Price Elasticity of Demand (PED)

Price elasticity of demand measures how much the quantity demanded of a product changes in response to a price change. In ecommerce and fashion retail, it is the foundation of every meaningful pricing decision: understanding elasticity determines which products can bear price increases without sacrificing volume, which need aggressive discounting to clear inventory, and where the margin opportunity sits hidden in the long tail.

---

How it works

Formula

Point elasticity:

PED = (% change in quantity demanded) / (% change in price)
    = (ΔQ / Q) / (ΔP / P)

A negative result is the norm — demand falls as price rises. Absolute magnitude above 1 = elastic (demand is price-sensitive). Below 1 = inelastic (demand is relatively price-insensitive). Exactly 1 = unit elastic. (Source: Practical Ecommerce, Armando Roggio, April 2020)

Midpoint (arc) elasticity — preferred for ecommerce use because it produces the same result regardless of which direction you calculate:

PED = [(Q2 - Q1) / AVERAGE(Q1, Q2)] / [(P2 - P1) / AVERAGE(P1, P2)]

(Source: Practical Ecommerce, April 2020)

Non-linearity

Elasticity is not constant along the demand curve. It tends to be higher (in absolute terms) near major competitor price points, and changes dramatically across the product lifecycle. A product does not have a single elasticity — it has different elasticities at full price, on promotion, and in clearance. Treating them as one number is a common source of margin leakage. (Source: 7Learnings, "A guide to price elasticity," 2026)

Types

Five theoretical types (Source: Shopify, April 2025):

• Perfectly elastic — consumers walk away at any price increase
• Relatively elastic — demand is sensitive to price changes (|PED| > 1)
• Unit elastic — PED = 1; quantity falls exactly proportionally
• Relatively inelastic — demand is insensitive to price (|PED| < 1)
• Perfectly inelastic — quantity purchased does not change with price

Veblen/Giffen goods — luxury items where higher price can increase demand (signaling effect). Relevant in fashion at the premium end; price reductions can harm brand equity. (Source: Shopify, April 2025; Subrata Mukherjee, Medium, October 2020)

---

Benchmarks (as-of 2026)

Cross-category ecommerce benchmarks

Category	PED (elasticity)	Classification	Source
Aggregate US ecommerce	-1.34	Elastic	Kwon & Hartley (2026)
Electronics	-1.72	Elastic	Kwon & Hartley (2026)
Fashion/apparel	-0.89	Inelastic (category aggregate)	Kwon & Hartley (2026)

Source: Kwon & Hartley (2026), "Pricing Strategies and Consumer Price Sensitivity in E-Commerce: Evidence From U.S. Online Retailers, 2022-2025." American Impact Review, DOI 10.66308/air.e2026015. Panel of 89 US online retailers, Q1 2022–Q4 2024. Note: publisher is relatively new; treat as directionally informative pending replication.

Macro sensitivity shift: elasticity was -1.52 when CPI inflation exceeded 8% (late 2022) vs -1.21 when inflation moderated to ~3% (2024). Price sensitivity rises during high-inflation periods. (Source: Kwon & Hartley 2026)

Fashion subcategory benchmarks

Mean own-price elasticity ranges from 1.17 for T-shirts to 2.21 for Trousers across apparel subcategories, with price sensitivities between -0.24 and -0.12 per euro across six product categories. (Source: Kalla, "Price Elasticity Model for Fashion Products," Semantic Scholar, dataset from European fashion retailer. Full text requires institutional access.)

Illustrative segment ranges (uncorroborated)

• Loyal customers: ~-0.4 elasticity (inelastic)
• Price-sensitive shoppers: ~-3.0 (highly elastic)

[!unverified] These segment-level figures appear in multiple secondary sources but without a named primary source. Treat as illustrative only, not validated benchmarks.

Profit sensitivity

1% price improvement = 6% effect on profitability for a typical S&P 500 company (assuming no volume loss). (Source: McKinsey, "How to navigate pricing during disinflationary times," May 2024)

1% price increase = 8.7% jump in operating profits (separate McKinsey citation via Revology Analytics, Armin Kakas, November 2024). Both figures are directionally consistent but use different baselines — the spread may reflect different company profiles or methodological differences.

---

Fashion-specific dynamics

The fashion data problem

Price elasticity estimation requires multiple data points at different price levels. A single fashion SKU — seasonal, not replenished — may only exist at one price point for most of its life, making individual-SKU elasticity estimation essentially impossible without attribute-based or cluster-based methods. This is the structural reason rule-based pricing persists in fashion. (Source: Subrata Mukherjee, Medium, October 2020)

Categories of fashion product for elasticity purposes (Source: Mukherjee 2020):

• Core/year-round (basics, essentials): more data points, more reliable elasticity, more inelastic
• Seasonal (SS/AW collections): limited data per SKU; elasticity changes dramatically across the season
• Trend-driven/one-collection: almost no usable own-price history; must estimate from attribute similarity

Seasonal elasticity dynamics

• Early season: moderate elasticity. Buyers purchasing for planned use. A 10–15% markdown can drive meaningful volume lift.
• Mid-season: lower elasticity. In-season = functional need still present; full price is acceptable.
• End of season (clearance): elasticity spikes then collapses. Swimwear in early June is elastic (15% markdown triggers demand spike); swimwear in mid-August can remain inelastic even at 50% off — the customer's need has passed, not just their willingness to pay.

(Source: Peak.ai, Tom Summerfield, September 2025)

20% of markdowns are unnecessary — products would have cleared without discount. (Source: Peak.ai, Tom Summerfield, September 2025 — vendor practitioner claim)

Size-level elasticity

Markdown decisions made at the style level can systematically over-discount sizes that would have cleared at full price while still holding inelastic excess in others. AI platforms that model size-level demand signals can recommend differentiated pricing by size within the same style. (Source: Quicklizard, "Dynamic Pricing for Fashion Retail: How to Choose a Platform in 2026")

Discount training effect

Repeating reactive markdowns season after season trains customers to wait for discounts on the sizes they want. This erodes full-price sell-through over time and compounds elasticity measurement difficulty because observed demand at full price is suppressed by learned customer behaviour. (Source: 7Learnings, 2026)

---

How retailers use elasticity

KVI tiering — the core framework

Key Value Items (KVIs) are highly price-visible, comparison-shopped, and typically elastic. They anchor the retailer's price image. KVIs typically account for "10 to 20 percent of a retailer's sales" but drive price perception disproportionately. (Source: McKinsey, "Pricing and promotions: The analytics opportunity," June 2021)

Long-tail/background SKUs are less visible and more inelastic — this is where the margin opportunity sits. Modest price increases pass unnoticed because customers are not actively comparing.

Amazon maintains its low-price reputation by "undercutting competitors on top-selling, high-visibility products, while protecting margins by charging more for less price-sensitive items" — an explicit elasticity-tiering strategy. (Source: McKinsey, "How retailers can drive profitable growth through dynamic pricing," March 2017)

One international online retailer sold "about 80 percent of its assortment below the list price" before an elasticity-analytics overhaul — evidence of widespread under-use of elasticity tiering. (Source: McKinsey, June 2021)

Measurement approaches in ecommerce (as-of 2026)

1. Historical transaction regression: log-log demand models with retailer and time fixed effects applied to past price/volume data.
2. A/B price testing: show different prices to two user segments simultaneously; measure conversion and AOV.
3. Switchback/time-split testing: alternate price weekly in a control/variant cycle over 4–6 weeks. Recommended over standard A/B for ecommerce because a list price propagates to Google Shopping, retargeting, and email simultaneously — showing different prices per-user in real time creates inconsistency and ad policy risk.
4. Conjoint analysis: survey-based willingness-to-pay estimation; useful for new products without sales history but systematically underestimates elasticity vs live price tests — practitioners warned not to raise prices based solely on apparent conjoint inelasticity.
5. ML cluster models: learn elasticity across similar product clusters so data-poor SKUs inherit signal from data-rich ones — the key mechanism enabling fashion SKU-level estimation.

(Sources: 7Learnings 2026; Speero, Paul Randall, August 2026; Professional Pricing Society / Quantiz, André Koeppl, 2015)

Practitioner warning (Speero, 2026): On one RS PRO project, teams assumed prices were too high and costing volume. Customer research revealed the brand was already priced below competitors and perceived as high-quality for the price. Elasticity testing confirmed no problem to fix. Understand price perception before running a test.

Breakeven volume calculation

Before any price test, define the breakeven volume lift required to make the price cut profitable:

Breakeven volume lift = price reduction % / (margin % − price reduction %)

At 30% margin and 5% price cut: breakeven = 20% volume lift (implies PED ~-4.0). At 50% margin: breakeven = 11% volume lift (implies PED ~-2.2). Higher-margin products tolerate price cuts more easily. (Source: Speero, Paul Randall, August 2026)

Tariff pass-through — an applied elasticity case

US apparel import tariffs rose from 14.7% (December 2024) to 35.1% (December 2025). (Source: shenglufashion.com, March 2026 — as-of December 2025; volatile, subject to ongoing litigation)

An Oliver Wyman (2025) case: specialty retailer used broad sub-5% increases across tens of thousands of SKUs, crossed price thresholds, and lost volume that never recovered. Alternative elasticity-targeted approach (larger increases on 4,000 inelastic SKUs with pricing power; price investment on ~200 elastic/price-visible items): +9% margin growth and +5% sales growth. (Source: 7Learnings citing Oliver Wyman 2025)

---

Channel and segment elasticity

DTC vs marketplace

The same product consistently shows higher price elasticity on a marketplace than on a brand's own DTC site. Marketplace shoppers can compare competing offers side-by-side within seconds; on a brand's own site, the comparison set is narrower and brand relationship provides an elasticity buffer. Pricing both channels with a single rule ignores a structural difference in customer behaviour. (Source: 7Learnings 2026; Omnia Retail DTC Strategy Guide)

Some brands deliberately price DTC below marketplace to drive customers to their highest-margin channel.

Segment variation by fashion tier

(Source: Subrata Mukherjee, Medium, October 2020)

• Mass/economy fashion: lower elasticity (buyers have few switching options or are brand-agnostic at very low price points)
• Mid-market and premium fashion: highest elasticity in fashion — most likely to respond to end-of-season sales and price comparison
• Luxury fashion: lower elasticity (social signalling reduces price sensitivity; deep discounts can harm brand equity)

Cross-price elasticity (substitution effects)

Cross-price elasticity = % change in demand for product A / % change in price of product B.

Within a fashion portfolio, cross-price elasticities for close substitutes (different colourways of the same style) often range from 0.1 to 0.5. Values above 0.5 indicate significant cannibalization risk. (Source: Revology Analytics, "Cross Price Elasticities: 7 Practical Steps To Win With")

Cannibalization risk: marking down one colourway can pull demand away from full-price sibling SKUs rather than expanding the total market. Cross-elasticity modelling in pricing tools (e.g., Competera) explicitly handles this. (Source: 7Learnings 2026; Competera product page 2026)

---

Dynamic pricing and AI tooling (as-of 2026)

Dynamic pricing intensity increased measurably: coefficient of variation in daily prices rose from 0.042 (Q1 2022) to 0.067 (Q4 2024), a 59.5% increase across 89 US online retailers. (Source: Kwon & Hartley 2026)

Despite this, 64% of retailers describe their pricing as "largely manual and experience-led" and only 4% use fully predictive and automated pricing (as-of 2026). (Source: 7Learnings citing RetailHive Pricing Benchmark — vendor-commissioned benchmark; directional finding consistent with McKinsey observations)

Retailers using AI/ML pricing grew sales 14.2% from 2023 to 2024 vs 6.9% for non-AI retailers. (Source: Kwon & Hartley 2026)

Key vendor claims (not independently audited)

Vendor	Claimed outcome	Source
Competera	+3–7% revenue growth, +2–5pp margin	Vendor product page (2026)
7Learnings	10–15%+ average profit uplift (fashion)	Vendor case studies (2026)
7Learnings — INTERSPORT Krumholz	+118% profit, +52% revenue (A/B test)	Vendor case study (2026)
7Learnings — Outletcity	+26% profit, +11% revenue (end-of-season)	Vendor case study (2026)
Peak.ai	30–50% fewer deep markdowns; 200–500bps margin improvement	Vendor claim (Sept 2025)

Promotional frequency: diminishing returns set in beyond three promotional events per quarter. (Source: Kwon & Hartley 2026)

Charm pricing: prices ending in .99 improve conversion rates by 3.2% relative to round-number prices. Charm pricing prevalence is highest in fashion (68.4% of products) and lowest in grocery (31.2%). (Source: Kwon & Hartley 2026)

---

Consumer price sensitivity shifts (as-of 2024)

• "Over a third of consumers have tried different brands, and approximately 40% have switched retailers in search of better prices and discounts" (as-of mid-2024). (Source: McKinsey ConsumerWise survey, 15,000+ consumers, 18 markets, June 2024)
• 76% of consumers reported trading down in Q3 2024, rising to 86% among Gen Z and millennials. (Source: McKinsey US consumer sentiment 2024)
• 36% of consumers plan to purchase private-label products more frequently — indicates high cross-price elasticity toward private label. (Source: McKinsey ConsumerWise survey 2024)
• Sustainability premiums eroded: percentage of young consumers willing to pay a premium for sustainability claims declined by up to 4 percentage points across product categories (Europe and US, early 2024 vs 2023). (Source: McKinsey ConsumerWise June 2024)

---

McKinsey five-factor pricing framework

McKinsey (2021) recommends linking pricing analytics across five factors: base pricing, markdown, promotions, competitor pricing, and price communication. Companies that activate broad margin management levers across pricing, portfolio, mix, promotions "outperformed peers on gross and EBITDA margins by up to 50%." (Source: McKinsey, May 2024)

Bain (2019) survey of ~1,100 companies across consumer industries: "78% of respondents say their pricing decisions could be improved." "Creating positive price perceptions is the single most important pricing capability for retailers, by a factor of more than 2x" — more important than setting the lowest absolute prices. "At least 20% of promotional activity" falls into categories that do not yield meaningful incremental sales or hurt the bottom line. (Source: Bain, "The Pricing Is Right," September 2019, updated April 2026)

---

Key terms

Term	Meaning
PED	Price Elasticity of Demand: how much demand changes with price
Elastic	|PED| > 1: demand responds strongly to price changes
Inelastic	|PED| < 1: demand is relatively insensitive to price
KVI (Key Value Item)	High-visibility product driving price perception; typically elastic
Cross-price elasticity	How demand for A changes when price of B changes (substitution/complement)
Arc elasticity	Midpoint formula; preferred over point elasticity for ecommerce
Markdown elasticity	Elasticity specific to clearance price reductions (typically different from regular price elasticity)
Charm pricing	Prices ending in .99 or .95; +3.2% conversion vs round numbers (Kwon & Hartley 2026)
Discount training effect	When repeated markdowns train customers to wait for discounts
Switchback testing	Weekly alternating price experiment; preferred over A/B for list-price testing
DML	Double Machine Learning: causal elasticity estimation approach

---

What practitioners report

Five practitioner myths from real project experience (Source: Professional Pricing Society / Quantiz, André Koeppl, 2015):

1. Elasticity is static — it is not. Technology products' estimates can become obsolete in 6 months; re-estimate on a 12–24 month rolling basis.
2. One number fits all — discount elasticity is consistently higher than regular price-change elasticity; distinguish which type of move you are measuring.
3. Elasticity is linear — non-linear; abrupt changes occur at price thresholds.
4. All methods converge — they don't. Conjoint shows lower elasticity than live price tests for the same product.
5. One-variable function — own-price-only model had quadratic error sum of 52.786; model incorporating cross-elasticities from three competitors had error sum of 0.085 — 623× more accurate.

---

Next frontier topics

Price Perception · Value-Based Pricing · Promotional Elasticity · Willingness to Pay · Demand Forecasting (link) · Key Value Items (KVIs) · Competitor Price Monitoring · Markdown Calendar