Did you know that Amazon earns more than one-fifth of its North America retail revenue because local stores can’t forecast accurately?

Did you know that Amazon earns more than one-fifth of its North America retail revenue because local stores can’t forecast accurately? Customers try to purchase the product at a store in these scenarios, but the stores are out-of-stock and so shoppers look to Amazon.

Clearly, forecasting essential, but we should note that it’s more than just predicting demand for your products. Demand forecasting also helps businesses effectively manage cash flow and maintain lean operations. If you’re carrying extra stock or don’t have enough to meet demand, you’re losing money.

When working with one large retailer, Harve Light, managing director at Conway MacKenzie, and team learned that a 10% increase in forecast accuracy could increase profitability by more than $10 million. And while not ALL retailers have the same opportunity, neglecting to forecast could be detrimental to your business. One study found that retailers lost $1.75 trillion to overstocks and out-of-stocks in a single year.

“This is especially relevant if you’re working with an outside manufacturer,” says Abby Perkins, director of content and communications at Glew.io. “You need to know [when] to reorder your product, and in what quantity, before you sell out.”

Not sure where to begin? We’ve put together your demand forecasting 101 guide to help you find the optimal stock levels.

Related:

• The Ultimate Inventory Management Resource Guide: Everything You Need to Know About Stock Control
• 8 Inventory Management Techniques to Help You Stay on Top of Stock Control
• Avoiding Stockouts: 5 Preventable Causes of Out-of-Stocks and What to Do About Them

What is demand forecasting?

In retail, demand forecasting is the practice of predicting which and how many products customers will buy over a specific period of time. Demand forecasting is typically done using historical data (if available) as well as external insights (i.e. the weather, consumer trends, etc.).

Forecasting helps retailers understand when they need to order new merchandise, and how much they’ll need to get. When determining this timeframe, you’ll need to consider the necessary lead time to help inform your reorder point.

How to forecast inventory demand

Understanding how to forecast inventory demand can be intimidating at first, and for good reason. It can be a complicated process, and it’s difficult to get it right. But the proper tools and approach, you can make the process much easier.

“Retail demand forecasting is one of the hardest analyses to get right: Forecast too little and you have empty shelves, and forecast too much and you have inventory gluts to work through,” says Carlos Castelán, managing director of The Navio Group, a retail consulting firm that’s worked with Whole Foods, CVS and Kraft Heinz. “It’s a mix of both art and science.”

It also depends on the size and type of retailer, says Light. “Large retailers have entire industries that help them improve their forecasting methods. Small retailers use basic spreadsheets,” he says.

Forecasting methods and techniques

There are several forecasting methods and techniques, some of which can be used simultaneously. Mainly, though, forecasting can be broken down into four main types:

• Qualitative

• Time series analysis

• Causal

• Simulation

Qualitative forecasting: AccountingTools.com defines qualitative demand forecasting as follows: “Qualitative forecasting is an estimation methodology that uses expert judgment, rather than numerical analysis. This type of forecasting relies upon the knowledge of highly experienced employees and consultants to provide insights into future outcomes.” Rather than using historical data alone, as in a quantitative approach, qualitative forecasting accounts for different factors that will impact future demand.

“Many retailers and brands adjust stock levels and orders based on the previous year’s output and sales,” says Marc Gingras, CEO of Foko Retail. “They often focus on data that’s readily apparent while ignoring what’s less quantifiable. That’s fine if you’re a small-to-mid-sized retailer just trying to stay afloat, but not if you want to be the next big name in retail.

Time series analysis: The time series analysis for demand forecasting skews closer to the quantitative approach. Towards Data Science says, “Time series analysis comprises methods for analyzing time series data in order to extract meaningful statistics and other characteristics of the data. Time series forecasting is the use of a model to predict future values based on previously observed values.”

“Data cannot be ignored.”

– Marc Gingras, CEO, Foko Retail

 

“Retailers should use an analytical approach, examining sales channels, suppliers and the demand placed on both, to accurately predict inventory needs,” says Gingras.

Causal: Causal forecasting pays special attention to the relationship between different events or variables. The weather is a big one, for example. The official definition for causal forecasting, according to BusinessDictionary.com, is: “Estimating techniques based on the assumption that the variable to be forecast (dependent variable) has cause-and-effect relationship with one or more other (independent) variables.”

Examining causal relationships helps you forecast more accurately because you can predict and account for external factors that affect demand. Light likes to categorize these as complements and cannibalization. “When a retailer puts dress shirts on sale, they will likely experience some increase in the sale of t-shirts. These are complements,” he says. “When a retailer puts one brand of t-shirts on sale, the other brands carried will suffer a decline in sales. This is cannibalization.” Remember to account for everything that’s happening in your store (and online!)

Simulation: Simulation forecasting is the approach where all methods are mixed together. It accounts for both qualitative and quantitative insights to provide a more holistic outlook. However, this is also arguably the most complicated forecasting technique to DIY, because of its complicated nature. Simulation also accounts for internal and external factors — those elements identified in your causal forecasting.

Demand forecasting tips

Establish a baseline for data

If you’re new to forecasting, one of the first things you’ll want to do is establish a baseline. Without data, it’s difficult to make informed forecasting decisions and predictions.

“The simplest way to build a forecast is to pull in sales from the year prior and then factor in the growth rate for your business year to date to get a baseline of what to expect,” says Joanna Keating, head of marketing and ecommerce at United By Blue, which operates three brick-and-mortar locations in New York and Philadelphia. “It also helps to plan your sales by the day, which allows you to react quickly if something doesn’t meet your expectations.”

Understand your customer and local market

“To effectively forecast demand, it’s most important to understand your customer well and their shopping tendencies,” says Castelán. Some questions to ask:

• Do my customers shop seasonally or is it consistent year round?
• What sizes and/or colors do my customers prefer?
• Are shoppers partial to certain brands?
• What do shoppers in my local area like?
• How quickly do trends catch on with consumers in my store’s area?

Lilly Pulitzer, for example, is very popular in the southeastern U.S. Compare that to an outdoor brand like Smartwool, which reigns supreme in the western states of Montana, Colorado and even Alaska.

Lily Pulitzer on Google Trends

Smartwool on Google Trends

You likely already have lots of this data, much of which can be captured through your point-of-sale (POS) terminal. “All of this information can be gathered through a past sales analysis,” says Castelán. “It’s helpful to have strong product attributes or product information management (PIM) to analyze performance in relation to product attributes as well as through customer data points.”

Analyze your KPIs

To analyze against your baseline, there are a few key metrics to track. “One of the key metrics of the forecasting process is sell-through rate, which is the percentage of non-clearance items that you will sell in relation to on-hand product for a given time period,” says Castelán.

“To identify the right sell-through rate and forecast demand, retailers often work collaboratively with suppliers to forecast demand (and their purchases) based on market information they might have along with promotional plans,” he says. “Work with suppliers to develop contingency plans [if your predictions are inaccurate].”

Remember external factors

We touched on this when discussing causal relationships in forecasting demand, but it’s so important that we’re stressing it: Remember to always consider external factors. While you know your own marketing and promotions schedule, plus the annual busy selling seasons during the holidays, there are other things you can’t predict or control.

“A big challenge is unknown events,” says Perkins. “You can have an accurate forecast that gets totally thrown off by something like a viral event in your industry, a related product launch or innovation, or even a weather event. Being nimble and able to adapt to unknown events is key.” That’s where the contingency plans come into play. Be prepared for the “If X happens, then Y product will be in demand” scenario.

Keating at United By Blue also advises having a plan, as well as adopting a more cautious approach to forecasting. “All areas of the business benefit from having a plan in place,” she says. “I always suggest to err on the conservative side to ensure all teams have the resources they need to handle a high sales period.”

Leverage technology and automation to increase accuracy

15% of inventory distortion issues happen because software can’t talk to each other. And when we don’t use tech, we make ourselves more susceptible to data discrepancies caused by human error. “Today, there are also several scaled-down versions of tools that the large retailers use available to smaller retailers at more reasonable costs,” says Light. With technology being so accessible, there’s no reason not to take advantage of it.

There are two key goals to building a tech stack ecosystem that facilitates forecasting and other inventory management-related processes:

1. Centralize your data: Centralized data is a fancy term for having all of your metrics housed and accessed in a single location. This is especially helpful for retailers with multiple locations and/or team members — that way, everyone is looking at the same information and making decisions based off the same numbers.

“One of the biggest challenges retailers face when it comes to forecasting is having to look for data in multiple places,” says Perkins. “It requires more manual effort and leaves a lot of room for human error.” When you leverage tools and tech to centralize the information, you know the data is accurate, formatted consistently and calculated in the same way across the board.

Perkins’ advice? “Get a reporting platform that houses all your data — ecommerce, POS, marketing, shipping, etc.,” says Perkins. “It makes it a lot easier to forecast accurately, and keep track of key metrics like sell-through rate that help with forecasting.”

2. Automate processes and workflows: Another way to reduce human error and preserve the validity of your data is through automations. “Use tools that have automation and alerts to keep you updated about products that are about to sell out (or not selling as quickly as expected) so you can adjust your forecast accordingly,” says Perkins. “We have one customer who uses automated alerts to let him know any time a product is within 60 days of selling out, since it takes 60 days to get his product back in stock.”

LowCarb Canada operates two brick-and-mortar locations and two online stores. At more than 2,000 SKUs, forecasting was a tedious and time-consuming process that they used to do manually. When they upgraded their technology, they used automated sales velocity reports to stay on top of stock levels and forecasting.

Over to you

Have you begun basic forecasting for your retail business? What are your biggest challenges when it comes to forecasting demand accurately? What advice do you have for others?