sellers, page-16

@adam: agreed, good personal experiences with the product would be a good start - just as Mark O'Brien has said.
But a well constructed research trial would be so much more powerful, and easy to execute

It shouldn't be that hard:
Get 20, well trained cyclists.
Give 10 the product and 10 a placebo. Have them perform a cycling test to exhaustion or a time trial over a set distance and measure time. Do it indoors on a wattbike or similar to remove external variables.
Get them back a week later and swap the product and the placebo. Bang - good data.

The beetroot juice trails are well known for a similar study design and showing good research results in cyclists. ( http://www.beet-it.co.nz/uploads/3/0/9/9/3099302/lansley_et_al_2011_acute_dietary_nitrate_supplementation.pdf )
And sales of Beetroot juice/shots have boomed since.