Key takeaways
- Forecasting becomes harder as soon as ecommerce businesses add more channels, locations, and SKUs.
- Traditional forecasting methods were not designed for frequent sales events, omnichannel fulfillment, or product variants.
- As complexity increases, forecasting errors compound faster and become harder to correct.
- Ecommerce businesses can respond by acknowledging forecasting as a structural challenge and using technology designed to absorb complexity
You’re not imagining it. Inventory forecasting really has become more difficult for ecommerce leaders over the past few years.
The challenge is not that teams are less capable or that demand is more unpredictable than ever before. The real issue is that ecommerce businesses now operate with far more complexity than the forecasting methods they still rely on.
As channels multiply, fulfillment models expand, and sales events happen year-round, even well-run businesses begin to feel strain. Forecasting does not fail all at once. It degrades quietly, then suddenly shows up as missed revenue, excess stock, or margin erosion.
Table of contents
- How ecommerce complexity has changed forecasting
- Why this matters for ecommerce businesses right now
- What we are hearing from the market
- What business leaders can do about it
- How to get started
- FAQs
How ecommerce complexity has changed forecasting
Forecasting errors are not new. What has changed is how often businesses are exposed to them. A decade ago, many ecommerce businesses sold through a single website or marketplace. Inventory lived in one or two locations, and promotions followed a predictable annual calendar.
Today, that model is the exception, not the rule. Brands sell across marketplaces, direct-to-consumer sites, and physical retail locations. Inventory may sit in multiple warehouses, stores, or third-party fulfillment networks.
According to Zaeem Batavia, VP of Global Sales for Descartes Ecommerce, this shift has fundamentally changed how businesses must approach forecasting.
“We’ve moved to an omnichannel type of world where we’ll sell 70% of our product online, but then we have maybe 30% in multiple retail stores dotted around the country,” he explains. “How do we ensure that we have the right inventory in the right locations?”
Layer in product variants like size, color, and style, along with seasonality and returns, and forecasting quickly becomes a multidimensional problem. Traditional tools were never designed to handle this level of complexity.
Why this matters for ecommerce businesses right now
Many ecommerce brands now run multiple promotional events each year rather than relying on a single peak season. Each event creates a new forecasting decision, and each decision carries risk.
“What a lot of our customers have [historically] done is just ordered more than they actually need to ensure they have enough to sell,” Batavia says. However, that approach is becoming increasingly costly, especially in a slower ecommerce market.
After peak periods, excess inventory often leads to heavy discounting. Margins shrink, brand value erodes, and capital remains tied up in unsold stock. At the same time, underestimating demand can mean stockouts during critical sales windows.
When forecasting struggles at this level, it becomes more than an operational inconvenience. It affects profitability, customer experience, and brand survival.
What we are hearing from the market
As forecasting grows more complex, many ecommerce businesses are discovering that their internal processes no longer scale.
Batavia points out, “Manual workarounds, people working in an office trying to figure out how much to purchase, just isn’t scaling with the demand volume that they’re seeing today.”
He notes that while teams have invested heavily in automating warehouse workflows and order processing, forecasting has quietly lagged behind.
“Customers have found that they’ve woken up, and they’ve gone like, ‘Our back-office headcount has just grown significantly,’” he says.
Often, that bloated back-office growth is driven by hiring more and more people to manage manual planning workflows in spreadsheets.
“[Leaders] are realizing they actually don’t know what people are doing,” Batavia explains. And the manual process leaves ample room for human errors.
At that point, forecasting stops being a straightforward, tactical task. Instead, it becomes an overcomplicated constraint on the business.
What business leaders can do about it
Ecommerce leaders who successfully navigate forecasting complexity tend to take a similar approach:
- Stop trying to solve forecasting with spreadsheets. First, they acknowledge that forecasting challenges are structural, not personal. More spreadsheets or more people rarely solve the underlying problem.
- Identify areas where better visibility and control are needed. Next, they examine how many variables their forecasting process must account for, including channels, locations, seasonality, returns, and fulfillment options.
- Leverage technology to automate complex forecasting. Finally, they implement technology-based solutions to handle complexity and automate decision-making, rather than continuing to rely on inefficient processes.
How to get started
Does forecasting feel harder than it used to? If so, your business has likely outgrown its current approach.
To address this, start by mapping the areas of complexity in your forecasting process today. Then, start exploring solutions.
When you’re ready to see options for inventory forecasting tools, Descartes is here to help you find the right solution.
FAQs
Growth introduces more channels, products, locations, and sales events. Each variable adds complexity that traditional forecasting methods struggle to manage.
Omnichannel selling requires inventory to be positioned across multiple locations. Forecasting must account for where demand will occur, not just how much demand exists.
Overbuying ties up capital and often leads to markdowns after peak periods, reducing profitability and increasing financial risk.
If forecasting requires significant manual effort, frequent adjustments, or creates recurring stockouts or excess inventory, it may be time to reassess the approach.