't spend money on formal research (SurveyMonkey Audience, Qualtrics) until after the ¥30-50M angel round. The informal version with real prospects is more predictive at this stage than a formal panel of random respondents.

08Six B2B SaaS pricing mistakes, ranked

The six traps that kill early-stage B2B SaaS pricing, ordered by how often they actually happen and how much damage they do.

1. Pricing from COGS Most common failure
Founder calculates "it costs us ¥50/user, add 3× margin, charge ¥150/user." In SaaS, marginal cost is near-zero. This calculation anchors the price at 1/10 of what the buyer would gladly pay.
Mitigation: treat COGS as a gross-margin sanity check only (keep margin >70%). Set price from value-based reasoning. See §01 Framework 1.
2. Wrong pricing metric
Charging per-seat when value scales with a different population; charging per-meeting when governance-adjacent buyers plan against annual budgets and expect low invoice volatility. Produces invoice arguments and churn for reasons the team can't name. Kashi's fix is to bill per covered employee in scope (see §02 definition) and to explicitly reject per-meeting consumption billing as category-wrong for compliance-adjacent buyers.
Mitigation: explicit metric-fit analysis in §02. Ask prospects how they budget for comparable tools; match the metric to that mental model.
3. Publishing Enterprise prices
Prospect sees the number, skips the conversation, self-disqualifies. Removes the negotiation ceiling that justifies high-touch Enterprise ACV.
Mitigation: publish Starter and Professional. Enterprise = "contact us." Exception: publish an example ACV band (¥15-25M) with a note that actual pricing depends on scope.
4. Changing prices too often
Pricing instability in the first 2 years signals "we don't know what we're doing." Kills trust with prospects who saw the old price, and kills renewal conversations with customers on the old price.
Mitigation: commit to a 12-month pricing freeze from first publication. If you must change, grandfather existing customers and notify 90 days in advance.
5. Free tier without conversion path
Starter covers the actual need of 90% of free users, so none ever upgrade. Free tier becomes an expensive charity.
Mitigation: Kashi's Starter ≤20 employees is defensible because the value at that scale is genuinely small (no Mirror, no Exec Brief). Monitor: if Starter users who hit the 20-employee cap convert at <15% within 6 months, retune the gating.
6. Permanent founding-customer discount
"I'll give you 50% off forever" — the single worst concession in early-stage pricing. Locks in a below-cost price as revenue scales.
Mitigation: founding-customer discount is time-boxed (2 years) and explicit in the contract. After the window, auto-steps up to list price. Never "forever."

0990-day pricing validation plan

Concrete actions, week by week, to move from "guessed number" to "defended number."

Expand week-by-week checklist 7 steps, Weeks 1-12 + Month 3-12
Week 1-2 — Draft pricing page + Van Westendorp script
Draft a `/pricing` page with Starter (free) + Professional (¥800 per covered employee/mo, provisional) + Enterprise (contact-led — do not publish a number). Write the 4 Van Westendorp questions into the standard pilot-conversation script. Frame WTP responses as directional input only.
Week 3-4 — Run 5 pilot conversations with WTP questions
Target: CEO or CHRO at 50-300-person JP companies (see jp-launch-runbook.html §02). Run the 4 questions toward the end of each conversation. Record the numbers.
Week 5-6 — Run 5 more conversations
Get to 10 data points. Plot the four Van Westendorp curves. Identify the [PMC, PME] band.
Week 7 — Decision point
If ¥800 per covered employee/mo is inside [PMC, PME] for the core target segment, publish the pricing page. If it's obviously outside, revise (and update business.html §4 + deck.html slide 10). Treat this as a coarse gate on obvious mispricing, not a statistical validation. Lock for 12 months.
Week 8-12 — First pilot commercial conversations
3 pilot candidates see the published pricing. Founding-customer discount (¥400 per covered employee/mo for 2 years) is the lever. Measure: how many balk at ¥400, how many at ¥800, how many at "Enterprise = contact us." Track the reason for the first real pushback — number, metric, scope, trust, deployment burden, or category confusion — more than the number itself.
Month 3-6 — First paid customer signs
Realized price for the first signed customer is the most important data point. Compare to the provisional Pro price. If realized > provisional, the provisional is too low. If realized < provisional after negotiation down, the provisional may be at the right level for buyers to feel like they negotiated well.
Month 12 — Formal pricing review
With 5-10 paying customers, re-run the analysis. This is where the ¥30-50M angel round has typically closed, so formal Van Westendorp research (50+ respondents, paid panel) becomes affordable. Revise prices if data warrants; grandfather existing customers.

10What this doc does not cover