Pricing
Good pricing is a packaging decision before it's a number. This skill picks the value metric (what you charge on), chooses a model, sets good-better-best tiers, and pressure-tests willingness-to-pay with Van Westendorp before you commit to a price.
Grounded in: Monetizing Innovation — Madhavan Ramanujam: design around willingness-to-pay and the value metric before building.
Go deeper (The Product Channel): The Art of Pricing
When to use this
- Launching a new product or paid tier and need to set prices from scratch.
- Suspect you're underpriced or leaving money on the table at the top end.
- Deciding between per-seat, usage-based, or flat tiered pricing.
- Repackaging existing plans because customers cluster on one tier or churn at upgrade.
- Building a pricing page and need three tiers with a clear anchor and a "most popular" pick.
Before you start (gather these)
- Value metric candidates — what scales with the value the customer gets (seats, projects, API calls, GB, transactions, contacts).
- Cost-to-serve per unit of that metric (to set a price floor and protect margin on usage plans).
- Willingness-to-pay signal — at minimum a few customer quotes or a Van Westendorp survey (the four price questions below); failing that, competitor price points.
- Segments — who buys (solo / team / enterprise) and the job each segment hires the product for.
- Current numbers if repricing — plan distribution, ARPA, upgrade/churn points.
If 2+ of these are missing or vague, ASK 2-4 sharp questions before proceeding, e.g.: "What single thing grows as a customer gets more value from this?" / "Do you have any WTP data, or should I anchor off competitors?" / "Which segments are you packaging for?" / "What's your cost to serve one unit?" If the data is provided, proceed and state assumptions inline (e.g., "Assuming cost-to-serve is negligible for seats; flag if not").
Process
Pick the value metric. Score 3-4 candidates against: does it grow as the customer gets more value? is it easy to understand and predict? hard to game? cheap to meter? Choose the one that aligns your revenue with their success. A good metric makes price feel fair because the customer only pays more when they're getting more.
Choose the model. Per-seat when value scales with people (collaboration tools). Usage when value scales with consumption and is spiky (infra, API). Flat tiered when value is bundled and you want predictable revenue + simple buying. Hybrid (platform fee + usage) when you need a floor plus upside. Default to tiered unless usage clearly tracks value better.
If per-seat, choose the billable unit explicitly: provisioned seats (every seat purchased) vs. active seats (only seats that logged in / used the product in the period). This choice materially changes churn: provisioned billing maximizes near-term revenue but inflates the bill with unused seats, which surfaces at renewal as "we're paying for 200 and using 60" and drives downgrades or churn. Active-seat billing lowers headline ARPA but ties the invoice to realized value, so renewals defend themselves. State which unit you're billing and why.
Estimate willingness-to-pay with Van Westendorp — per segment. Ask (survey or proxy from interviews) the four questions: at what price is it too cheap (quality doubt), a bargain, getting expensive, too expensive. Plot the cumulative curves and read off the canonical intersections:
- Point of Marginal Cheapness (PMC) = "too cheap" × "expensive" — the lower bound of the acceptable range.
- Point of Marginal Expensiveness (PME) = "bargain" × "too expensive" — the upper bound.
- Acceptable range runs PMC → PME.
- Optimal Price Point (OPP) = "too cheap" × "too expensive" — the price with the fewest people rejecting on either side.
- Indifference Price Point (IPP) = "bargain" × "expensive" — where as many call it cheap as call it expensive (often the median/competitor anchor).
Run the four questions separately for each segment (solo / team / enterprise) — willingness-to-pay diverges sharply across them, and one blended curve hides the spread that justifies your tiers. Use the range, not a single number.
Set good-better-best tiers. Build three tiers around the WTP range. Good = lands the core job for the price-sensitive segment (the anchor's floor). Better = the target plan most buyers should pick; load it with the features the median segment values and mark it "most popular." Best = a deliberately premium anchor that makes Better look reasonable and captures high-WTP buyers. Differentiate tiers by the value metric + a small number of meaningful features, not a long checklist.
Anchor intentionally. Order tiers high-to-low or put the anchor (Best, or an Enterprise "Contact us") at the top so Better reads as the sensible middle. The decoy effect is real: a slightly-worse, similarly-priced option pushes buyers to your target tier.
Set the gaps. Price the jump between tiers so upgrading is an easy yes. The common ~2-3x rule of thumb refers to total plan price (Better's monthly bill ≈ 2-3x Good's), not per-unit/per-seat price — per-unit rates usually fall as tiers rise (volume discount), so don't apply the multiple to the unit rate. It's a defensible starting ratio, not a rule. If everyone clusters on one tier, your fences (the things that force an upgrade) are in the wrong place.
Stress-test for money left on the table. Check: is Best capped (or "Contact us") so you don't cap your biggest accounts? Is the cheapest tier capturing low-WTP buyers without cannibalizing the middle? Does any single tier hold >70% of customers (sign of bad fences)? Is the value metric something a growing customer will naturally consume more of?
Define expansion. Name the moment a happy customer outgrows their tier (more seats, more usage, a gated feature they now need). Expansion revenue beats acquisition; the metric should make it automatic.
Output template
# Pricing & Packaging: [Product]
## 1. Value Metric
**Chosen metric:** [e.g., active projects]
**Why:** [grows with value, predictable, hard to game, cheap to meter]
| Candidate metric | Scales w/ value? | Easy to predict? | Hard to game? | Cheap to meter? | Verdict |
|---|---|---|---|---|---|
| [seats] | [Y/N] | [Y/N] | [Y/N] | [Y/N] | [keep/drop] |
| [projects] | [Y/N] | [Y/N] | [Y/N] | [Y/N] | [chosen] |
| [API calls]| [Y/N] | [Y/N] | [Y/N] | [Y/N] | [drop] |
## 2. Pricing Model
**Model:** [tiered / per-seat / usage / hybrid]
**Rationale:** [why this fits how value is delivered]
**Price floor:** [cost-to-serve per unit] → never price below [floor]
## 3. Willingness-to-Pay (Van Westendorp — per segment)
Run the four questions separately per segment (median responses):
| Segment | Too cheap | Bargain | Getting expensive | Too expensive |
|---|---|---|---|---|
| Solo | [$X] | [$X] | [$X] | [$X] |
| Team | [$X] | [$X] | [$X] | [$X] |
| Enterprise | [$X] | [$X] | [$X] | [$X] |
Derived points per segment (canonical intersections):
| Segment | PMC (too cheap × expensive) | PME (bargain × too expensive) | Acceptable range (PMC→PME) | OPP (too cheap × too expensive) | IPP (bargain × expensive) |
|---|---|---|---|---|---|
| Solo | [$X] | [$X] | [$low]–[$high] | [$X] | [$X] |
| Team | [$X] | [$X] | [$low]–[$high] | [$X] | [$X] |
| Enterprise | [$X] | [$X] | [$low]–[$high] | [$X] | [$X] |
- **Source:** [n responses per segment / interview proxy / competitor anchor — state confidence]
- **Note:** never blend segments into one curve — the spread across segments is what justifies the tier prices below.
## 4. Good–Better–Best Tiers
| | [Good] | [Better] ★ Most popular | [Best] |
|---|---|---|---|
| **Price** | [$X/mo] | [$Y/mo] | [$Z/mo or "Contact us"] |
| **Value metric** | [up to N units] | [up to M units] | [unlimited / custom] |
| **For** | [segment] | [segment] | [segment] |
| **Key features** | [core job] | [core + 2-3 high-value] | [everything + premium] |
| **Upgrade trigger** | [hits N units / needs X] | [needs SSO / scale] | — |
**Anchor:** [Best / Enterprise] is positioned to make [Better] the obvious choice.
**Target tier:** [Better] — expect [~%] of buyers here.
**Tier gaps:** [Good→Better ~Nx], [Better→Best ~Nx] — *(total plan price, not per-unit rate)*.
### 4b. Usage / Hybrid variant (use instead of the table above if model is usage or hybrid)
| | [Good] | [Better] ★ Most popular | [Best] |
|---|---|---|---|
| **Platform fee** | [$X/mo or $0] | [$Y/mo] | [$Z/mo or "Contact us"] |
| **Included units** | [N units/mo] | [M units/mo] | [custom commit] |
| **Overage rate** | [$/unit above N] | [$/unit above M, lower] | [negotiated $/unit] |
| **Metered on** | [value metric] | [value metric] | [value metric] |
| **Billing basis** | [active not provisioned, if seats] | [active not provisioned] | [committed-use / true-up] |
| **For** | [segment] | [segment] | [segment] |
| **Upgrade trigger** | [consistently over N units] | [needs higher commit / lower rate] | — |
- **Per-unit rate falls as tiers rise** (volume discount) — keep every unit rate ≥ cost-to-serve floor.
- **Floor:** platform fee guarantees minimum revenue; overage captures upside.
- **Metered billing:** if pure pay-as-you-go (no platform fee), set included units to 0 and lead with the per-unit rate; protect margin with the floor on every unit.
## 5. Money-Left-on-the-Table Check
- [ ] Top tier uncapped or "Contact us" (don't cap biggest accounts)
- [ ] No single tier holds >70% of customers (fences are working)
- [ ] Cheapest tier captures low-WTP without cannibalizing [Better]
- [ ] Best tier captures high-WTP buyers (someone should buy it)
- [ ] Price ≥ floor on every tier
## 6. Expansion Path
**Outgrow moment:** [customer crosses N units / needs gated feature]
**Mechanic:** [auto-prompt upgrade / overage / seat add]
**Expected expansion:** [net revenue retention target / direction]
## 7. Open Questions & Next Test
- [biggest assumption to validate]
- [test: e.g., A/B the Better price, or run Van Westendorp on n=50]
Avoid (anti-patterns)
- Cost-plus pricing. Marking up your costs ignores what the value is worth — it caps your upside and signals low value. Price to WTP; use cost only as a floor.
- Charging on a metric that doesn't track value (e.g., per-seat for a tool one admin runs for the whole org). The customer feels nickel-and-dimed and games it.
- A single price from Van Westendorp. It gives a range; pick within it based on positioning, don't treat the OPP as gospel.
- Feature-checklist tiers. Twenty checkmarks per column hide the actual fence. Differentiate on the value metric plus 2-3 features that genuinely matter.
- Capping your top tier. A hard-numbered "Enterprise: $X" leaves your biggest, highest-WTP accounts paying less than they would. Use "Contact us."
Tips
- The middle tier should be your hero. Design Good and Best to sell Better — Good is a little too limited, Best is a little too much, Better is just right.
- Round numbers signal premium; charm prices signal value. $99 reads "deal," $100 reads "professional." Match the price ending to the segment.
- Annual upfront beats monthly for cash and retention. Offer ~2 months free on annual; it pulls forward cash and cuts churn.
- When repricing, grandfather existing customers or migrate them gently — a surprise price hike on loyal users costs more in goodwill and churn than it gains in ARPA.
1---2name: pricing3description: Designs pricing and packaging using value-metric selection, Van Westendorp willingness-to-pay analysis, and good-better-best tiering. Picks the metric you charge on, chooses the model (per-seat, usage, tiered), sets three tiers with anchoring, and stress-tests for money left on the table. Use when you say "how should we price this," "what should our tiers be," "are we underpriced," "per-seat vs usage," or "package our plans."4---56# Pricing78Good pricing is a packaging decision before it's a number. This skill picks the **value metric** (what you charge on), chooses a **model**, sets **good-better-best** tiers, and pressure-tests willingness-to-pay with **Van Westendorp** before you commit to a price.910**Grounded in:** *Monetizing Innovation* — Madhavan Ramanujam: design around willingness-to-pay and the value metric before building.11**Go deeper (The Product Channel):** [The Art of Pricing](https://sidsaladi.substack.com/p/week-41-the-art-of-pricing-11-essential)1213## When to use this14- Launching a new product or paid tier and need to set prices from scratch.15- Suspect you're underpriced or leaving money on the table at the top end.16- Deciding between per-seat, usage-based, or flat tiered pricing.17- Repackaging existing plans because customers cluster on one tier or churn at upgrade.18- Building a pricing page and need three tiers with a clear anchor and a "most popular" pick.1920## Before you start (gather these)21- **Value metric candidates** — what scales with the value the customer gets (seats, projects, API calls, GB, transactions, contacts).22- **Cost-to-serve** per unit of that metric (to set a price floor and protect margin on usage plans).23- **Willingness-to-pay signal** — at minimum a few customer quotes or a Van Westendorp survey (the four price questions below); failing that, competitor price points.24- **Segments** — who buys (solo / team / enterprise) and the job each segment hires the product for.25- **Current numbers if repricing** — plan distribution, ARPA, upgrade/churn points.2627If 2+ of these are missing or vague, ASK 2-4 sharp questions before proceeding, e.g.: "What single thing grows as a customer gets more value from this?" / "Do you have any WTP data, or should I anchor off competitors?" / "Which segments are you packaging for?" / "What's your cost to serve one unit?" If the data is provided, proceed and state assumptions inline (e.g., "Assuming cost-to-serve is negligible for seats; flag if not").2829## Process301. **Pick the value metric.** Score 3-4 candidates against: does it grow as the customer gets more value? is it easy to understand and predict? hard to game? cheap to meter? Choose the one that aligns *your* revenue with *their* success. A good metric makes price feel fair because the customer only pays more when they're getting more.312. **Choose the model.** Per-seat when value scales with people (collaboration tools). Usage when value scales with consumption and is spiky (infra, API). Flat tiered when value is bundled and you want predictable revenue + simple buying. Hybrid (platform fee + usage) when you need a floor plus upside. Default to tiered unless usage clearly tracks value better.3233 **If per-seat, choose the billable unit explicitly: provisioned seats (every seat purchased) vs. active seats (only seats that logged in / used the product in the period).** This choice materially changes churn: provisioned billing maximizes near-term revenue but inflates the bill with unused seats, which surfaces at renewal as "we're paying for 200 and using 60" and drives downgrades or churn. Active-seat billing lowers headline ARPA but ties the invoice to realized value, so renewals defend themselves. State which unit you're billing and why.343. **Estimate willingness-to-pay with Van Westendorp — per segment.** Ask (survey or proxy from interviews) the four questions: at what price is it *too cheap* (quality doubt), *a bargain*, *getting expensive*, *too expensive*. Plot the cumulative curves and read off the canonical intersections:35 - **Point of Marginal Cheapness (PMC)** = "too cheap" × "expensive" — the lower bound of the acceptable range.36 - **Point of Marginal Expensiveness (PME)** = "bargain" × "too expensive" — the upper bound.37 - **Acceptable range** runs PMC → PME.38 - **Optimal Price Point (OPP)** = "too cheap" × "too expensive" — the price with the fewest people rejecting on either side.39 - **Indifference Price Point (IPP)** = "bargain" × "expensive" — where as many call it cheap as call it expensive (often the median/competitor anchor).4041 Run the four questions **separately for each segment** (solo / team / enterprise) — willingness-to-pay diverges sharply across them, and one blended curve hides the spread that justifies your tiers. Use the range, not a single number.424. **Set good-better-best tiers.** Build three tiers around the WTP range. **Good** = lands the core job for the price-sensitive segment (the anchor's floor). **Better** = the target plan most buyers should pick; load it with the features the median segment values and mark it "most popular." **Best** = a deliberately premium anchor that makes Better look reasonable and captures high-WTP buyers. Differentiate tiers by the value metric + a small number of meaningful features, not a long checklist.435. **Anchor intentionally.** Order tiers high-to-low or put the anchor (Best, or an Enterprise "Contact us") at the top so Better reads as the sensible middle. The decoy effect is real: a slightly-worse, similarly-priced option pushes buyers to your target tier.446. **Set the gaps.** Price the jump between tiers so upgrading is an easy yes. The common ~2-3x rule of thumb refers to **total plan price** (Better's monthly bill ≈ 2-3x Good's), *not* per-unit/per-seat price — per-unit rates usually *fall* as tiers rise (volume discount), so don't apply the multiple to the unit rate. It's a defensible starting ratio, not a rule. If everyone clusters on one tier, your fences (the things that force an upgrade) are in the wrong place.457. **Stress-test for money left on the table.** Check: is Best capped (or "Contact us") so you don't cap your biggest accounts? Is the cheapest tier capturing low-WTP buyers without cannibalizing the middle? Does any single tier hold >70% of customers (sign of bad fences)? Is the value metric something a growing customer will naturally consume more of?468. **Define expansion.** Name the moment a happy customer outgrows their tier (more seats, more usage, a gated feature they now need). Expansion revenue beats acquisition; the metric should make it automatic.4748## Output template4950```markdown51# Pricing & Packaging: [Product]5253## 1. Value Metric54**Chosen metric:** [e.g., active projects]55**Why:** [grows with value, predictable, hard to game, cheap to meter]5657| Candidate metric | Scales w/ value? | Easy to predict? | Hard to game? | Cheap to meter? | Verdict |58|---|---|---|---|---|---|59| [seats] | [Y/N] | [Y/N] | [Y/N] | [Y/N] | [keep/drop] |60| [projects] | [Y/N] | [Y/N] | [Y/N] | [Y/N] | [chosen] |61| [API calls]| [Y/N] | [Y/N] | [Y/N] | [Y/N] | [drop] |6263## 2. Pricing Model64**Model:** [tiered / per-seat / usage / hybrid]65**Rationale:** [why this fits how value is delivered]66**Price floor:** [cost-to-serve per unit] → never price below [floor]6768## 3. Willingness-to-Pay (Van Westendorp — per segment)6970Run the four questions separately per segment (median responses):7172| Segment | Too cheap | Bargain | Getting expensive | Too expensive |73|---|---|---|---|---|74| Solo | [$X] | [$X] | [$X] | [$X] |75| Team | [$X] | [$X] | [$X] | [$X] |76| Enterprise | [$X] | [$X] | [$X] | [$X] |7778Derived points per segment (canonical intersections):7980| Segment | PMC (too cheap × expensive) | PME (bargain × too expensive) | Acceptable range (PMC→PME) | OPP (too cheap × too expensive) | IPP (bargain × expensive) |81|---|---|---|---|---|---|82| Solo | [$X] | [$X] | [$low]–[$high] | [$X] | [$X] |83| Team | [$X] | [$X] | [$low]–[$high] | [$X] | [$X] |84| Enterprise | [$X] | [$X] | [$low]–[$high] | [$X] | [$X] |8586- **Source:** [n responses per segment / interview proxy / competitor anchor — state confidence]87- **Note:** never blend segments into one curve — the spread across segments is what justifies the tier prices below.8889## 4. Good–Better–Best Tiers9091| | [Good] | [Better] ★ Most popular | [Best] |92|---|---|---|---|93| **Price** | [$X/mo] | [$Y/mo] | [$Z/mo or "Contact us"] |94| **Value metric** | [up to N units] | [up to M units] | [unlimited / custom] |95| **For** | [segment] | [segment] | [segment] |96| **Key features** | [core job] | [core + 2-3 high-value] | [everything + premium] |97| **Upgrade trigger** | [hits N units / needs X] | [needs SSO / scale] | — |9899**Anchor:** [Best / Enterprise] is positioned to make [Better] the obvious choice.100**Target tier:** [Better] — expect [~%] of buyers here.101**Tier gaps:** [Good→Better ~Nx], [Better→Best ~Nx] — *(total plan price, not per-unit rate)*.102103### 4b. Usage / Hybrid variant (use instead of the table above if model is usage or hybrid)104105| | [Good] | [Better] ★ Most popular | [Best] |106|---|---|---|---|107| **Platform fee** | [$X/mo or $0] | [$Y/mo] | [$Z/mo or "Contact us"] |108| **Included units** | [N units/mo] | [M units/mo] | [custom commit] |109| **Overage rate** | [$/unit above N] | [$/unit above M, lower] | [negotiated $/unit] |110| **Metered on** | [value metric] | [value metric] | [value metric] |111| **Billing basis** | [active not provisioned, if seats] | [active not provisioned] | [committed-use / true-up] |112| **For** | [segment] | [segment] | [segment] |113| **Upgrade trigger** | [consistently over N units] | [needs higher commit / lower rate] | — |114115- **Per-unit rate falls as tiers rise** (volume discount) — keep every unit rate ≥ cost-to-serve floor.116- **Floor:** platform fee guarantees minimum revenue; overage captures upside.117- **Metered billing:** if pure pay-as-you-go (no platform fee), set included units to 0 and lead with the per-unit rate; protect margin with the floor on every unit.118119## 5. Money-Left-on-the-Table Check120- [ ] Top tier uncapped or "Contact us" (don't cap biggest accounts)121- [ ] No single tier holds >70% of customers (fences are working)122- [ ] Cheapest tier captures low-WTP without cannibalizing [Better]123- [ ] Best tier captures high-WTP buyers (someone should buy it)124- [ ] Price ≥ floor on every tier125126## 6. Expansion Path127**Outgrow moment:** [customer crosses N units / needs gated feature]128**Mechanic:** [auto-prompt upgrade / overage / seat add]129**Expected expansion:** [net revenue retention target / direction]130131## 7. Open Questions & Next Test132- [biggest assumption to validate]133- [test: e.g., A/B the Better price, or run Van Westendorp on n=50]134```135136## Avoid (anti-patterns)137- **Cost-plus pricing.** Marking up your costs ignores what the value is worth — it caps your upside and signals low value. Price to WTP; use cost only as a floor.138- **Charging on a metric that doesn't track value** (e.g., per-seat for a tool one admin runs for the whole org). The customer feels nickel-and-dimed and games it.139- **A single price from Van Westendorp.** It gives a *range*; pick within it based on positioning, don't treat the OPP as gospel.140- **Feature-checklist tiers.** Twenty checkmarks per column hide the actual fence. Differentiate on the value metric plus 2-3 features that genuinely matter.141- **Capping your top tier.** A hard-numbered "Enterprise: $X" leaves your biggest, highest-WTP accounts paying less than they would. Use "Contact us."142143## Tips144- The **middle tier should be your hero.** Design Good and Best to sell Better — Good is a little too limited, Best is a little too much, Better is just right.145- **Round numbers signal premium; charm prices signal value.** $99 reads "deal," $100 reads "professional." Match the price ending to the segment.146- **Annual upfront beats monthly for cash and retention.** Offer ~2 months free on annual; it pulls forward cash and cuts churn.147- When repricing, **grandfather existing customers** or migrate them gently — a surprise price hike on loyal users costs more in goodwill and churn than it gains in ARPA.
Sidsaladi9