Inventory forecasting is essential to make sure that you avoid having stockouts or over-ordering products. Forecasting becomes more complex when you’re running a seasonal business with highs and lows. We take you through what to consider to forecast customer demand correctly.
There’s no one forecasting strategy that every ecommerce business should be using. Instead, there are different options for different goals and data sets.
If you’re a seasonal business, you need a model that factors in changes in velocity. So, you need to make sure the solution you choose works for seasonality.
For example, the economic order quantity (EOQ) formula is popular among ecommerce businesses as it can reduce logistics and warehouse costs through larger orders. But it ignores seasonal fluctuations and assumes constant, regular demand. That could risk leaving a seasonal business with too much stock in low periods and not enough stock during the high ones.
Which forecasting technique you choose can largely depend on which data you have available and the forecasting goal.
Whether you have historical data or are a new business just getting started, there’s a forecasting technique that can suit your needs.
Qualitative forecasting uses external research to predict demand. Here’s the data you’ll need:
Qualitative forecasting is best used when historical data isn’t useful, relevant, or available. It’s often used by new businesses but requires a lot of subjectivity, so it's hard to create detailed and accurate forecasts. However, this method can still work well. Just make sure your research or third-party data factors that seasonality in.
Quantitative forecasting uses historical demand data and mathematical formulas to determine future demand. You’ll need historical data like:
Unlike trend or graphical forecasting - which relies purely on past sales data - this method combines past demand data with mathematical formulas to determine future demand. There are two main ‘time series analysis’ models (more on this below), both flexible enough to account for rare events and seasonality.
Graphical forecasting uses historical data to create a visual representation of demand. Here’s the data you’ll need:
This can be an effective way to identify patterns and trends. Due to the visual nature of this forecasting, it can help to visualise seasonal peaks and lows across time.
However, this method relies on past data, so it may not be possible for new businesses.
Trend forecasting uses historical data to project potential future trends. You’ll need at least two year’s worth of the following data:
This method isn’t as flexible as quantitative forecasting as it may not consider seasonal fluctuations. Due to the amount of historical data needed, it also wouldn’t suit a new business.
Forecasting models use one of the techniques mentioned above to determine different kinds of stock. Each model serves a specific goal. We’ve included models useful for seasonality that can help you see how much safety stock you should have, when you need to reorder inventory and predict demand.
Safety stock is extra inventory held to make sure you don't run out due to a supply chain issue, emergency or surge in demand.
(Maximum daily use x Maximum lead time) – (Average daily use x Average lead time)
You should use this formula with the Reorder point formula below. It’s always helpful, but especially during seasonality or global volatility.
Reorder points for vital shock keeping units (SKUs) help you determine when to reorder, so don’t end up with too much or too little of an item.
(Average daily use x Average lead time in days) + Safety stock
The reorder point formula is essential to avoid stock-outs or overstocking, especially during seasonality. Carry out this formula regularly to account for any changes.
Time series analysis examines patterns in past behaviour over time to forecast future behaviour. There are two main models. The first is ‘moving average forecasting’, which takes a previous period’s demand data and calculates the average demand over that period to forecast over a ‘moving’ period of time.
Demand over time period / number of months, weeks or days
Moving average forecasting is a great model to give you an idea of average demand. The model gives equal weight to each period and only considers data during the chosen period.
Although this can factor in seasonality, it can ignore the impact of random events if they happen outside the chosen time frame.
The second model is exponential smoothing. It looks at the actual demand of the current period and the forecast previously made for the current period. You can tailor the results to fit your company’s market positioning by adding a smoothing constant.
(D*S)+(F*(1-S))
D = most recent period’s demand
S = the smoothing factor represented in decimal form
F = the most recent period’s forecast
Exponential smoothing is a more advanced approach that overcomes moving average forecasting’s problems. Therefore, if you’ve got the time and data, this is the model you want to use. However, it can be particularly time-consuming to execute.
While inventory forecasting can help you see how seasonality affects your business, it’s also good to have a target turnover to help measure your success.
The consensus in the ecommerce industry is that a good stock turnover is a ratio of between 4 to 6, which means you’ve sold and restocked 4 to 6 times the amount of inventory that you’ve stored. This means that the average ecommerce store should be selling their entire inventory every 60-90 days.
In comparison, if you only have a turnover of 2, either your margins are very high and you’re selling luxury goods, or your customers aren’t choosing your product and you’re at risk of losing out to competitors.
Although inventory forecasting isn’t a certainty, it can help you plan for different possibilities and make informed decisions to prepare for seasonal shifts. It’s also important to be prepared for sudden shifts caused my macro economics, and forecast on a regular basis if there are any big changes.
Juni is here to help you take control of your cash flow with tools and intelligence you need to make smarter decisions.
Juni’s virtual cards, multicurrency accounts, and accounting and advertising integrations give you a unified view of your finances. Get increased visibility of your cash flow and transactions with insights and analytics tools. Get control of your transactions with custom notification settings and transaction reviews by assigning cards to managers. Juni can help you boost your cash flow with cashback on all your eligible spend and access flexible capital with extended terms*.
* Capital for cards is available for companies registered in UK, NL, SE, DE, FR, ES, IT and FI, upon eligibility. Capital for invoices is available for companies registered in NL, SE, DE, FR, ES, IT and FI, upon eligibility. Fees and terms and conditions apply. Find out more here.
Juni is not a bank – yet. We’re a financial technology company providing e-money services and as such are not part of a deposit guarantee scheme. Learn more about our license and how we keep your business and money safe.
Juni Technology AB (company number 559248-0908) is licensed and registered as an electronic money institution with the Swedish Financial Supervisory Authority (Sw. Finansinspektionen) (license no: 64025). Payment Card Solutions (UK) Limited (company number 05941947) and UAB B4B Payments Europe (company number 305539054) are regulated by the Financial Conduct Authority (ref: 930619) and the Bank of Lithuania (licence no: 76), respectively. The Juni Mastercard Corporate Card is issued by Payment Card Solutions (UK) Limited and UAB B4B Payments Europe pursuant to a license by Mastercard International Incorporated. Mastercard and the circles design are registered trademarks of Mastercard International Incorporated. Juni Technology AB. Masthamnsgatan 21, 413 28 Göteborg, Sweden. © 2025 Juni. All Rights Reserved.
