Demand Forecasting

Demand Forecasting for Small Apparel Brands

By Bellamy Grindl · Retailytics ·

Demand forecasting for a small apparel brand is the process of estimating future sales by style, category, and channel — so buying decisions are made against a projection rather than a gut feel. For most brands, the forecasting process evolves in three stages: the founder makes all purchasing decisions, someone close to the numbers takes it on, and eventually a dedicated process or person owns it.

Buying your first purchase order and hoping it sells is a perfectly reasonable approach when you have one product and one channel. The problem is that most brands keep that same approach as the assortment grows to 80 styles across three channels. What worked at $300K quietly stops working at $2M — and by the time the cracks show, they show up in cash flow.

Why Demand Forecasting Gets Harder as You Grow

Replenishment, new season buying, channel complexity, and assortment depth all compound the difficulty of forecasting as a brand scales. The mental model that worked for five styles doesn't scale to two hundred.

What changes as a brand grows: replenishment decisions require tracking velocity and lead times, not just watching sell-through. New season buying means projecting without history on specific styles. Each channel added creates different velocity patterns, margin profiles, and fulfillment timelines for the same inventory pool.

Most founders reach a point where the complexity has outgrown the time they can give to inventory decisions. Because they have a hundred other jobs, the forecasting process stays exactly as informal as it was on day one.

The 35-Variable Problem

A well-run demand forecast considers far more inputs than most founders realize — lead times, sell-through by style, size curves, weeks of supply, channel allocation, promotional calendar, year-over-year data, return rates, markdown timing, and more.

If inventory planning is not your dedicated job, you are almost certainly not considering all of those variables. You are buying the new and replenishing what sold, and calling that a plan.

The consequences show up later. Wrong inventory decisions do not announce themselves immediately. They show up in cash flow months later, in a markdown season that underperforms, in a missed launch because cash was locked in the wrong stock. Costly decisions compound over time when the planning process doesn't keep pace with the business.

How to Build a Demand Forecasting Process at Your Stage

Demand forecasting evolves in three stages: under $1M (document your assumptions), $1M–$5M (build the weekly tracking habit), $5M+ (build the process around a person).

Stage 1: Under $1M — Start Simple, Start Now

The goal is not a sophisticated forecast. It is a documented assumption. Before every buy, write down what you expect to sell by month, which styles you expect to be strongest, and how you're allocating budget between proven styles and new ones. That document is your forecast. It doesn't have to be right — it has to exist so you can compare it to what actually happened.

Stage 2: $1M–$5M — Build the Tracking Habit

At this stage you have enough history to forecast from data rather than just assumptions. Track weekly sales by category and top styles vs. last year, sell-through rate by style, weeks of supply on reorder candidates, and remaining open-to-buy for the season. A consistent weekly review — even 20 minutes — changes how buying decisions get made. Patterns become visible. Winners surface early. Slow styles get caught before they age into dead stock.

Stage 3: $5M+ — Build the Process Around a Person

At this stage the complexity requires someone whose primary job is inventory planning. Monthly OTB by category and channel, seasonal buys planned against a demand plan, reorders triggered by data rather than intuition. This is also the stage where AI-powered planning tools become worth evaluating — but only once the process and data governance are in place to support them.

The Most Common Demand Forecasting Mistakes in Apparel

The most costly mistakes: forecasting revenue instead of units, using only sales as an input, buying flat year-over-year, ignoring the promotional calendar, and planning new styles like proven ones.

Using only sales as an input. If inventory ran low or out of stock during a period, your sales data is artificially suppressed. Forecasting from that data means you will under-order because the demand was higher than the sales showed. Track inventory availability alongside sales to see the full picture.

Buying flat year-over-year. If a style was at 90% sell-through last season and you bought the same quantity, you're likely to stock out. If it was at 40%, you overbought the first time. Adjust quantities based on what the data says, not what the PO said.

Not accounting for the promotional calendar. A sale event will pull forward demand that would otherwise happen later. Forecast with the promo calendar in view, not around it.

Client Work — $20M Fashion Apparel Brand

A $20M fashion apparel brand engaged Retailytics over a nearly three-year engagement to build planning infrastructure across a growing team and multi-channel business. Retailytics implemented a data hierarchy and reporting system, established inventory benchmarks, reduced aged inventory penetration by 20 percentage points, and grew the Faire wholesale channel 25% in units year-over-year. As headcount scaled, Retailytics trained and onboarded new team members across planning, merchandising, and sales. The engagement culminated in a full S&OP process tailored to the brand's needs — enabling integrated planning across product development, financial planning, and inventory lifecycle management for the first time.

Want to build the full forecasting and planning system for your brand?

The live cohort walks through demand forecasting, OTB, and in-season management in four weeks using your actual data. View the Cohort →

Frequently Asked Questions

What is demand forecasting in apparel retail?

Demand forecasting in apparel is the process of estimating future unit sales by style, category, and channel — so buying decisions are made against a projection rather than a guess. It typically includes analysis of historical sell-through, seasonal patterns, promotional calendars, and channel mix.

How do small apparel brands forecast demand without historical data?

For new brands or new styles without sell-through history, forecasting relies on comparable style performance, category benchmarks, planned marketing activity, and conservative unit assumptions. The key is making the assumption explicit — written down before the buy — so you can compare it to what actually happens and improve the next forecast.

When should a small apparel brand invest in demand forecasting software?

Demand forecasting software makes sense when the brand has clean, consistent data across sales and inventory, an internal team member who understands retail planning fundamentals, and a process the software will support rather than replace. For most brands under $3M, a structured spreadsheet process is sufficient.

What data does an apparel brand need to forecast demand?

The most useful inputs are weekly sales by style and variant, current on-hand inventory by variant, historical sell-through rates, planned marketing and promotional activity, and open purchase orders with expected receipt dates. Shopify provides most of this through its reporting and inventory tools.

How far in advance should an apparel brand plan buying?

Most apparel brands plan seasonal buys 4–6 months in advance, given standard lead times of 3–5 months from order to delivery. Monthly reorder decisions happen on a shorter cycle based on in-season velocity and weeks of supply.

About the Author

Bellamy Grindl is the founder of Retailytics. She has built demand forecasting and inventory planning systems at brands across every revenue stage. Book a free 15-minute call to talk through where your brand is in this process.