The headline retail sales number is badly distorted by vehicle units and petrol prices, both of which swing wildly and neither of which tells you much about underlying demand. The control group strips them out, along with the two categories that are sourced differently in the national accounts.
Because it maps almost one-for-one into goods consumption, the control group is the number GDP trackers update on. It is also nominal, so in a high-inflation period a strong control group can mean price increases rather than real volume growth.
Example: headline retail sales fall 0.2% as petrol prices drop, but the control group rises 0.7%. GDP nowcasts revise up despite the negative headline, and the front end of the curve sells off.
Related: gdp, nowcasting, consumer-confidence, industrial-production, real-gdp