Pricing & Operations Optimization — BabyShip

BabyShip
Packaging Optimization

DIM-weight pricing, packaging, and checkout economics. Integer linear programming selects the best 7 box types for 17 top-selling strollers, balancing shipping cost, operational simplicity, and the pricing & conversion tradeoffs that come with bulky baby gear.

View Results Source Code
17 Strollers
43 Box Options
7 Optimal Boxes
6.5% Cost Penalty

Case Overview

Problem, approach, results, and business impact in one place

Problem

Shipping bulky baby gear involves tradeoffs between cost, pricing, and customer experience. Inefficient packaging and rigid pricing structures can raise landed shipping cost and push total checkout cost past thresholds where conversion drops sharply — for many e-commerce categories, orders above roughly $100 all-in see materially lower completion rates.

Technical layer: Carriers bill on dimensional weight when it exceeds actual weight (here: L×W×H / 250). Warehouses also need a limited box catalog, not one SKU per product.

Approach

  • Modeled shipping cost using dimensional-weight pricing: billable weight = max(actual weight, DIM weight).
  • Built an integer programming framework (Gurobi) to choose box assignments under fit rules, a 43-option catalog, and a cap of 7 active box types.
  • Pricing sensitivity: framed how shipping-weight penalties flow into published shipping fees and free-shipping thresholds; extension work uses Python (and in production, PostgreSQL-backed order/shipping extracts) to stress-test thresholds against real usage.
  • Compared the optimized 7-box system to a “perfect fit” baseline (best possible box per product) to quantify inefficiency from standardization.

Results

  • Flagged pricing regions where high DIM-driven shipping pushes all-in cost above ~$100, a band that typically hurts conversion unless offset by subsidies or product mix.
  • Quantified inefficiency from limited box sizes: only 6.5% total billable-weight penalty (104.7 lb across 17 SKUs) vs. perfect fit, while cutting box variety from 17 → 7 (59% fewer packaging SKUs).
  • Two outlier products (Vista V2, Fox 5) drive ~30% of total shipping weight via a 40×40×40 box — clear candidates for packaging redesign, surcharges, or threshold tuning.
  • Delivered a repeatable framework to evaluate operational simplicity vs. cost whenever leadership changes box policy or carrier DIM divisors.

Business Impact

  • Actionable recommendations for pricing strategy (shipping fees, free-shipping bar) and packaging decisions (which boxes to stock, which SKUs to redesign).
  • Surfaced ways to reduce landed cost while preserving customer experience — e.g. concentrating fixes on the two heaviest DIM outliers instead of blanket cuts.
  • Showed how operational constraints (max box types, carrier DIM rules) directly shape revenue outcomes through conversion and margin, not only through freight invoices.

Key insight

Optimizing pricing and packaging is not just a modeling exercise. It requires balancing carrier cost structures, customer behavior at checkout, and operational constraints so decisions stay credible in finance, ops, and product.

Why Box Selection Drives Cost

Dimensional weight, standardization, and the cost–complexity tradeoff

Dimensional Weight

Carriers charge the greater of actual weight and DIM weight (L×W×H / 250). Oversized boxes inflate billable weight even when the product is light — the core lever for shipping cost and pricing.

Box Standardization

Using a unique box per product is impractical. Warehouses need a limited set of box sizes that can cover all products efficiently.

The Trade-off

Fewer box types mean simpler picking and inventory, but often higher DIM penalty. The optimization finds the best 7-box portfolio that minimizes total billable weight subject to fit and operational limits.

Data Overview

17 popular US strollers and 43 candidate shipping boxes

ProductL (in)W (in)H (in)Weight (lb)
Triv Next2422.51518.4
Vista V217.325.73327
Cruz V216.522.829.525.5
Fox 335211521.8
Fox 517.323.635.421.1
Fox 5 (2 pieces)3520.913.428
Mixx Next2723.61928.3
City Mini GT23125.510.522.4
Single-to-Double3424.51827
Portable Playard10.3110.3127.9914
Close2Baby30181030
Breeze Plus Playard12123023
Swivel Sleeper Luxe34342332
Smart Sleeper382215.538
Lullaby Playard10.5193025
Travel Crib Light523.51913
Lotus Travel Crib2412814
Box IDL (in)W (in)H (in)Dim Weight (lb)

Stroller Dimensions Comparison

Model Formulation

Integer Linear Program solved with Gurobi optimizer

Decision Variables

xij ∈ {0, 1} — assign stroller j to box i

yi ∈ {0, 1} — use box type i in the system

Objective

Minimize Σi,j costij · xij

where cost = max(dimensional weight, actual weight)

Constraints

1. Each stroller assigned exactly one box

2. Assignment only if box type is selected

3. Stroller must physically fit in the box

4. At most 7 box types total

Fit Criteria

A stroller fits a box when all three dimensions (with 0.5″ padding) are ≤ the corresponding box dimensions.

Only 731 of 731 possible pairs actually fit

774Binary Variables
1,394Constraints
0.02sSolve Time
0.00%Optimality Gap

Optimal Results

7 box types selected to cover all 17 products

Selected Box Types

Box 6

15×15×48

DW: 43.2 lb

2 products

Box 8

40×40×40

DW: 256.0 lb

2 products

S-16765

36×24×18

DW: 62.2 lb

5 products

S-4478

36×36×24

DW: 124.4 lb

3 products

S-4659

24×24×36

DW: 82.9 lb

2 products

S-4866

28×28×20

DW: 62.7 lb

2 products

S-16766

40×30×30

DW: 144.0 lb

1 product

Product → Box Assignments

ProductAssigned BoxBox DimensionsShip Weight (lb)
Triv NextS-1676536×24×1862.2
Vista V2Box 840×40×40256.0
Cruz V2S-465924×24×3682.9
Fox 3S-1676536×24×1862.2
Fox 5Box 840×40×40256.0
Fox 5 (2 pieces)S-1676536×24×1862.2
Mixx NextS-486628×28×2062.7
City Mini GT2S-447836×36×24124.4
Single-to-DoubleS-447836×36×24124.4
Portable PlayardBox 615×15×4843.2
Close2BabyS-1676536×24×1862.2
Breeze Plus PlayardBox 615×15×4843.2
Swivel Sleeper LuxeS-447836×36×24124.4
Smart SleeperS-1676640×30×30144.0
Lullaby PlayardS-465924×24×3682.9
Travel Crib LightS-486628×28×2062.7
Lotus Travel CribS-1676536×24×1862.2

Cost Analysis

Comparing the optimized 7-box system against a perfect per-product baseline

1,718.0 Optimized Total (lb) 7-box system
1,613.3 Perfect Total (lb) Unique box per product
104.7 Total Penalty (lb) Only 6.5% increase

Shipping Weight: Optimized vs. Perfect

Extra Cost per Product

Box Utilization — How Many Products per Box Type

Dimensional Weight vs. Actual Weight by Assignment

Technical takeaways

  • S-16765 (36×24×18) is the workhorse box, covering 5 of 17 products efficiently.
  • Vista V2 and Fox 5 are the costliest to ship — they force Box 8 (40×40×40) and drive ~30% of total billable weight; prioritize these for packaging or pricing levers.
  • The 7-box system adds only 6.5% extra billable weight vs. a perfect per-SKU box — a strong ops simplicity trade for finance.
  • Most standardization penalty sits on two SKUs; the other fifteen pack near the theoretical minimum under the shared catalog.