# Hershey Uses AI to Forecast S'mores Season, Lifts Co-Brand Placement and Inventory Timing

*Chocolate maker deploys machine learning to optimize seasonal bundling and retail placement for three-SKU category.*

By **Jenny Huang Goodman MPA MSc MHSA, Principal** — The Stash Edge, Hako Shikin LLC.
Published 2026-08-24.

Canonical: https://www.pops4.com/stash/articles/hershey-2026-08-24t15-5
Subject: Hershey
Tags: ai forecasting, seasonal inventory, co-brand placement, category coordination, retail timing

---

The Hershey Company deployed AI forecasting tools to time and place s'mores ingredients—its own chocolate bars, plus marshmallows and graham crackers from partner brands—inside retail stores, according to Modern Retail. The move addresses a narrow seasonal window: s'mores purchases spike in late spring and summer, then vanish. Miss the window, and you carry dead inventory. Launch early, and you train retailers to ignore the category.

Hershey used machine learning to analyze regional weather patterns, campground occupancy data, and prior-year sales velocity by store. The system generates SKU-specific recommendations: which stores stock which pack sizes, when end-cap displays go live, and which co-brand marshmallow or cracker gets bundled at checkout. The AI layer runs continuously, adjusting projections as Memorial Day approaches or a cold snap delays camping season in the Midwest.

The mechanism is elastic seasonality married to three-party coordination. S'mores require three distinct product categories, often from competing suppliers. Hershey cannot control cracker or marshmallow replenishment, but it can use its own forecast to negotiate co-branded displays and shared margin with those suppliers. The AI tool gives Hershey a datapoint retailers trust: "Your region will see peak velocity in week 22, and this store format converts best with a checkout impulse bundle." That precision earns placement.

The underlying pattern works for any brand selling into a predictable seasonal spike with complementary SKUs outside their catalog. A hot sauce brand forecasting peak grilling season can approach chip and beer brands with co-promotion timing. A candle maker tracking holiday gifting curves can bundle with card and wrap suppliers. The AI removes guesswork and turns the conversation from "maybe we should team up" to "here is the exact week and store format where joint placement lifts both SKUs."

A small physical-product brand replicates this without enterprise AI by building a simple three-column spreadsheet: date, event trigger, and complementary product. If you sell cocktail bitters, your triggers are Derby Day, Cinco de Mayo, Father's Day. Complementary products are premium mixers, glassware, garnish kits. Two months before each trigger, you email those brands with a one-page PDF: "Our sales data shows [date range] converts at [X] percent above baseline. We propose a bundled checkout display at [retailer]. Split the slotting fee, share the margin." You are offering the other brand a forecast they do not have, which makes you the coordinator.

For solo founders, start with one regional grocer and one holiday. Pull your own prior-year sales by week, chart it, and identify the two-week spike. Approach one complementary brand—ideally a product already on that grocer's shelf—with the chart and a proposal: joint endcap, even split on costs, you handle the retailer conversation. Cost: zero beyond your time. The chart is your credential.

The broader implication is that AI-driven seasonal forecasting is now table stakes for any brand competing in categories with weather, holiday, or event triggers. Retailers expect you to arrive with timing data, not a hopeful pitch. If you sell a product that pairs with something you do not make, you either coordinate the category or watch a competitor do it first.

## The takeaway

Use historical sales spikes and event timing to approach complementary brands with co-promotion forecasts retailers cannot ignore.

---

## Publisher

**Hako Shikin LLC** — Virginia Beach, Virginia. Founded 1997. ASI 217876 · DUNS 18-204-6339.
Principal and author: **Jenny Huang Goodman MPA MSc MHSA**.

- Author: https://www.huanggoodman.com/about
- LLM context: https://www.pops4.com/stash/llms.txt
- MCP endpoint, for AI agents: https://mcp.pops4.com/mcp
- Client dashboard: https://dashboard.pops4.com/
- Catalogue: 70,000+ products, 200+ brands