Price Optimization Tool

Price Optimization Tool

Build an evidence-bounded price decision from seller economics and observed behavior, then recommend a reversible test or rollout with explicit uncertainty.

Installation

npx skills add nexscope-ai/eCommerce-Skills --skill price-optimization-tool -g

Capabilities

• Audit price, demand, traffic, promotion, cost, and inventory data for comparability.
• Calculate contribution economics and hard candidate-price constraints.
• Estimate directional or numeric elasticity only when the evidence supports it.
• Compare price candidates under base, downside, and upside demand scenarios.
• Optimize regular, promotional, bundle, quantity, and good-better-best price candidates.
• Design controlled price tests with hypotheses, guardrails, confounder controls, and decision rules.
• Produce a recommendation with uncertainty, approval requirements, and a reversible rollout.

Usage Examples

Evaluate these five price candidates using my cost and sales history.

Can this dataset support a price-elasticity estimate, and what should I test next?

Build a price experiment for my top five Shopify SKUs without misleading customers.

Compare separate-item, bundle, and quantity-tier pricing for these products.

Inputs and Collection

Use seller-supplied and inspected evidence first. Collect:

• SKU, variant, channel, market, currency, tax treatment, fulfillment method, lifecycle stage, and business objective;
• timestamped regular price, realized selling price, list or compare-at price, coupons, promotions, and seller-funded discounts;
• timestamped sessions or impressions, orders, units, net revenue, cancellations, returns, and inventory availability;
• COGS, inbound freight, duties, packaging, fulfillment, payment, referral, affiliate, ad, return, and other variable costs;
• traffic source, ad spend, content or listing changes, stock status, seasonality, events, and promotion windows;
• comparable competitor offers with source, capture time, variant, pack size, availability, shipping, seller, and fulfillment;
• bundle components, attach rates, cannibalization risks, tier thresholds, and operational constraints;
• target metric, approved floor and ceiling, test duration constraints, platform rules, approver, and risk tolerance.

If material inputs are missing, ask one consolidated follow-up. If they remain unavailable, provide a provisional candidate framework and test plan, not a fabricated optimal price.

Workflow

1. Define the Decision and Evidence Boundary

State the SKU, market, channel, objective, candidate range, time horizon, and decision owner. List inspected sources and label inputs:

• Confirmed: supported by inspected evidence.
• Assumption: an explicit scenario input, not an observed fact.
• Unknown: missing information that blocks a calculation or conclusion.

Choose one primary objective, such as contribution dollars, contribution per visitor, cash recovery, revenue, sell-through, launch learning, or a constrained balance. Do not silently optimize revenue when the seller asked for profit, or units when inventory is limited.

2. Audit and Align the Data

Build a time-aligned dataset at the most reliable common granularity. Check:

• realized price rather than list price alone;
• seller-funded discount and promotion stacking;
• currency, tax, pack size, product version, channel, and market consistency;
• stockouts, suppressed listings, missing traffic, cancellations, and returns;
• changes in ads, traffic mix, content, reviews, fulfillment, competitors, and seasonality;
• sufficient observations and meaningful price variation.

Exclude or flag non-comparable periods. Do not interpret a price-demand correlation as causal when other material variables changed.

3. Calculate Unit Economics and Constraints

For each observed or candidate price:

Net Revenue = Selling Price - Seller-Funded Discounts - Refund Allowance
Contribution $ = Net Revenue - COGS - Variable Selling Costs
Contribution % = Contribution $ / Net Revenue

When percentage fees apply to selling price:

Price Floor = (Unit Cost + Fixed Variable Costs + Target Contribution $) / (1 - Variable Fee Rate)

Run base, high-return, high-ad-cost, fee-change, and promotion-stack scenarios. Keep gross margin, markup, contribution margin, and net profit distinct. Remove candidates that violate approved economics, legal or contractual constraints, platform rules, or customer-trust limits.

4. Assess Whether Elasticity Is Estimable

Use a numeric estimate only when there is sufficient clean price variation, comparable exposure, reliable quantity or conversion data, and manageable confounding. A simple midpoint diagnostic is:

Price Elasticity = ((Q2 - Q1) / ((Q2 + Q1) / 2)) / ((P2 - P1) / ((P2 + P1) / 2))

Report the observation window, units, exclusions, uncertainty, and whether the result is descriptive or plausibly causal. Segment only when sample size and decision relevance justify it.

If evidence is weak:

• state that elasticity is not reliably estimable;
• use a range of explicitly labeled demand-response scenarios;
• recommend the smallest useful controlled test;
• never substitute an unverified category benchmark and call it product evidence.

5. Model Candidate Prices

Create a candidate grid that includes the current price, economically meaningful lower and higher options, and any approved bundle or tier. For each candidate, calculate:

Expected Units = Baseline Units × Demand Response Scenario
Expected Revenue = Candidate Realized Price × Expected Units
Expected Contribution = Contribution per Unit × Expected Units
Break-Even Unit Change = Baseline Total Contribution / Candidate Contribution per Unit - Baseline Units

Show base, downside, and upside cases. If elasticity is supported, translate the estimate into a bounded scenario rather than presenting a single precise forecast. Include inventory, capacity, cash-flow, return, cannibalization, and promotion implications.

For bundles and tiers, compare component economics, customer savings, incremental units, attach rate assumptions, fulfillment cost, and cannibalization. Do not use an inflated standalone reference price to manufacture savings.

6. Select the Decision Path

Choose one of three outcomes:

• Recommend: evidence is sufficiently strong and the candidate satisfies all gates.
• Test: the candidate is plausible but uncertainty is material and measurable.
• Hold and collect data: economics, data quality, policy, or authorization is inadequate.

Rank candidates against the declared primary objective and secondary constraints. Explain why the selected option wins and what evidence could reverse the decision.