Inventory Forecasting: Methods, Step-by Step Guide, and Best Practices For Success
If only you had a crystal ball that could tell you the perfect stock quantities to carry, running an ecommerce business would be a walk in the park.
Instead, your best-selling items are out-of-stock right before the holidays, yet your warehouse is brimming with products sleeping on its dusty shelves.
The problem is clear—you need a better approach to inventory forecasting.
Need expert help selecting the right tool?
With one-on-one help, we guide you to your top software options. Narrow down your software search & make a confident choice.
The solution is here. In this blog post, you’ll find a step-by-step guide and a list of best practices that will help you do just that.
What is Inventory Forecasting?
Inventory forecasting involves anticipating and optimizing stock levels based on customer demand over a specific period. It's a continuous process and is also commonly referred to as demand planning.
Here's why ecommerce businesses like yours can benefit from having more accurate predictions:
Better inventory forecasting gives you greater inventory control.
With just enough stock, you're less likely to encounter stockouts and overstock situations, which are the very culprits that cost retail businesses a projected loss of $1.77 trillion in 2023.
Keeping your inventory forecast as accurate as possible prevents you from tying up capital or compromising customer experience.
Inventory Forecasting Challenges
Accurately predicting inventory levels can save your business a fortune, but it’s not without challenges.
The most common ones are:
- Dealing with uncertain consumer demand. Past sales data alone isn’t enough to predict optimal inventory levels, because they don’t account for uncertainty.
- Managing seasonal trends. Seasonal forecasting models require accounting for adjustments like sales fluctuations and promotional effects.
- Integrating forecasting with supply chain management. Supplier delays, economic instability, and other disruptions trigger supply uncertainty. That’s why 51% of businesses plan to collaborate with suppliers to forecast demand accurately.
- Ensuring data accuracy. Undetected data quality issues impact 47% of decision-makers. Forecasts built on flawed data points can lead to poor inventory levels, decreased customer satisfaction, and reduced operational efficiency.
- Adapting to market trends and consumer behavior. Demand volatility can cause significant deviations from forecasted values. You may need to invest in varied, resource-intensive forecasting methods to manage such cases.
Managing these challenges sharpens your inventory forecasts, leading to smoother operations and a stronger market position.
5 Common Inventory Forecasting Methods
Going with your gut can leave your business wide open to inventory inaccuracies and inefficiencies.
While your intuition is valuable and shaped by experience, it’s still susceptible to bias.
Pairing it with these demand forecasting methods can make your inventory forecasts more objective and reliable.
Here are 5 types of inventory forecasting methods:
| Forecasting method | How it works |
|---|---|
| QUANTITATIVE FORECASTING | ---- |
| Time Series Analysis | Analyzes historical data to detect sales trends and seasonality. |
| Regression Analysis | Uses relationships between variables to predict future demand. |
| Economic Order Quantity (EOQ) | Calculates the right inventory quantity to minimize costs. |
| QUALITATIVE FORECASTING | ---- |
| Market Research | Evaluates consumer preferences and market conditions. |
| Delphi Method | Gathers expert opinions through iterative surveys. |
| Expert Opinion | Relies on informed predictions by industry experts. |
| MIXED FORECASTING | ---- |
| Demand Sensing | Adjusts forecasts with real-time data. |
| Collaborative Planning, Forecasting, and Replenishment (CPFR) | Collaborates with supply chain partners for accurate joint forecasting. |
| TREND FORECASTING | ---- |
| Long-term Forecasting | Includes general factors like social or cultural trends influencing customer behavior. |
| Short-term Forecasting | Factors in seasonal trends. |
| GRAPHICAL FORECASTING | Creates graphical representation of historical sales data to see market trends and sales patterns. It can be a line graph, histogram, or pie chart. |
Inventory Forecasting Metrics and Formulas
Metrics enable you to measure your forecast's performance against a set of goals or standards. They exist to support better long-term planning and strategy.
Use this quick guide to set inventory metrics and formulas that you can use when evaluating inventory planning strategies.
| Metric | Definition | Formula |
|---|---|---|
| Inventory Turnover Ratio | How often inventory is sold over a period; a low ratio suggests overstock | Cost of Goods Sold (COGS) / Average Inventory |
| Sales velocity (monthly) | The rate of sales factoring out stockouts | (365 days sales / # of days in stock during 365 days) x 30 days |
| Days Sales of Inventory (DSI) | How long it takes to sell your stock | (Average Inventory / COGS) × 365 |
| Service Level | Percentage of customer orders fulfilled on time | (Number of Orders Fulfilled / Total Orders) × 100% |
| Stockout Rate | Percentage of unfulfilled orders due to out-of-stock items | (Number of Stockouts / Total Orders) × 100% |
| Economic Order Quantity (EOQ) | The optimal number of units to purchase with the lowest total costs | √2DS/H, where:D = annual demand in unitsS = order cost per purchaseH = annual holding cost per unit |
| Reorder Point | Inventory level to trigger replenishment | (Average Daily Usage × Lead Time) + Safety Stock Units |
| Safety Stock | Buffer inventory held to prevent stockouts | (Maximum Daily Usage × Maximum Lead Time) - (Average Daily Usage × Average Lead Time) |
| Mean Absolute Deviation (MAD) | Average magnitude of forecast errors; lower MAD means more accurate forecasts | ∑(Forecasted Value−Actual Value) |
| Mean Squared Error (MSE) | Averages the squared errors to emphasize larger errors | ∑(Forecasted Value−Actual Value)^2 / N |
With these benchmarks set, we can now dive into the actual forecasting process.
How to Forecast Inventory, Step-by-Step
Whether you’re relying on pen and paper or cutting-edge software, inventory forecasting always boils down to the same fundamental steps because you need the same elements to come together.
Here’s the blueprint:
1. Collect historical sales data
In inventory forecasting, you can't look forward without looking back.
Your sales history shows your inventory movement, giving you insights into important factors like product performance.
Matt Bellerose, founder of seafood online retailer Lobster Order spoke on how effective it is to assess this type of information:
Using past sales data, we’ve reduced overstocking by an average of 15%.
So, how far back do you have to go?
The consensus is at least two years. This gives you ample time to see how well your product does on the market at different times of the year.
You can collect these data from various systems, including your ecommerce platforms and CRM software.
2. Analyze your sales data
You’ve built a comprehensive and accurate data set—well done!
Next up is poring over it to uncover patterns, trends, and anomalies that influenced your past performance.
As a first step, make a list of the information you need. Some good things to watch out for are:
- Product lifecycle stages. Understand a product's overall market lifespan by tracking its performance from launch to decline.
- Seasonality. Study sales patterns that fluctuate with the season or certain times of the year.
- Promotional impacts. Analyze the effect of discounts or special offers on customer demand.
- Out-of-stock situations. Monitor how often and when products have been unavailable.
These detailed evaluations can help you identify which products perform well (or not) at specific times, so you know when and which products to stock up on.
3. Choose the most suitable forecasting method
Earlier, we talked about the various forecasting techniques, and how each one has its strengths and weaknesses.
For this very reason, Amanda Bunch, the CEO of piercing jewelry online retailer BodyArtForms, opts for a hybrid method:
In my experience, quantitative methods have pros, such as you can make decisions based on facts and numbers (so long as they’re accurate), and you tend to feel more confident in your decision.
On the other hand, qualitative data is more agile in its approach, and it makes more sense for us as a company dealing with jewelry for piercings to choose this method as it works hand-in-hand with customer behavior and trends in the industry.
However, the accuracy is not as dependable as quantitative means.
Given how our company operates, we opt for a hybrid method of forecasting.
Your business needs, products, and the current economic conditions will dictate your forecasting technique.
Like BodyArtForms, however, many ecommerce brands prefer to combine qualitative and quantitative methods.
Doing so provides flexibility and adaptability that even online retailers of consumable goods like Lobster Order can rely on it.
When the pandemic hit in 2020, the seafood retailer’s quantitative models couldn’t absorb a 30% increase in delivery orders to homes. So the company adopted qualitative methods to navigate the unpredictable event.
In early 2020, our team’s gut feeling about the market told us that supply chain disruptions were coming, so we increased our inventory buffer by 20%.
As it turned out, this decision was critical when shipping delays hit.
We have used a mixed approach to best effect. Quantitative methods are the baseline, but qualitative insights from our sales and customer service teams help fine-tune the predictions.
Over the last two years, this hybrid model has improved our forecast accuracy by 22%.
If you’re dealing with seasonal products, multimodel forecasting techniques also often yield the most accurate results.
For instance, a study found that using Multiple Linear Regression (MLR) and Random Forest Regression (RFR) together improves seasonal forecasting accuracy.
MLR is good at identifying linear relationships between variables, while RFR captures complex, non-linear interactions. This hybrid approach cut forecast errors by up to 20%.
Last but not least, you need to select a timeframe for your forecast as well, which can range from monthly to annually.
JT Gill, marketing manager of supplement brand Wellness Extract, explained how the duration influences the best forecasting method you choose:
Short-term forecasts can benefit from quantitative precision, while long-term forecasts might require the flexibility of qualitative inputs.
Forecast periods also affect accuracy.
The longer the time period, the less accurate your forecasts are likely to be because you can’t account for any possible market changes.
4. Incorporate market trends and seasonality
Checking your previous sales can show you how the market fared in the past—but you can’t rely on the future to behave the same way.
Inventory forecasts require you to account for many moving parts, such as seasonality, market fluctuations, market initiatives, competitors, and customer behavior.
Doing so improves your projections and business adaptability.
Identify emerging trends early by staying up-to-date with industry reports and customer feedback.
Additionally, you can keep track of competitor activity and update your forecasts as the market shifts.
And for good measure, check your sales to see which circumstances or events are unlikely to repeat themselves and weed these outliers out of the equation.
Example: Seeing increased product demand after a viral social media campaign, a short-term influencer collaboration, or a natural disaster.
5. Identify optimal lead time demand, reorder points, and safety stock levels
Inventory forecasting takes into account more than just sales history and trends.
You also need the following core metrics: lead time demand, reorder points, and safety stock levels.
- Lead time demand is how much inventory you need to have on hand after placing an order to maintain sufficient stock until the next batch of inventory arrives.
- Reorder points trigger a new purchase order when you reach a certain level of inventory.
- Safety stock refers to the amount of inventory you keep on hand in anticipation of high demand or supply chain disruptions. You need a healthy amount to prevent stockouts, but not too much so that you don't build up a stockpile.
Knowing what the recommended level for each can help you better predict the best inventory level per SKU.
They’re usually computed as follows:
| Metric | Most common formula |
|---|---|
| Lead time demand (LTD) | LTD = Average Item Sold In A Day × Average Supplier Lead Delivery Time |
| Safety Stock (SS) | SS = (Maximum Daily Sales × Maximum Lead Time) – (Average Daily Sales × Average Lead Time) |
| Reorder point (ROP) | ROP = (Lead Time × Demand Rate) + Safety Stock Level |
6. Implement your forecasts
Once you've gathered the pieces, it’s time to put them together to make a forecast.
You can make manual calculations, use Excel, or rely on a software tool.
If you opt for Excel, here’s a quick guide:
- Input your historical sales data into a new spreadsheet in two data series. One for the timeline (e.g., dates) and one for corresponding sales figures.
- Go to the Data tab and click “Forecast Sheet” to generate your forecast from the selected cells.
- Choose a line or column chart for visualization, set your desired end date, and click “Create” to finalize.
Forecasts are used to help you guide stocking and purchasing decisions.
Identify and check your predictions against the inventory forecasting metrics you’ve set to measure how accurate your forecasts are.
7. Streamline the process with inventory management software and advanced forecasting tools
If all the tasks above sound laborious and time-consuming, it’s because they are. When you're forecasting inventory, the adage “nothing beats the classics” doesn’t apply.
Managing inventory is so much easier with technology.
For instance, inventory management software provides real-time insights into inventory status and syncs stock levels across multiple sales channels—both of which can dramatically improve data accuracy, and consequently, forecasting accuracy.
It also gives you detailed analytics and reporting and supports integration with other business systems, so you can gather relevant information such as historical sales data with ease.
If you opt for more advanced tools, such as Sage X3, Linnworks, Cin7, and Shipbob, then you can automate demand forecasting to know how much stock can meet customer demand.
For instance, Linnworks has the Stock Forecasting functionality, which predicts when you’ll need to reorder stock for an item based on how much stock is consumed per item over the coming year.
So if you’re looking to have better inventory forecasts, you need an IMS to set things up correctly.
I’ve already mentioned some of the best tools out there, but we have more for you to consider:
To give your forecasting an extra punch of accuracy, however, you can combine your inventory management solution with an AI-powered inventory forecasting tool.
The use of inventory management systems gives you the real-time data you need to make informed decisions.
Meanwhile, an inventory forecasting tool gives you the flexibility to adjust to trends and unexpected demand faster with machine-learning algorithms that improve forecast accuracy.
Cin7 CEO, Ajoy Krishnamoorthy reported that customers who have implemented its smart inventory forecasting tool, Inventoro, notice results almost immediately.
In just three months, after adopting Inventoro by Cin7 and following their recommendations, HairCo has seen a 10% increase in sales.
The solution became HairCo's right-hand man.
It accounted for lead times, minimum order quantities, and past sales data—predicting inventory needs three to six months in advance, an unprecedented window for the company.
Krishnamoorthy explained that inventory forecasting software enables sellers to make fast, informed decisions that translate insight into actions:
With Inventoro, for instance, you can achieve 99% product availability, while freeing up 20% in inventory capital and saving valuable time.
8. Monitor and adjust forecasts regularly
Inventory forecasting isn’t a one-off process, even when you have a tool to automate so many of its nuts and bolts.
Keeping forecasts up-to-date ensures their relevance and accuracy.
Plus, you’ll have to compare your forecasts to their performances religiously to see if they’re stacking up.
How accurate is it? Are the reorder points and safety stock able to cover lead time demand, including prolonged ones?
Use the gleaned insights to adjust and refine your forecasting approach and replenishment process.
Here’s a good rule of thumb for tracking forecast performance:
Plan in months, measure in weeks; plan in weeks, measure in days.
It’s super important to know how the plan is working, progressing, and course correct as needed.
Best Practices for Accurate Inventory Forecasting
Pair the steps above with these recommended practices, so you'll be able to create more proactive and reliable forecasts:
Review and adjust data regularly
Since your sales history forms the basis for your inventory predictions, you need to keep your data clean.
Poor-quality data can only produce off-target forecasts.
Establish a routine for reviewing your sales data, on-hand stock quantities, and other records to catch and correct inaccuracies.
Demand Planning Manager Nipun Sawhney emphasized the importance of a physical stock count to correct inaccuracies. “There should be a wall-to-wall cycle count where all items are physically counted,” he advised.
Simplify inventory checkups by introducing barcode scanners or RFID technology into your inventory management system.
Collaborate with stakeholders
Getting accurate inventory forecast isn’t only in the hands of supply chain teams.
Ecommerce businesses like Lobster Order bring together folks from sales, operations, and finance to get a well-rounded assessment of future inventory needs.
According to Bellerose:
Monthly meetings with sales, marketing, and operations representatives ensure that all views are accounted for.
This collaborative method has reduced our stock-outs by 35% year-over-year.
Aside from this, you can work closely with your suppliers to determine whether JIT inventory management is an option for you.
Focus on continuous improvement
Are there any gaps or inconsistencies in your forecasts? If yes, conduct post-mortem analyses to understand the root cause (e.g., data entry, model assumptions, or external factors).
Use metrics like Mean Absolute Deviation (MAD) and Mean Squared Error (MSE) to quantify forecast errors.
Doing so will help you identify patterns or recurring issues in the future.
Alternatively, an inventory management tool’s real-time data analytics can help you continuously monitor forecast performance and make adjustments on-the-fly.
Market conditions and other factors constantly shift.
With or without discrepancies, strive for continuous improvement in your forecasting practices to stay ahead of changes.
Enhance your inventory forecasting by leveraging accurate, up-to-the-minute data from a perpetual inventory approach.
Develop a contingency plan
Demand isn’t linear. While you do your best to meet it by sharpening your inventory forecasting prowess, it’s still a good idea to seal all the gaps.
Tightening your supply chain is one way to do this.
Work on improving supplier relationships and finding backup suppliers to help reduce stockouts and satisfy forecasted inventory.
Each supplier will have different levels of flexibility; determine what areas they’re willing to adjust for you, such as delivery methods.
Optimize inventory turnover
For ecommerce brands, a good inventory ratio is between four and six.
Using inventory management software is one of the best ways to meet this ideal range. Your tool can perform ABC analysis and generate reports that identify high and low-turnover items.
It can also set automatic reorder alerts for high-demand SKUs.
Phasing out stagnant stock also offers a cleaner dataset for your forecasts. It frees up resources for your top performers, reduces holding costs, and improves cash flow.
Train staff on forecasting methods
Develop your employees' skills in inventory forecasting by introducing them to various forecasting methods.
For ecommerce companies using inventory management tools, ensure they receive the necessary training on setup and operation, as well as how to maintain data quality.
Keeping everyone aligned and familiar with the process contributes to better forecasting.
These best practices all work together to improve your forecasts.
According to PierrePark CEO:
Using a cross-functional approach—integrating insights from sales, marketing, and logistics—regularly reviewing and adjusting forecasts based on market changes, and maintaining contingency plans have helped us handle unexpected demand spikes.
This holistic approach ensures we stay agile and responsive to market conditions.
Final Thoughts
Learning how to forecast inventory better is a lot to take in, especially since it’s an ongoing process.
Lean on ecommerce inventory management tools and forecasting software to simplify essential steps such as:
- historical sales data collection and analysis,
- optimal lead time demand, reorder points, and safety stock levels determination,
- and algorithm-based forecasting methods.
Taking these tasks off your plate allows you to channel your energy toward increasing forecast accuracy and improving inventory management.
Need more ecommerce inventory management tips? Perform at a higher level by learning the tricks of the trade here.
Keeping up with ecommerce trends requires constant learning and exposure to the latest insights. Join our newsletter to receive tips for growing your ecommerce business today.
Inventory Forecasting FAQs
Before we bid you adieu for this topic, let’s get into some final questions you may need answered.
What role does inventory management software play in forecasting?
Inventory management software features include advanced analytics and reporting tools that enhance forecasting accuracy. It automatically generates reports based on real-time stock levels, sales trends, and demand patterns.
Automation is one of the top benefits of IMS, giving the tools the ability to handle much of the routine work and support data-driven decisions.
However, RFID data collection and barcode scanning are necessary for improved inventory data accuracy, as well as more advanced inventory forecasting tools for more accurate and agile predictions.
What are some advanced forecasting tools available for ecommerce businesses?
Some ecommerce stores like Balance One Supplements use sales forecasting tools to “set appropriate stock levels, and prevent overstocking of products that might have shorter demand cycles.”
CEO James Wilkinson recently launched a range of immune health supplements that came just before cold and flu season:
“We used our sales forecasting tool to predict an upsurge in demand, which helped us to pre-stock inventory and efficiently meet the demand.”
But most advanced inventory management tools are comprehensive enough to track sales history and predict its trajectory alongside your inventory levels.
Many of them integrate with stand-alone inventory forecasting tools, such as Inventoro by Cin7 and Inventory Planner by Sage.
This type of software leverages algorithms to analyze data and adapt to changing circumstances. With flexible forecasting, brands can achieve high product availability, improved replenishment practices, and proper resource allocation.
