The paradox of the Paradox of Choice

There’s a famous experiment, conducted by Mark Lepper and Sheena Iyengar in the late 90’s, which purports to demonstrate that too much choice is bad for us. They set up a jam stall within a supermarket, and on different days offered different varieties of jam; sometimes six different flavours, sometimes 24. They found, to their surprise, that the six flavour stand sold more product than the 24 flavour stand, and concluded that too much choice was bad for consumers. This story has been told and retold countless times, and is often cited as an example of applying Behavioural Economics to websites.

The problem is that it’s not true.

The results from the experiment are correct, sure, but the conclusions drawn from those results aren’t. The experiment itself has been repeated and the results have not been duplicated. There’s really no strong correlation between the number of choices consumers are presented with and the amount of sales. 

This isn’t too surprising. Amazon are hardly known for their lack of choice, nor for their lack of sales.

We need to be careful when extrapolating results from small numbers to apply to large numbers, because people are not equivalent to numbers on a spreadsheet, and context is key. Experiments may not be repeatable under different circumstances, with different audiences, products and purchasing contexts, but in embracing an experimental approach you may find an approach which fits for your particular circumstance and ends up paying dividends.

When considering this particular problem, perhaps a better way of looking at it is ‘how much work does the customer have to do in order to make a decision?’. If a rushed customer trying to do their weekly shop has to sample 24 different jam products to find the one they prefer, perhaps they won’t buy. If I have to look through 24 similar looking products on your website to find the one for me, I probably won’t buy. But if you can create a usable way of guiding and helping me find the right one out of many, using strong UX and cues from behavioural psychology, experimenting along the way, perhaps it doesn’t matter how many I have to choose from.