We boost the profit margins of trading companies ranging from a single warehouse to national networks
Demand Forecasting and Stock Management Software
We reduce order placement time from several hours to 20–30 minutes.
Guaranteed 10% reduction in inventory reserves without any loss of sales.
Our software has been successfully deployed at 70+ companies.
Sales
Procurement ignores sales plans and seasonality, so customers are unable to find the products they need. This ends in frustration and customers going to a local competitor.
Finance department
The company is losing money due to its dead stock and bloated safety stock. It’s hard to say exactly how much without considering all factors.
Warehouse logistics
Half of the warehouses are overloaded, whereas others lack products that they need. As a result, some of the products exceed their expiration date.
Marketing
Most likely, the promotion failed due to a stocking shortage. Nobody really knows, because it is difficult to measure its effectiveness.
Over the last 10 years we have seen many inventory management techniques.
And each has potential drawbacks…
The procurement department suffers as a result: They work overtime and hire expensive experts.
Forecast NOW! helps purchasing work more efficiently, improving the company's bottom line.
Sales
The system has compiled and coordinated a demand forecast with the sales department plan.
Finance department
There is a long-term procurement plan for assessing the financial budget. The most cost-effective level for satisfying demand has been calculated, taking into account cost and storage risks.
Warehouse logistics
Stocks have been distributed correctly between warehouses, taking into account moving costs and price differences.
Marketing
You can plan the effectiveness of marketing campaigns while taking into account cannibalization of demand. And you can compare the results with the plan.
Eyeball estimation
Recommendation of suppliers
ABC analysis as a procurement tool
Assigned experts decide on the methodology
Forecast based on a single number or an interval
Various departments have their own methodology
Manual determination of the minimum and maximum stock levels expressed in units
Order placement when inventories are out of stock
No calculation of the optimal level to satisfy demand
Buyer calculation according to his or her own method
Personal decision of a safety stock expressed in units
Average daily sales forecasts concerning YZ products
Utilization of the theory of constraints for products with intermittent demand and long delivery wait times
Ordering based on shipping data from main warehouse to distribution centers, as opposed to the final sales from distribution centers to end consumers
Algorithms originally invented for scientific applications are repurposed
Point forecast for demands with normal distribution, excluding FMCG
Counting of errors on the basis of statistical criteria as opposed to monetary criteria
Over the last 10 years we have seen many inventory management techniques.
And each one of these has its own
flaws…
Sales
Procurement did not take into account the sales plan and sales season. Customers consequently cannot find the products they need. They express their frustration, and they go to a nearby store instead.
Finance department
The company is losing money due to its dead stock and bloated safety stock. It’s hard to say exactly how much without considering all factors.
Warehouse logistics
Half of the warehouses are overloaded, whereas others lack products that they need. As a result, some of the products exceed their expiration date.
Marketing
Most likely, the promotion failed due to a stocking shortage. Nobody really knows, because it is difficult to measure its effectiveness.
Forecast NOW! helps purchasing work more efficiently, improving the company's bottom line.
Sales
The system has compiled and coordinated a demand forecast with the sales department plan.
Finance department
There is a long-term procurement plan for assessing the financial budget. The most cost-effective level for satisfying demand has been calculated, taking into account cost and storage risks.
Warehouse logistics
Stocks have been distributed correctly between warehouses, taking into account moving costs and price differences.
Marketing
You can plan the effectiveness of marketing campaigns while taking into account cannibalization of demand. And you can compare the results with the plan.
It forecasts product stocks,
though not demand.
Most companies are still using demand forecasting methods that fell out of favor 10 to 15 years ago, including exponential smoothing, ARIMA, moving averages, the Holt-Winters method, and others. These are not only outdated, but they are also ineffective for solving inventory management problems for 94% of the product range in the fast moving consumer goods (FMCG) segment as well as for almost all non-FMCG products. This has been proven by many scientific studies.
Why is it necessary to forecast product stocks rather than demand?Classical forecasting means one-number forecasting, regardless of whether it is a daily, weekly or monthly forecast. Such methods are based on various algorithmic approaches, such as average daily sales, moving averages, the Holt-Winters method or ARIMA. In order to assess the accuracy of the forecast, mathematical methods are used that determine the forecast error as a percentage. They have no relationship to money.
This approach assumes the use of a safety stock unless the user deliberately decides not to add one, in accordance with their expert evaluation. In fact, this model cannot be managed. For example, if the goal is to ensure a certain level of demand, it is impossible to take any action that could influence this.
The output is a demand forecast for a given interval (week, month or year). For example, we will sell between 6 and 12 product units in a week. Or, in other words, there is a 95% probability that sales will not exceed 12 units. The interval estimate improves upon the classical approach by expanding the safety stock, which is adjusted either up or down in order to ensure demand with a given level of service. At this stage we encounter the concept of the type of distribution, which describes the safety stock.
In the vast majority of cases, this is the Normal distribution and Poisson distribution. The problem is that 96% of the products at a supermarket that just sells groceries do not have a normal distribution, and in fact this approach is not suitable for grocery retail and other areas of trade and distribution. An ideal application for this approach is a very deep planning horizon and regular sales that are stable day in and day out. Unfortunately, under modern market conditions this situation is almost impossible to achieve.
Here we can already calculate the distribution for each specific product, and as a result we get an estimate of the volume of demand and the probable distribution of sales. This approach has already yielded many more opportunities to manage demand. We already have the full-fledged ability to independently manage the service level, that is, to find the optimal balance between how much demand needs to be met and how many products are stored in the warehouse at any given moment in order to maximize profits. But due to the large number of calculations and limited resources, low-probability demand levels (<1%) are not taken into account.
This introduces a certain margin of error, since exceptional cases on which the business could have made money given the risk balance are lost. Another drawback of this approach, which is true of all of the previous ones as well, is the linear estimate of residual product stocks as of the date of receipt of the order. Nevertheless, such models have proven themselves to be effective for selling spare parts as well as in the aviation industry.
These models make it possible to estimate the distribution of even the smallest probabilities of demand. They provide probabilistic forecasting of remaining inventory and probability estimates of lead times. And here we are able not just to generate a forecast, but to describe all kinds of risks (shortages and late deliveries) and exceptional situations (one-time large sales) that can be used to benefit the business.
Full probabilistic models provide a significant boost to business. They help you to assess whether you need to invest money in stock, or instead to hold off where the likelihood of sales is negligibly small and doesn't justify the potential profit or unavoidable costs.
Hours spent
developing
the mathematical model
500 +
Studies
based on real data
in real industries
70 +
Successful
deployments
10
Years of developing
the methodology