Sales forecasting is important for any business. However, for businesses that are involved in selling ‘perishable’ products such as groceries, concert tickets, train tickets, airline seats, and much more, miscalculating future sales can be a very costly exercise. Accurate sales forecasting can help a business to better staff its facilities, maintain adequate inventory levels, and ultimately be more profitable.

The problem with traditional sales forecasting

The traditional sales forecasting system involves attaching a percentage probability to the success of every single deal in the sales pipeline. This is then multiplied by the revenue value of that deal. All these individual estimates are added together to create the final sales forecast.

The problem is that the weighted pipeline method is flawed on many fronts. In the first place, it does not address the reality that sales is a zerosum game. If a $100 000 deal is in an early stage, it could be assigned a close probability of 25 percent. In the sales forecast, this will give it a value of $25 000 ($100 000 x 25%). In practice, however, the deal will either bring in $100 000 or nothing never $25 000.

Secondly, the weighted pipeline method is extremely subjective. Sales reps often tend to err on the side of optimism. From a sales perspective, this optimism is good, but it can lead to grossly inflated sales expectations.

How AI can help with better sales forecasting

According to research carried out by the Aberdeen Group, companies that are able to accurately forecast sales are 10 percent more likely to experience regular annual revenue growth. They are also 7 percent more likely to hit their quotas.

AI is showing great promise in improving sales forecasting because it addresses several of the inherent weaknesses of traditional forecasting methods such as the weighted pipeline. In many companies, it is already outperforming human forecasters by speeding up the decision-making and planning processes. This is one of the reasons why a number of well-known brands have already integrated AI into their sales planning.

Below are a couple of examples:

Perishable products. According to Professor John Paul B. Clarke from Georgia Tech, just about all perishable products (see above) will sooner or later use AI-based sales forecasting. The reason is that these industries need to adjust sales and pricing on the fly to make sure all their available products are sold before they expire.

He quotes the example of a concert where tickets are selling faster than expected. If this was your business, you could increase the price of the remaining tickets without fear, because you know that they will sell even at the higher price. If, on the other hand, the AI system predicts that a certain flight will be half-empty, the airline could slash the price of tickets before the time to make sure the plane is full.

Sales optimization. E-commerce colossus Amazon is currently using AI to optimize sales forecasting. In its case, it does a deep data analysis to pinpoint vital information about customers’ online buying habits and then suggests products that will fulfill those needs.

The AI system also issues a recommended reply for any prospect, depending on the data it collected via their emails and other correspondence. The system is able to improve user experience and therefore also boosts online sales.

The AI can also calculate the percentage probability that a customer will in fact make a purchase better than traditional systems. This improves the company’s short-term and middle-term sales forecasts.