Pricing Strategy

Overview

Price is a structural decision, not a calculated number. Cost-plus and competitor-matching both ignore 50 years of pricing research: what people pay is shaped by reference points, anchoring, loss aversion, and offer structure — not by cost. Kahneman & Tversky (Econometrica 1979): losses hit ~2× harder than equal gains. Thaler (1980): endowment effect and mental accounting drive consumer pricing behavior.

Compose with: first-principles · probabilistic-thinking · pareto-principle · pmf-crossing-the-chasm.

When to Use

Apply when: setting initial prices; planning a price change (especially raising); designing freemium/tiered/usage structures; sales asks for discounts >1/week; competitors' price is the only input; pricing an AI product against volatile/falling inference costs and choosing seat- vs. usage- vs. outcome-based models, protecting gross margin as AI capex and model releases shift the cost floor, or defending price against AI-native competitors pricing off the same collapsing token cost.

When NOT to use: no demonstrated value (use lean-startup); price regulated; purely tactical single-deal discount; LTV/CAC already working and question is execution only.

Coaching Novices (Adaptive Front Door)

• Engine mode: user has segment + value data → run The Process directly.
• Coach mode: vague or unfamiliar → guide step by step.

In Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.

1. What it is. Pricing is a designed signal — what people pay depends on reference points and loss aversion far more than on cost.
2. Check fit. No demonstrated value → lean-startup. Regulated → wrong scope.
3. Elicit value. What outcome does the customer get, in their metric? ("save 8 eng-hours/week" not "makes them faster")

[WAIT — do not advance until user responds]

4. Walk The Process one step at a time with their input.

[WAIT — do not advance until user responds]

5. Close with one price, one anchor, one structure, one 60-day experiment.

[WAIT — do not advance until user responds]

The Process

Run the Pricing Audit. Value-first, anchor, structure, frame, test.

1. Articulate value to the customer in customer units. Not "our software does X" but "saves them $50K/year in engineering-hours." If you cannot name the value in the customer's metric, pricing is guesswork.
2. Estimate WTP per segment. Use Van Westendorp Price Sensitivity Meter (4-question survey), conjoint analysis, or direct evidence from paid pilots (B2B — most reliable).
3. Choose an anchor. First price seen frames every subsequent price. Anchor high if defensibly justified; the high anchor lifts the entire tier structure.
4. Design the tier structure. 3 tiers (Starter / Pro / Enterprise): Starter makes Pro feel affordable; Enterprise anchors and captures top WTP; Pro is the target sweet spot.
5. Frame for loss aversion. "Save $X by paying annually" beats "monthly costs $Y more" ~30%. Money-back guarantee converts ~10–30% higher than free trial. Per-user vs. flat-fee: per-user for SMB, flat-fee captures more enterprise value.
6. Stress-test the anchor. Show prices to 5–10 target-segment buyers: "expensive but worth it?" / "too expensive?" / "I'd buy at this price." Adjust if segment reference point is materially below anchor.
7. Run a 60-day measurement. Track: conversion rate, upgrade rate (Starter → Pro), discount-request frequency, revenue per tier. High upgrade rate + low discount-request frequency = right structure.

Output: Pricing Audit (fill after each step)

Field	Your answer
Value to customer (customer units)
Segment WTP range + method
Chosen anchor + justification
Tier structure (Starter / Pro / Enterprise)
Loss-aversion framing chosen
Anchor stress-test result (5–10 buyers)
60-day metrics + re-evaluation date

→ Method in Action: De Beers and the Engagement Ring (1947 → ongoing) · Netflix's Qwikster Failure (2011)

→ 2026 lens: Pricing an AI product under volatile inference costs — seat vs. usage vs. outcome (2023–2026)

Pricing Packs

Domain patterns (anchors / tier structure / key framing / dominant failure):

• B2B SaaS: TCO-comparison / 3-tier Team-Pro-Enterprise / "save 20% annual" / underpricing Enterprise.
• Consumer subscriptions: Category norms / 2-tier Free-Paid / free-trial auto-conversion / free tier too generous.
• Enterprise services: Peer-engagement / project-priced / cost-of-NOT-engaging / hourly billing.
• Luxury/signaling: Social reference points / quality-attribute ladder / identity framing / anchor commodifies.

Applying It Well

• Price is a designed signal. Cost-plus leaves money on the table; competitor-matching anchors you to someone else's (possibly wrong) positioning.
• The anchor matters more than the price. First number frames every subsequent one. Anchor high if defensibly justified.
• Loss aversion ~2× gain framing. Frame upgrades around what's lost by not upgrading.
• WTP variance is larger than founders expect. Top-segment WTP often 5–10× bottom for the same product.
• Discount frequency is the leading indicator. >1/week = structure is wrong. Fix structure, not the deal.

→ Primary sources: references/sources.md

Common Rationalizations

[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.

Fake move	Reality
[D] Cost-plus pricing	Cost is unrelated to value. Systematically underprices high-value products, overprices commodity ones.
[D] "Competitors charge X so we should too"	Anchors you to their (possibly wrong) positioning for their segment. Use as data, not anchor.
[D] Discounting deal-by-deal	Erodes the public anchor, training all customers to negotiate. Fix structure, not the deal.
[D] Single-tier pricing	Misses segment-WTP variance. Tier structure captures multiple WTP points without changing the product.
[D] "We should just charge more"	Without identifying which segment pays more for what value, this is wishful thinking.
[D] Underpricing to "establish" first	Initial pricing anchors permanent expectations. Raising later triggers loss aversion. Launch at intended price.
[D] Free tier too generous	Eliminates the loss-aversion upgrade lever. Free = enough to taste, not enough to satisfy.
[D] Ignoring loss-aversion framing	"Save $200/year" converts ~30% better than "monthly costs $200 more" for identical economics.
[D] No WTP measurement	Setting price without Van Westendorp, paid pilot, or conjoint is guessing. Free methods exist.
[D] No 60-day re-evaluation	Pricing is a hypothesis. Conversion data should refine the structure, not be ignored.
→ Add [O] entries here after each real use — paste the actual failure pattern	What went wrong and why

Red Flags

• Rationale is "cost + markup" or "competitor matching" without value reasoning
• Single tier only; no A/B test on framing; no WTP measurement
• Discount requests >1/week; free tier delivers most of the product's value
• No re-evaluation in 12+ months despite material business changes

Verification

• Value in customer units (not features); WTP estimates with method named
• Anchor explicitly chosen and justified; stress-tested with 5+ buyers
• Tier structure: middle tier is the designed sweet spot
• At least one loss-aversion framing applied
• 60-day measurement plan with re-evaluation date

---

Part of deciqAI Knowledge Skills — 227 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. See it run → https://www.deciqai.com/c/pricing-strategy · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.

Agents: latest version & machine-readable metadata → https://www.deciqai.com/s/pricing-strategy.json