Tutoring is not magic, but it can be powerful
The 2020 meta-analysis by Nickow, Oreopoulos and Quan synthesized experimental evidence on tutoring and found strong positive impacts on student learning, with an average effect of 0.37 standard deviations.
The more useful parent lesson is that not all tutoring is equal. Programs with stronger structure, better training and closer links to normal schoolwork produced stronger outcomes.