Timing Markdowns & Clearance — Interactive Pricing Theory

Optimizing end-of-season price reductions to maximize revenue from a fixed stock of perishable or fashion goods, balancing the tension between selling early at high prices and risking leftover inventory.

The Problem

Every fashion retailer faces the same dilemma at the end of a season: shelves are stocked with garments that will soon be obsolete, and the clock is ticking. Unlike commodity goods that can sit in a warehouse indefinitely, seasonal products have a hard deadline. After the season ends, unsold units go to off-price channels at a fraction of their original value—or are written off entirely.

The traditional response is the markdown: a permanent price reduction intended to accelerate sales. But the timing and depth of markdowns matter considerably. Cut too early or too deep, and revenue is lost from customers who would have paid full price. Cut too late or too modestly, and excess inventory remains when the season ends.

Definition — Markdown Management Problem

Given a fixed initial inventory

I_0

, a finite selling horizon of

T

periods, and no opportunity for replenishment, choose a sequence of prices

p_1 \ge p_2 \ge \cdots \ge p_T

(monotonically non-increasing) to maximize total expected revenue over all periods.

The problem has several distinctive features that set it apart from other pricing models:

No replenishment. Inventory only decreases. Every unit sold today is one fewer available tomorrow.
Finite horizon. The selling season has a definite end. Unsold inventory has a known salvage value (often zero).
Monotone pricing. Prices can only decrease—raising prices mid-season would confuse customers and damage the brand.
Demand uncertainty. Customer arrival rates and willingness-to-pay are stochastic, making the sell-through rate at any given price uncertain.

This setup is ubiquitous in fashion retail, consumer electronics at end-of-life, and any industry where products have short life cycles.

Mathematical Formulation

We build up the markdown model in stages, starting from the simplest single-period case and extending to a full multi-period dynamic program. Throughout, we assume a population of customers whose willingness-to-pay (WTP) follows a normal distribution with mean $\mu$ and standard deviation $\sigma$ . At price $p$ , the expected demand in a single period is:

d(p) = m \cdot \left[1 - \Phi\!\left(\frac{p - \mu}{\sigma}\right)\right]

(1)

where $m$ is a market-size scaling factor and $\Phi$ is the standard normal CDF. Lowering the price brings more of the WTP distribution above the threshold, thereby increasing demand.

Single-Period Model

In the simplest case, the retailer sets a single price $p$ for the entire season and cannot adjust it. The problem becomes:

\max_{p} \; R(p) = p \cdot \min\bigl(d(p),\, I_0\bigr)

(2)

The minimum reflects the inventory constraint: it is not possible to sell more than is available. If demand exceeds inventory (a stock-out), some willing customers leave without purchasing. The optimal price balances:

Price effect: Higher $p$ increases per-unit revenue.
Volume effect: Higher $p$ reduces the number of units sold.

Single-Period Intuition

Suppose there are 100 units and the WTP distribution is

\mathcal{N}(50, 15^2)

. Setting

p = 80

means only customers with WTP

\ge 80

purchase—roughly 2.3% of the market. Revenue per unit is high but volume is minimal. Setting

p = 20

captures 97.7% of the market, but at a low margin. The optimal price lies between these extremes, depending on the inventory level relative to market size.

Two-Period Model

Now allow exactly one markdown. The retailer chooses an initial price $p_1$ for the first half of the season and a lower price $p_2 \le p_1$ for the second half:

\max_{p_1, p_2} \; p_1 \cdot s_1 + p_2 \cdot s_2

(3)

where

s_1 = \min\bigl(d(p_1),\, I_0\bigr), \quad s_2 = \min\bigl(d(p_2),\, I_0 - s_1\bigr)

(4)

The two-period model introduces the core tradeoff of markdown management: the first-period price determines how much inventory is carried over into the markdown period. A high initial price preserves margin but risks a heavy inventory burden in period 2, forcing a deeper markdown.

Optimality Condition (Two Periods)

At the optimum, the marginal revenue of the last unit sold in period 1 equals the marginal revenue of the first unit sold in period 2. The retailer should mark down at the point where continuing to sell at the high price produces less incremental revenue than switching to the lower price. This is the markdown analogue of the equal-marginal-revenue principle in price differentiation.

T-Period Optimization

The general formulation considers $T$ selling periods with a monotonically decreasing price path. Let $I_t$ denote inventory at the start of period $t$ . The dynamic programming recursion is:

V_t(I_t) = \max_{p_t \le p_{t-1}} \left[ p_t \cdot s_t + V_{t+1}(I_t - s_t) \right]

(5)

with boundary condition $V_{T+1}(I) = v \cdot I$ , where $v$ is the salvage value of unsold units (often zero). The sales quantity $s_t = \min(d(p_t), I_t)$ enforces the inventory constraint.

Solving this exactly requires discretizing both the price space and the inventory state space. With a price grid of size $K$ and inventory levels up to $I_0$ , the computational complexity is $O(T \cdot I_0 \cdot K)$ , which is tractable for practical problem sizes.

A key property of the optimal solution: prices decrease faster early in the season when inventory is high relative to remaining demand, and slow down as inventory approaches a sustainable sell-through rate. This contrasts with the naive approach of equal-sized markdowns, which often leaves too much inventory for the final periods.

Interactive Explorer

Experiment with the markdown simulator and related visualizations. Compare optimal schedules against naive strategies, and observe how the number of markdown periods affects total revenue.

The blue line shows the optimized price schedule. The gray dashed line shows a naive linear strategy. The amber line tracks inventory depletion.

Key Insights

Several important patterns emerge from the markdown optimization model:

Front-loading revenue is optimal. The optimal strategy typically holds prices higher early in the season (when all inventory is available) and accelerates markdowns as the deadline approaches. This contrasts with naive equal-step strategies that give away margin too early.
Inventory-to-demand ratio drives markdown depth. When initial inventory is large relative to expected demand, the optimal strategy requires deeper and earlier markdowns. When inventory is scarce, the retailer can afford to hold prices longer.
WTP dispersion matters. A wider WTP distribution (larger $\sigma$ ) means there are more customers at extreme valuations. This generally favors more aggressive markdown strategies because the incremental demand from lowering prices is larger.
More periods help, but with diminishing returns. Adding markdown opportunities (more periods) always weakly improves revenue, but most of the benefit comes from the first few markdowns. Beyond 4–5 periods, the incremental gain is often small.
The salvage value sets a floor. If unsold inventory has positive salvage value, the retailer can afford to be less aggressive with markdowns. Zero salvage value forces a sell-at-any-price approach in the final period.

Extensions

The basic model captures the core tradeoffs, but practical markdown management involves several additional considerations:

Strategic Customer Behavior

Customers are not passive price-takers. Sophisticated shoppers anticipate future markdowns and may wait strategically for lower prices, even if the current price is below their WTP. This markdown cannibalization effect erodes early-period revenue: every customer who delays a purchase is a lost full-price sale. Models that account for strategic waiting typically yield more moderate markdown paths and sometimes recommend committing to a no-markdown policy or using limited-time offers to create urgency.

Price Ladder Constraints

Retailers often impose additional constraints beyond monotonicity:

Minimum markdown depth: Each price reduction must be at least a certain percentage (e.g., 20%) to be worth the operational cost of repricing.
Maximum markdown depth: A single markdown cannot exceed some threshold (e.g., 50%) to protect brand perception.
Discrete price points: Prices must fall on a predefined ladder ($79.99, $59.99, $39.99, etc.).
Coordinated markdowns: Products in the same category must be marked down simultaneously to maintain assortment consistency.

Multi-Product Interactions

When a retailer manages an assortment of related products, marking down one item can cannibalize demand for substitutes or drive traffic that creates cross-selling opportunities for complements. Full-assortment markdown optimization is a large-scale combinatorial problem that requires heuristic decomposition methods.

Demand Learning

At the start of the season, the retailer has imperfect knowledge of the demand parameters. Early-season sales data provides information that can be used to update demand forecasts (Bayesian learning) and adjust the markdown schedule. This creates an explore/exploit tension: aggressive early markdowns generate data but sacrifice margin.

Capstone: Fashion Season Manager

This capstone puts you in the role of a markdown manager for a Zara-like fast fashion retailer. In Play mode, you manage 500 units of a new collection over a 12-week season, deciding each week whether to hold prices, mark down, or run clearance. Strategic customers may wait for markdowns, creating game-theoretic tension. In Design mode, you configure the demand model and compare your strategy against optimal dynamic programming solutions and heuristic policies.

References

Aviv, Y. & Pazgal, A. (2008). “Optimal Pricing of Seasonal Products in the Presence of Forward-Looking Consumers.” Manufacturing & Service Operations Management, 10(3), 339–359.
Lazear, E. P. (1986). “Retail Pricing and Clearance Sales.” American Economic Review, 76(1), 14–32.
Pashigian, B. P. (1988). “Demand Uncertainty and Sales: A Study of Fashion and Markdown Pricing.” American Economic Review, 78(5), 936–953.
Phillips, R. L. (2021). Pricing and Revenue Optimization, 2nd ed.. Stanford University Press.
Smith, S. A. & Achabal, D. D. (1998). “Clearance Pricing and Inventory Policies for Retail Chains.” Management Science, 44(3), 285–300.
Talluri, K. T. & van Ryzin, G. J. (2004). The Theory and Practice of Revenue Management. Springer.