How to get results from research for farming

Some R&D work with farming clients over the last couple of years has given some clues about how to get results from farm research.
Most people now realise that research for farming has not lifted productivity over the last couple of decades. Difficult seasons may be one reason, but my data indicates that productivity has not improved, even if converted to water use efficiency, which should remove most of the dry season effect.
Another objection to this claim has been that no-till has improved farm profitability, and this has resulted from farm research. To this, I would explain that no-till had already happened in peoples minds, more than two decades ago, and similar developments in mind-set have not been occurred since.
I contend, that we have had the wrong approach, in relying increasingly on somebody outside of farming doing the research and providing the expertise. GM cropping is one example, but the increase in number of agronomists is another.
The R&D being done by clients is a long story in itself, but is showing that it is possible to do better.
At the risk of criticising our current research institutions, these ideas need to be developed further. Hopefully I can contribute to this, amongst other day-to-day crises.