The article shows how the combination of trends in production customisation and digitalisation requires the use of new methods of forming a customer pool and implementation of a strategy of concentrating on prospective customers. Small and medium machine-building enterprises of Sverdlovsk oblast were examined. The study introduces the indicator «customer potential coefficient», which determines an increase in sales volume from the next orders of a customer relative to the first placed order. To calculate the customer potential coefficient, including new ones, multiple regression models are used, data for which are taken from open sources publishing external information and enterprise reports. This makes it possible to create an algorithm and automate the calculation process when developing appropriate data integrators.