Price and package a product

To: CEO From: Head of Product Date: January 2027 Subject: Alternative to the 30% Across-the-Board Price Increase: Shifting to Usage-Based Pricing

1. Diagnosis: We Have a Metric Problem, Not a Price-Level Problem

Tom is right to sound the alarm on our gross margin falling from 41% to 24.4% ($10,775 margin on $44,127 monthly revenue). However, a flat 30% price hike across seats and enterprise contracts will not fix our unit economics, and it risks triggering churn among our highest-margin customers.

Our fundamental issue is a mismatch between our value metric (seats) and our cost driver (pages processed): * Startups (50 accounts): Generate $9,999/mo in revenue and cost $1,297 to serve (87.0% gross margin). They average 21.6k pages/month. * Mid-Market (24 accounts): Generate $15,147/mo and cost $8,454 to serve (44.2% gross margin). Three high-volume accounts (P, Q, V) are already gross-margin negative. * Enterprise (6 accounts): Generate $19,000/mo and cost $23,232 to serve (-22.3% gross margin). Four of our six Enterprise customers are underwater. Northwind Logistics alone processes 4.82M pages, costing us $5,784 while paying $3,200.

Because our API is integrated by a few developers and runs headless in production, customers scale page volume massively without adding seats. Even if we could magically apply Tom’s 30% hike to all Enterprise contracts tomorrow (ignoring that four don’t renew until Mar–Nov), Enterprise revenue would rise to $24,700 against $23,232 in costs—a dismal 5.9% gross margin. Meanwhile, we would penalize our 87%-margin Startups, giving them a reason to look at Formwise ($79/seat).

---

2. Critique of the December Price Test

We cannot use the Growth team’s price test to justify a broad 30% increase for four reasons: 1. Severe Sample Ratio Mismatch (SRM): The test intended a 50/50 split across 2,110 visitors, but routed 1,240 (58.8%) to Control and 870 (41.2%) to Variant ($\chi^2 = 64.9, p < 0.0001$). Our randomization or tracking broke—likely due to caching, bot traffic, or self-selection before the event fired. The readout is invalid. 2. Significantly Underpowered: With only 104 total conversions, the 95% confidence interval for the difference is roughly $\pm 1.9\%$. The test could not detect anything smaller than a ~38% relative drop in conversion. "Not statistically significant" here means "insufficient data," not "no effect." 3. Wrong Cohort & Seasonality: A two-week window in mid-December captures holiday anomalies and tests initial sign-up conversion of small self-serve accounts, telling us nothing about 90-day retention, seat expansion, or Enterprise contract elasticity. 4. Ignoring Usage Economics: Raising the seat price to $129 still leaves unlimited pages uncapped, inviting adverse selection from heavy API users.

---

3. Proposed Pricing Model: Platform + Usage (Hybrid)

Competitors price on pages ($0.008–$0.015/page) or cap pages per seat (Formwise at 50k/seat). Our cost to serve is $0.0012/page (falling 10%/quarter). We must align revenue with page volume while retiring the "seat" tax for API products.

The New Model: * Starter (Self-Serve): $199/month platform fee. Includes 50,000 pages/mo and unlimited seats. Overage: $0.006 per page. * Growth (Self-Serve): $499/month platform fee. Includes 150,000 pages/mo and unlimited seats. Overage: $0.004 per page. * Enterprise (Annual Contract): $1,500/month ($18k/yr) platform fee. Includes 500,000 pages/mo and unlimited seats, plus SLA/SSO. Tiered overage: * 500k – 2M pages: $0.0035 per page * 2M+ pages: $0.0025 per page

Why this works: Even our lowest Enterprise tier ($0.0025/page) yields a 52% gross margin on marginal pages today, expanding to 67% over the next year as inference costs drop 10% per quarter. Meanwhile, we remain 50% to 70% cheaper per page than DocuLens ($0.008–$0.012) and ExtractIQ ($0.015).

---

4. Segment & Financial Impact

The table below compares our current state, Tom’s 30% hike (assuming 100% immediate realization and zero churn), and the proposed model (assuming customers optimize for the lowest-cost tier at current volumes):

(Note: Table sums exact per-account figures from the export; CSV rounding accounts for minor differences from the $44,127 / 24.4% total).

Customer Impact by Segment: * Startups (Neutral/Positive): Every single Startup processes $\le 40,000$ pages. On Starter ($199/mo), all 50 accounts are 100% covered by the 50k included allowance. The 17 accounts with 3 seats see a 33% price drop (from $297 to $199); the 19 accounts with 2 seats pay the exact same ($198 vs. $199); the 14 single-seat accounts pay $100 more (which we will grandfather for 6 months). Removing seat limits encourages whole-team adoption. * Mid-Market (Rebalanced to Value): All 24 accounts optimize onto the Growth plan ($499 base + $0.004 overage). Low-volume, high-seat accounts (e.g., Mid-market L at 7 seats / 134k pages) actually see a $194/mo price reduction, protecting them from churn. High-volume, unprofitable accounts (e.g., Mid-market Q at 544k pages, currently -$59 GM) move from $594/mo to $2,075/mo (68.5% GM). * Enterprise (Profitability Restored): Average effective price becomes $0.0033/page—still ~65% below DocuLens, giving us massive pricing power while turning a -$4.2k monthly loss into a +$41.4k gross profit.

---

5. Rollout & Enterprise Renewal Playbook

Self-Serve Rollout (February 1): 1. New Sign-ups: Launch Starter and Growth tiers immediately on the pricing page. 2. Existing Self-Serve: Migrate accounts whose bill decreases or stays flat (37 of 74 accounts) immediately in February—this builds goodwill and unlocks unlimited seats. For accounts whose bill increases (14 single-seat Startups and 23 high-volume Mid-Market accounts), grant a 90-day transition window at legacy rates, paired with new dashboard usage alerts and an annual-commit discount option.

Enterprise Renewals (Account-by-Account): We cannot unilaterally hike annual contracts in January, as 5 of 6 renew later in the year. However, we have two immediate renewals and strong competitive leverage: * Copperfield Freight (Renews Jan 2027 | 2.38M pages | Currently $2,400/mo, -$456 GM): Immediate priority. Under the new card, their usage prices at $7,700/mo ($92.4k/yr). Playbook: Offer a 1-year transition bridge at $5,500/mo ($66k/yr)—a 48% gross margin for us, and still less than a third of what DocuLens ($19.5k/mo) would charge them. * Northwind Logistics (Renews Feb 2027 | 4.82M pages | Currently $3,200/mo, -$2,584 GM): Their pages tripled in a year and they want a 2-year lock at the same rate. Locking in -$31k/yr in losses for two years is a non-starter. Standard new pricing is $13,800/mo. Playbook: Show them their 3x volume growth and competitor benchmarks ($39k/mo on DocuLens). Offer the 2-year lock they want in exchange for a stepped commitment: $8,500/mo in Year 1 (47% GM today, rising to >60% as costs fall) and $11,500/mo in Year 2. * Arcadia Health (Renews Mar 2027 | 3.31M pages | Currently $4,500/mo, +$528 GM): Evaluating a competitor. Paying $0.00136/page today. Playbook: Any competitor they are evaluating will quote them $25k–$40k/mo for 3.3M pages. Position our new model aggressively: waive the $1,500 platform fee for Year 1 and lock in a flat $0.0028/page ($9,268/mo) on a 2-year deal. We double our revenue, hit a 57% gross margin, and remain 70% cheaper than the competitor they are evaluating. * Halden (Jun), Meridian (Sep), Brightwater (Nov): Begin QBRs 120 days prior to renewal, sharing usage telemetry that highlights their cost-per-page savings vs. market benchmarks before introducing the new rate card.