"I think some research suppliers oversell a bit they want us to do too much market research. For example, one of the biggest recently published data on a 'blind' versus 'identified' product test. Here are the results:
Branded..........................................Unbranded
Prefer A 55.5% .............................. Prefer A 45.6%
Prefer B 44.5% ............................... Prefer B 54.4%
Prefer A 68.0 ..................................... Prefer A 60.7
Prefer C 32.0 .....................................Prefer C 39.3
Prefer B 64.4
Prefer C 35.6
I'm told the differences were highly significant statistically. The research firm concluded that there was no choice between blind and identified but that both should be used in just about every case where there was any reason to even suspect an effect of branding. Do you agree?"