Stocking Retail Assortments Under Dynamic Consumer Substitution

Stocking Retail Assortments Under Dynamic Consumer Substitution Motivation Sample path with T sequential customers Retailer RETAILER Customer t with preferences Ut=(Ut0, Ut1,…, Utn) Category with n substitutable variants Motivation • Retail consumers might substitute if their initial choice is out of stock – Retailer’s inventory decisions should account for substitution effect • Consumers' final choice depends on what he/she sees available “on the shelf”. – In most previous models demand is independent of inventory levels. • Contribution of this paper: – Determination of initial inventory levels (single­ period) taking into account dynamic substitution effects Outline • Brief Literature Review • Model Formulation • Structural Properties – Component-wise Sales Function – Total Profit Function – Continuous Model • • • • Optimizing Assortment Inventories Numerical Experiments Price and Scale Effects Conclusions Brief Literature Review • Demand models with static substitution: – Smith and Agrawal (2000) / van Ryzin and Mahajan (1999): 1. First choice is independent of stock levels 2. If first choice is out of stock, the sale is lost (no second choice) 3. Consequently, demand is independent of inventory levels • Papers that model dynamic effects of stockouts on consumer behavior – – Anupindi et al. (1997): concerned solely with estimation problems (no inventory decisions) Noonan (1995): only one substitution attempt Model Description • Assumptions: –Sequence of customers is finite w.p.1 –Each customer makes a unique choice w.p.1 • Some special cases: -Multinomial Logit (MNL): -Markovian Second Choice: U t j = u j + ξt j q j = P(U t[1] = U t j ) P(U t[2] = U tk | U t[1] = U t j ) = p kj U t0 > U t[3] > … > U t[ n ] -Universal Backup: all customers have an identical second choice -Lancaster demand: attribute space [0,1] and customer t has a random “ideal point” Lt, then U t j = a − b Lt − l j Model Formulation • Profit function: – hj(x,w) = sales of variant j given x and w. – Individual profit: pj(x ,w) = pj hj(x,w) – cjxj. – Total profit: p(x ,w) = S pj(x ,w). • Retailer’s objective: max E[π ( x, ω )] x≥0 • Recursive formulation: – System function: f ( xt , U t ) = xt − e d ( xt ,U t ) = xt +1 – Sales-to-go: ηt j ( xt , ω ) = xtj − f j ( xt ,U t ) + ηt j+1 ( xt +1 , ω ) – Border conditions: ηTj +1 ( xT +1 , ω ) = 0 x1 = x Structural Properties • Lemma 1: – x¥y ï xt¥yt for all sample paths. • Decreasing Differences: – h:SµQö√ satisfies decreasing differences in (z,q) if h(z’,q) – h(z,q) ¥ h(z’,q’) – h(z,q’) for all z’¥ z , q’¥ q. – Lemma: if max{h(z,q): zœS} has at least one solution for every q œ Q, then the largest maximizer z*(q) is nonincreasing in q. Structural Properties • Theorem 1: The function hj(z·ej+ q,w) satisfies: (a) Concavity in z for all w. (b) Decreasing differences in (z,q) for all w. • Corollary 1: (a) A base-stock level is optimal for maximizing the component-wise profits. (b) The component-wise optimal base-stock level for j is nonincreasing in xi (i≠j). ï Usual newsboy problem Structural Properties • Let: –T(x ,w)=∑ hj(x,w) = total sales –Hj(z ,w)= T(z·ej+ q,w) • If all variants have identical price and cost, then: p(x ,w) = p·T(x ,w) – c·∑ xj • Theorem 2: There exists initial inventory levels x and sample paths w for which: (a) T(x ,w) is not component-wise concave in x. (b) Hj(z ,w) does not satisfy decreasing differences in (z,q). Structural Properties • Counterexample for (a): • Continuous model: –Customer t requires a quantity Qt of fluid with distribution Ft(·) –Redefine sample paths: w = {(Ut, Qt): t=1,…,T} • Theorem 3: There exist sample paths on which p(x ,w) is not quasi-concave Optimizing Assortment Inventories • Lemma 3: If the purchase quantities Qt are bounded continuous random variables then ∇E[h(x,w)] = E[∇h(x,w)] Numerical Experiments • Heuristic policies (with T~Poisson) Let qj(S) be the probability of choosing variant j from S 1. Independent Newsboy: demand for each variant is independent of stock on hand. xIj = λ q j ( S ) + z j 2λ q j ( S ) j∈S 2. Pooled Newsboy: customers freely substitute among all available variants. q(S ) = ∑ q j (S ) j∈S q j (S ) x = x( S ) q( S ) j P x ( S ) = λ q ( S ) + z 2λ q ( S ) Numerical Experiments • Assumptions: -T~Poisson(30). -Qt~exp(1). • Example 1: –Assume MNL utilities: υj q ( S ) = P(U t = max{U : i ∈ S}) = ∑υ i + υ 0 j j i t i∈S uj /µ ⎧ e ⎪ where υ j = ⎨ u / µ ⎪⎩e 0 j∈S j=0 –Equal cost and prices. –Result from van Ryzin and Mahajan: assume v1≥v2≥…≥vn, the optimal assortment is a set Ak={1,2,…,k} with k=1,…,n. –Simulation: profit within ±1% with 95% confidence. –Several starting points tested. Numerical Experiments • Performance Comparison for Example 1. -Independent newsboy is biased: underestimates popular items, over estimates less popular items. Numerical Experiments • Example 2: –Two variants: variant 1 less popular but with high margin, and variant 2 more popular but lower margin. –Sample Path Gradient policy induces customers to “upgrade” to the highmargin variant. Numerical Experiments • Example 3: –Lancaster demand model: U t = a − b Lt − l j –Lt~U[0,1], a=0.2, b=1 Price and Scale Effects • Measure of “evenness”: – y is more “fashionable” than z, if z majorizes y: n ∑y i =1 k n [i ] = ∑ z[i ] i =1 k ∑ y[i] ≤ ∑ z i =1 [i ] k = 1,… , n − 1 i =1 • Observations: 1) If v is more fashionable than w, then w is more profitable 2) Price or volume increase ï higher variety is offered Conclusions • Concluding remarks – General choice model. – Improvement upon the existing literature. – Interesting numerical experiments with valuable insights. • Comments – Assortment or inventory problem? – Assumes full information. – Single replenishment: price decision might be relevant. – Industry evidence (field study).