With the introduction of carbon peaking and carbon neutrality targets, the influx of renewable energy into the market has triggered the needs for demand-side response. However, demand-side users are difficult to be predicted for their responsiveness due to the volatility of their electricity consumption behaviour and the difficulty of obtaining comprehensive electricity consumption and socio-economic information. Therefore, for the electricity sellers or aggregators which is difficult to incentive the demand-side resources to making them participating in the auxiliary service market. In the paper, a customer segmentation method based on the responsiveness has been proposed, and the smallest input data set is figured out by the support vector machine algorithm under the premise of guaranteeing a certain classification accuracy. Additionally, a significance analysis is carried out by Analysis of Variance algorithms has been used to summarize what electricity consumption characteristics and socio-economic characteristics of the users who are able to carry out effective demand-side response under the tariff packages with different peak-to-valley price differentials. The conclusions of this paper have certain value for the electricity retailer or aggregators to participation in the energy auxiliary market, and to incentive their users to dig their own flexibility adjustment ability. The conclusions also valuable in recommending the most suitable time-of-use tariff package for the customers.