Frequently asked questions

First of all, we focus specifically on demand forecasting for grocery retail. We take into account assortment characteristics and industry-specific tasks such as fresh food demand forecasting, safety stock optimization, sales promotion forecasting, and intelligent pricing to get measurable results through optimization and automation of decision-making processes.

Secondly, we utilize machine learning as a technological core of our solutions. Our team has 10+ years of experience in applying machine learning to solve business problems, and we are proud to say that DSLab services demonstrate higher levels of forecast accuracy and business performance than competitors' solutions.

Thirdly, we deliver measurable results. Our solutions are designed to bring tangible business value. We help you make money out of the data you have, improving on-shelf availability, turnover, and revenue.

We provide demand forecasting services to help grocery retailers increase operational efficiency and revenue while reducing costs. Our services deal with all sorts of forecasting tasks: fresh food forecasting, safety stock optimization, sales promotion forecasting, intelligent pricing, regular demand forecasting, and automatic replenishment.

With machine learning technologies at their core, our services are end-to-end and are integrated into existing business processes. Contact us to find out more

Our forecasting algorithm calculates daily demand forecast, required safety stock and an optimal amount of each product to be ordered. The results of the forecast are transferred to the retailer's replenishment system and the order is placed automatically.

To make sure this never happens and our predictions are realistic, we add an automated control module to the forecast. The module is based on the expertise of retail professionals about certain demand characteristics for all types of goods.

No. You don't need to have any specific data science or machine learning expertise in-house to start working with us. From data cleaning and preparation to model training and regular updates: we'll take care of everything.

We use data provided by the client, such as sales volumes, price fluctuations, ongoing sales promotions, available stock, delivery dates, delivery unit, batch size, store and location characteristics, and many more. When applicable, we use external and third-party data – from weather forecasts to nearby traffic.

The more the better. As a rule of thumb, to successfully apply machine learning, the size of the dataset starts from tens of thousands of entries (i.e. sales). However, for some tasks, our approach requires fewer amounts - months rather than years - of data.

We provide several options: the service can be deployed as a SaaS or directly on your premises. In a SaaS mode, data transfer is required and both model building and data processing happen in the DSLab private cloud or on-premise. We recommend using the SaaS mode as it tends to demonstrate the best price/effectiveness ratio.

However, if the data transfer isn't feasible (i.e. due to performance issues or government regulations), we deploy service on your premises reusing existing infrastructure.

We ensure security by end-to-end encryption of data. Potentially sensitive data (i.e. data on particular stores and SKUs) can be anonymized through hashing or use of random identifiers. All data transfers are conducted through a secure VPN channel.

Our services integrate smoothly into the existing infrastructure via API. The services line up with standard data sources such as Hadoop or SQL-based data storage, as well as replenishment systems. On average, it takes one month to get the service up and running.

Our services don't have a visual interface. Integrated via API, they are embedded in the existing workflow. No need to accustom yourself to yet another interface, everything is familiar!

We provide 24/7 customer support. We also include daily model re-training as part of our service to improve the quality of a forecast when new data is available.

We use a subscription model based on the number of stores in a retail chain. The more stores you have, the lower the subscription fee for each store. No matter the size of a retail chain, our services are n-fold cheaper than those of big software corporations.

We conduct A/B testing to compare business value delivered by our service to the business value of existing or competitors' solutions. We use business metrics, such as on-shelf availability and write-offs to measure the quality of the forecast. With these metrics, return on your investment to DSLab services can be easily calculated. Request demo