Good comment, I agree. We need more information about what works to improve learning outcomes. This can only come from rigorous impact evaluations. While we still have a long way to go, there are some things that we know. For instance, for rapid improvers:
Poland: 1999 education reform paid off; extending comprehensive education helped improve learning outcomes considerably; see: http://ideas.repec.org/p/wbk/wbrwps/5263.html
Chile: controversial, but the 1980 reforms which extended learning opportunities and extended choice might be paying off: see for example: http://ddp-ext.worldbank.org/EdStats/CHLprwp08.pdf; http://www.economia.puc.cl/politicas-publicas?docid=5525
Mexico: Government intentionally targeted PISA scores and enacted policies to expand opportunities (eg, through the conditional cash transfer program, see for example: http://ideas.repec.org/p/got/iaidps/122.html) and improve school quality (see for example: http://www.sciencedirect.com/science/article/pii/S0304387811000927).