Why retail planning is moving beyond seasonal forecasts

For decades, retail planning revolved around seasons: buy in advance, lock in plans, and hope demand followed expectations. But between social media-driven trends, shorter product lifecycles, and rising cost pressures, that model is increasingly misaligned with reality.

What retailers are grappling with now is not simply better forecasting, but how to make planning systems responsive enough to support daily decisions without breaking financial guardrails.

“That ability to plan everything all at once with a customer dimension and all the data you need to understand that is probably best described as integrated business planning,” said Matt Hopkins, global retail lead at AI-powered enterprise planning platform Board.

While IBP is not yet common language on the retail show floor, Hopkins sees it as the logical destination for retailers trying to reconcile merchandising ambition with supply chain execution.

The limits of seasonal planning

Traditional seasonal planning assumes relative demand stability. But fashion and retail increasingly operate in an environment where demand can shift overnight.

“If they’ve got a million followers and they put it on TikTok,” Hopkins said, “there’s nothing you can do about that.”

That volatility exposes the limits of long-range planning alone. Retailers still need pre-season plans, but those plans must now coexist with short-term signals, particularly social sentiment and real-time demand sensing.

“Demand sensing is really powerful,” Hopkins said, “but it’s still a very narrow piece of optimization.”

Without a strong baseline forecast, demand sensing can amplify errors rather than correct them. “If you’ve got a really crappy forecast baseline sitting there,” he said, “demand sensing is not going to help.”

Moving from forecasts to decisions

Hopkins, speaking with Supply Chain Management Review at the recent NRF Retail Big Show in New York City, argues that retailers are beginning to treat forecasting less as an end goal and more as an input to better decisions.

“What we’re seeing now is people moving away from forecasting as a key input to a decision framework,” he said. “It’s actually saying, did that improve a decision, yes or no?”

That shift matters because retail outcomes are rarely driven by forecasts alone. Promotions, substitutions, price elasticity, and customer behavior all influence whether inventory decisions lead to margin preservation or erosion.

AI and machine learning, Hopkins said, are helping retailers break forecasts into components that can be evaluated and reused. Those insights then feed both short-term execution and longer-term planning cycles.

“You take those insights from day-to-day decisions,” he said, “and you push them back up into your longer plan cycles and they replace assumptions.”

The execution gap

Even as retailers improve forecasting and sensing, Hopkins sees a persistent bottleneck: execution systems that are too rigid to respond dynamically.

“You have all this AI going along—super sophisticated, super automated and until it hits the ERP system…and it just rounds it up,” he said.

Fixed parameters such as minimum order quantities, lead times, and replenishment rules often negate the value of upstream intelligence. The result is a disconnect between what planners know and what the supply chain can actually execute.

“If you’ve understood what customers really want in this store,” Hopkins said, “and that then goes through replenishment and it just gets rounded up, you’ve not achieved anything.”

Hyperlocal planning meets supply chain reality

Retailers increasingly differentiate through hyperlocal assortments, but that strategy carries cost and complexity.

“Hyper localization … for supply chains is difficult,” Hopkins said, “and it’s also bad for commercial planning because the costs of having local assortments is very, very high.”

Board’s approach focuses on making those tradeoffs explicit. Retailers can evaluate the cost to serve a hyperlocal assortment, assess supplier readiness, and determine whether fulfillment can be executed profitably.

“The challenge,” Hopkins said, “is attaching those super fine-tuned hyperlocal plans to execution.”

Without that connection, even the most sophisticated assortment strategies remain theoretical.

AI as an always-on planner

One of the more practical shifts Hopkins sees is how AI changes the cadence of retail decision-making. Instead of periodic reviews, AI enables continuous evaluation—what he described as “an always-on supply chain planner and always-on merchandising assistant agent.”

But automation alone is not the goal. Trust and explainability remain critical.

“You’ve got to take the merchants and buyers and supply chain planners with you,” Hopkins said. “They hold that box.”

To build that trust, leading retailers monetize decisions. Rather than presenting abstract alerts, AI systems quantify the value of acting now versus later, or not at all.

That framing turns AI from a black box into a decision partner grounded in financial outcomes, Hopkins said.

From curiosity to confidence

Hopkins believes 2026 will mark a turning point, not because AI suddenly becomes perfect, but because retailers now see how it aligns with core retail objectives.

“They can probably now see that it actually serves a very, very strong purpose,” he said, “which is customer experience.”

Rather than replacing jobs, AI augments them, helping planners, merchants, and supply chain teams operate with better information, faster feedback loops, and clearer tradeoffs.

“The outcomes are very traditional retail outcomes. Store, customer service, customer excellence,” Hopkins said.

For retailers navigating an increasingly volatile market, the path forward is less about chasing the latest AI feature and more about building decision frameworks that connect demand signals, financial guardrails, and execution, every day, not just once a season.

SC
MR