Computed Commitment Pricing: A Research-Derived Pricing Model for Outcome-Oriented Language Training

# Computed Commitment Pricing: A Research-Derived Pricing Model for Outcome-Oriented Language Training

Mervine Gowry — Founder, Voicely Language

Working Paper · June 2026

1. Introduction

Language-learning subscription products present buyers with a menu: Basic, Pro, Premium, Annual. The labels vary; the structure is universal. Price differences reflect storage limits, offline access, or live tutor minutes — not the learner's starting level, target level, language pair, or required study hours. A francophone learning Spanish pays the same rate as a monolingual English speaker learning Japanese, despite the Foreign Service Institute's documented finding that these two learners face training timelines that differ by a factor of 2.5 (FSI, 1973/2011).

This decoupling is commercially convenient and pedagogically dishonest. It creates the impression that language acquisition scales uniformly with monthly spend, when the research is unambiguous that it does not.

This paper proposes an alternative: Computed Commitment Pricing (C2P), a pricing model in which the program duration and total cost are computed algorithmically from four learner-specific parameters rather than selected from a tier menu. The computation is transparent, research-derived, and communicated to the learner before payment. The result is a single offer — not a menu — that reflects what the learner's specific situation actually requires.

2. The Problem with Tier Menus

2.1 Price-Outcome Disconnection

A tier menu prices access, not outcome. "Premium" grants additional features; it does not specify what level of proficiency those features will deliver, over what timeline, for which language. The learner has no basis for comparing the offer to an outcome — only to other tier labels within the same product.

This is not merely a consumer protection concern. It is an alignment failure. When commercial success is measured by subscription renewals and engagement metrics (streaks, XP), not by CEFR advancement, the product's incentive structure actively works against measurable acquisition. A learner who remains permanently at A2, completing daily streaks for eighteen months, is a high-value customer under engagement-based metrics and a product failure under acquisition-based ones.

2.2 Language Distance Is Ignored

The U.S. Foreign Service Institute (FSI) has documented, over decades of training data, that language distance from English is not linear (FSI, 1973/2011). Category I languages (French, Spanish, Italian, Portuguese, Dutch) require approximately 600–750 hours to professional working proficiency. Category IV languages (Arabic, Japanese, Korean, Mandarin) require approximately 2,200 hours — a 3.6× difference in training load from the same starting point.

No major consumer language subscription product prices in this difference. A learner beginning Japanese pays the same monthly rate as a learner beginning Spanish, despite a projected acquisition timeline that is 2.5 times longer. The product receives 2.5× more subscription revenue from the Japanese learner — not for delivering more value, but for the learner's language choice.

2.3 Commitment as an Unpriced Variable

Nation (2001) establishes a minimum threshold of 10–16 encounters per vocabulary item for durable acquisition. Krashen's comprehensible input hypothesis (1982) requires sustained exposure calibrated just above the learner's current level. DeKeyser's ACT-R application to SLA (2007) demonstrates that proceduralization of grammar requires massed, distributed practice — not occasional review. All three converge on a minimum weekly study commitment for measurable acquisition: approximately 5 hours per week (1 hour per day, five days) for meaningful CEFR progression within a defined program.

Tier menus do not specify commitment. A $9.99/month plan and a $29.99/month plan can both be used for 15 minutes a week. The product receives different revenue; the learner's acquisition outcome is identical — and poor.

3. Computed Commitment Pricing (C2P)

3.1 Model Definition

C2P computes a program offer from four parameters:

1. Current CEFR level — established by diagnostic assessment (e.g., pronunciation assessment, VOICECAST™ placement)

2. Target CEFR level — declared by the learner

3. Language distance factor — FSI-derived multiplier for the target language (Category I: 1.0×; Category II: 1.3×; Category III: 1.5×; Category IV: 2.5×)

4. Weekly commitment — fixed at 5 hours/week (1h/day Mon–Fri); this is not a variable but a non-negotiable program minimum

The computation proceeds as follows:

Step 1 — Baseline acquisition hours:

Using the British Council / Council of Europe cumulative hour estimates (CEFR Companion Volume, 2020):

Baseline hours needed = CEFR_hours[target] − CEFR_hours[current]

Step 2 — Voicely efficiency adjustment:

Voicely's mode-specific efficiency multipliers (derived from DeKeyser 2007, Nation 2001, Krashen 1982) reduce required clock hours by approximately 50% versus traditional guided instruction. Efficiency-adjusted hours = baseline hours / 2.

Step 3 — Language distance adjustment:

Efficiency-adjusted hours × FSI language multiplier.

Step 4 — Program duration:

At 5 hours per week: Program months = (adjusted hours / 5) / 4.33

Rounded to the nearest whole month, labeled as V{n} (e.g., V7 for a 7-month program).

3.2 Example Computations

The Japanese learner's program is 57% longer than the French learner's for the same A0→B1 journey. This is not a pricing decision — it is a measurement of what the research says the journey requires.

3.3 Single Offer, Not Menu

C2P produces one offer. After HEXI computes the learner's program, the output is:

"Your V7 program: 7 months at $149/month. French, A0→B1. 1 hour/day, Mon–Fri."

There is no "Basic V7" and "Premium V7". There is one program at one price. The learner either commits to it or does not. This eliminates the upsell pressure inherent in tier menus and eliminates the perverse incentive to recommend longer programs for commercial gain.

4. Commitment as Product Architecture, Not Policy

4.1 The Non-Negotiable Minimum

Nation (2001) identifies 5 hours per week as a practical floor for measurable vocabulary acquisition within a defined timeline. Below this threshold, the spacing intervals required by spaced repetition systems (Ebbinghaus, 1885/1913; Wozniak, 1990) collapse — items are forgotten faster than they are reviewed. DeKeyser (2007) adds that proceduralization of grammar patterns requires massed distributed practice; below a weekly threshold, declarative knowledge fails to convert to automatic production.

C2P does not offer a 2-hour-per-week plan not because such a plan is uncommercial, but because such a plan does not work. Offering it would be selling a product known to fail.

4.2 Commitment as a Filtering Signal

Behavioral economics research on commitment devices (Ariely & Wertenbroch, 2002; Bryan, Karlan & Nelson, 2010) demonstrates that pre-commitment to a specific, measurable behavior improves follow-through rates significantly over open-ended intentions. The C2P commitment requirement (1h/day Mon–Fri, non-negotiable) functions as a purchase-time commitment device: a learner who is unwilling to commit to the minimum selects out before payment, reducing the false-positive enrollment rate and improving program completion metrics.

This is the opposite of the freemium funnel, which maximizes enrollment volume at the cost of outcome quality. C2P inverts the funnel: fewer enrollments, higher commitment, measurable outcomes.

4.3 Transparency as Trust

When the computation behind the offer is made legible — "HEXI computed your V7 program from your A0 French placement, the FSI Category I multiplier, and your B1 goal" — the learner is not accepting a product's recommendation on faith. They are reading a derivation. The research citations are part of the offer. The program duration is not a product decision; it is a measurement.

This shifts the commercial relationship from persuasion to disclosure. The product is not claiming it can get the learner to B1 in 7 months; it is citing the same research that governments and universities use to set foreign language training timelines and showing how that research maps to the learner's specific situation.

5. Comparison to Existing Pricing Models

5.1 Engagement-Based Subscription (Duolingo, Babbel)

Engagement-based subscriptions price access to a content library. Revenue is maximized by maximizing daily active users, not by maximizing learner advancement. The streak mechanic is a retention mechanism, not an acquisition mechanism. Learner outcome is not measured because it is not a business KPI.

C2P inverts this: the learner's CEFR advancement (confirmed by HEXI) is the trigger for the guarantee reward (30% discount on next V-program for V3–V17). The product's commercial incentive and the learner's acquisition outcome are aligned.

5.2 Course-Based Fixed Pricing (Pimsleur, Rosetta Stone)

Fixed-course products sell a defined content bundle (30 lessons, 90 days) at a fixed price. The bundle is the same for all learners regardless of starting level, target level, or language distance. A Pimsleur Japanese and a Pimsleur Spanish course have the same price despite wildly different acquisition timelines. There is no mechanism for computing a learner-specific program.

5.3 Tutor Marketplace (iTalki, Preply)

Tutor marketplaces price individual sessions ($15–$80/hour). Total cost is uncapped and unpredictable. No program structure, no CEFR milestone tracking, no minimum commitment requirement. The learner manages their own learning architecture — a task that research (DeKeyser, 2007) shows most adult L2 learners are not equipped to do without structured guidance.

C2P offers the predictability of a fixed-price program, the personalization of a computed offer, and the outcome accountability of a defined CEFR milestone — at a price point below the total cost of equivalent tutor hours.

6. Implications for EdTech Pricing Design

6.1 Outcome-Linking as a Design Constraint

C2P requires that the product be able to make a meaningful outcome claim. A product that cannot specify what level of competence it will deliver to which learner type over which timeline cannot implement C2P — the computation has no target variable. This design constraint is productive: it forces the product to define its acquisition theory before defining its pricing.

6.2 Diagnostic as Sales Architecture

In C2P, the diagnostic (CEFR placement assessment) is not a free trial — it is the pricing mechanism. The learner discovers their offer through their assessment result, not through a tier comparison table. This repositions the diagnostic from a marketing gimmick ("see what level you are") to a product-level commitment ("this is what your situation requires, here is why, here is the price").

Voicely's VOICECAST™ implements this: a voice-recorded CEFR placement that outputs a V-program recommendation with FSI research citation, before any payment is requested. The learner knows exactly what they are buying before the purchase decision.

6.3 Language Distance Pricing Parity

C2P produces longer — and therefore more expensive total-commitment — programs for linguistically distant languages. A Japanese learner's V11 program costs $1,639 in total versus a French learner's V7 at $1,043. This difference reflects the learner's actual training load, not a premium tier. If the Japanese learner reaches B1, they have received the same outcome as the French learner — at a higher total cost, because their journey was longer. This is honest.

The alternative — charging both learners the same monthly rate with no program structure — extracts more revenue from the Japanese learner while providing no additional outcome. C2P redistributes this revenue legitimately: more months means more program, not the same program with a price premium.

7. Conclusion

Computed Commitment Pricing is not a pricing gimmick. It is the logical consequence of taking acquisition research seriously at the product level. When the research specifies how many hours a particular learner needs for a particular outcome in a particular language, and when that research is public and verifiable, there is no principled basis for offering a tier menu instead of a computed program.

C2P produces:

• One offer derived from the learner's specific situation, not selected from a category
• Transparent derivation citing the same research used by Cambridge and the FSI
• Non-negotiable commitment grounded in Nation (2001) minimum weekly exposure thresholds
• Commercial incentive alignment — product revenue correlates with learner program length, which correlates with acquisition complexity, not with willingness to pay for "Premium"
• Outcome accountability — the guarantee (advancement confirmation → discount reward) creates a measurable loop between program completion and commercial outcome

The tier menu is an artifact of an era in which EdTech was priced like SaaS. C2P treats language training as what the research says it is: a defined, measurable, learner-specific process with a derivable timeline and an honest price.

References

Ariely, D. & Wertenbroch, K. (2002). Procrastination, Deadlines, and Performance: Self-Control by Precommitment. Psychological Science, 13(3), 219–224.

Bryan, G., Karlan, D. & Nelson, S. (2010). Commitment Devices. Annual Review of Economics, 2, 671–698.

Council of Europe (2020). Common European Framework of Reference for Languages: Companion Volume. Council of Europe Publishing.

DeKeyser, R. (2007). Practice in a Second Language: Perspectives from Applied Linguistics and Cognitive Psychology. Cambridge University Press.

Ebbinghaus, H. (1885/1913). Memory: A Contribution to Experimental Psychology. Teachers College, Columbia University.

Foreign Service Institute (1973/2011). FSI Language Difficulty Rankings. U.S. Department of State.

Krashen, S.D. (1982). Principles and Practice in Second Language Acquisition. Pergamon Press.

Nation, I.S.P. (2001). Learning Vocabulary in Another Language. Cambridge University Press.

Skehan, P. (1998). A Cognitive Approach to Language Learning. Oxford University Press.

Wozniak, P.A. (1990). Optimization of learning. Unpublished master's thesis, University of Technology, Poznan.