Inventory forecasting is crucial to the financial success of any retail business. It helps strike a balance between sinking too much cash into inventory at once, while ensuring demand can always be satisfied without going out of stock.
However, this is also one of the most difficult aspects of inventory management to get right. So in this chapter, we outline all the techniques and best practices retailers can use to forecast inventory requirements.
Get the complete guide to Inventory Management
Download all 10 chapters of this inventory guide in one handy PDF - so you can easily upscale inventory management in your retail business.
What is inventory forecasting?
Inventory forecasting is essentially making an informed projection on how much stock will be needed to satisfy demand over a given time period. It starts with a simple demand forecast, then uses what’s already in stock to plan how much inventory is required going forward.
It’s important to note that forecasting will always be educated guesswork. No forecast is set in stone, and there are several factors that can affect accuracy – such as seasonality and sales history.
Setting forecast boundaries
The first step in predicting your inventory requirements is to create a simple forecast of your expected sales. For this, you’ll need to set some forecast boundaries.
1) Forecast period
A forecast period is the specific amount of time into the future that a forecast will be attempting to predict.
We recommend at least the following three periods:
- 90 day.
- 30 day.
These should then be reviewed each month. If market trends or actual sales performance is different than expected, then upcoming sales forecasts can be adjusted accordingly.
2) Base demand
Base demand is simply the exact current demand for a product at the specific point a forecast is due to begin from.
For example, a company may be doing a 30 day forecast for white Nike sneakers. If they sold 37 units over the previous 30 days, then base demand would be 37.
Incorporating trends and variables
It’s not enough to forecast inventory based purely on current demand. There’s a whole host of factors that could impact the data going forward.
Most businesses will therefore need to take a variety of trends and variables into account in order to achieve the most accurate inventory forecasting possible.
1) Sales velocity
Stock-outs shouldn’t happen, but in reality they sometimes still do. And sales velocity takes this into account when it comes to looking over your past sales performance for inventory forecasting.
Maybe you only sold 20 units of a product in the past 90 days. But it doesn’t tell the whole story if you actually had it listed as ‘out of stock’ for 89 of those days.
Forecast better for the upcoming period and you could sell much more.
We can therefore use the following calculation to omit out of stock days:
Note: The time periods are obviously not fixed here. You could work out 10 day sales velocity from 180 days of sales, but larger data sets tend to yield more reliable results.
All things being equal, sales velocity gives an indication of how much a product should sell if continuously in stock over a 30 day period. Something that can be very useful for forecasting inventory requirements into the future.
However, there are several other factors that can impact this too…
2) Marketing activity
It’s also essential that you consider marketing and advertising activity when it comes to inventory forecasting. This should be taken into account in two ways:
- Past marketing activity when looking at sales history.
- Planned marketing activity when planning for future sales.
The main thing to watch out for is whether planned marketing activity is conceivably different or scaled up/down compared to past marketing activity.
It could be realistic to expect a 25% sales uplift if Facebook Ads worked well last Q3, and you decide to increase budget by 25% in this Q3. So you’ll need to account for this when inventory forecasting.
Seasonality is absolutely critical for forecasting your stock requirements.
Winter coats tend to not sell well in summer months. Good gifts will tend to pique around the Holidays. Items you discount will possibly go through the roof during Black Friday.
But there may also be more subtle, not so obvious variations in demand for certain products.
This is where 12 month sales data is powerful.
You’ll need to look back over the previous year (several years if possible) to see which specific months and periods in time certain items:
- Start to trend up.
- Start to trend down.
This allows you to make data-backed decisions on seasonality, and how much inventory you might require during these periods.
4) Unexpected publicity
Unexpected media attention or publicity may be unlikely. But it’s still something you’d need to forecast for if it happens.
You’ll probably need to forecast an uptrend in sales if Kim Kardashian is pictured wearing some of your jewellery. Or be prepared for a possible downtrend if you get some bad press in a national newspaper.
Either way, unexpected publicity is definitely something you should be aware of and reacting to with your forecasts.
5) Industry-specific effects
Events and goings on within your industry and marketplace in general can also impact demand for products.
This could be a whole range of things, such as:
- A major competitor going out of business.
- A large company diversifying into your niche.
- Major changes to pillar marketing channels (if Facebook banned your Ad account, for example).
- Law changes in states/countries you operate/sell in.
Forecasting for new products
As stated earlier, inventory forecasting is always going to be somewhat of a guess. And that becomes even more so when it comes to new products with limited sales data available.
The key is to try to inform your guesswork as much as possible.
So consider things like:
- Trends of similar products you’ve launched.
- Trends of other products within that category you’ve launched.
- Trends of all previous products you’ve launched.
- Using a tool like Google Trends to see seasonality of when people search for specific products most.
- Making use of market research, surveys and focus groups.
Inventory planning and replenishment
Sales and demand forecasting is one thing. But true inventory forecasting needs to go a step further and actually plan out how you’ll replenish stock for the upcoming period.
This means considering:
- Current stock levels. How much is currently on-hand? There’s no point purchasing 40 units to cover 40 forecasted sales if you already have 27 units on-hand.
- Pipeline inventory. How many units have already been ordered and en route as pipeline inventory? You don’t want to double buy stock.
- Lead time. How long will it take for new stock to be delivered, received into inventory and made ready for sale? We’ll cover lead time and reorder points with more detail in Chapter 4: Purchasing Inventory of this guide.
Automated inventory forecasting
Forecasting demand, sales and, in particular, inventory can be an incredibly complex task.
Luckily, one option is to use automated inventory forecasting.
Tools like this won’t do all the work for you – you’ll still need to analyse data and consider any unexpected sales up or downturns. But they’ll be able to take previous sales performance and run accurate reports on estimated inventory requirements.
Veeqo, for example, helps take the guesswork out of the entire process.
Just select a sales history period to use, how far you want to forecast forward and any expected up or downturns and Veeqo will calculate your inventory requirements:
Discover how Veeqo can transform your inventory forecasting
Start a 14-day free trial today, no credit card required.
As you can see, there’s a lot of estimating that goes into forecasting inventory. But the more you can base decision making on data, the more accurate you’re going to be.
Many guides on inventory planning simply look at reorder point and economic order quantity (EOQ) calculations, without addressing how to actually forecast demand like done here. However, we do look at reorder points and EOQ in the next chapter on purchasing inventory.