4 ms·

Waiter, There's Pseudo-Science in My Soup

15 points by alexandros 17y ago

rabidsnail 17y ago
I know nothing about shipping physical products, but how difficult would it be to make 50 different kinds of labels, ship each kind to a different set of stores, and measure the sales of each store? That seems like it would be more reliable than trying to detect differences in people's "emotional responses".
- clistctrl 17y ago
  maybe, but I think there are forces involved besides the label which could alter the results. For instance, what if the weather is colder in Test Space 1 so soup sales increase not because of label, but because of desire for some warm soup. Maybe soup sales are down at test space 2 because that particular location had a super sale last week so a large portion of the local consumers are no longer in the soup market this week.
  - tel 17y ago
    You'd use the statistics from the likely huge amount of sales data that Campbell almost certainly already has to adjust for things like that. You'd also have a magnificent sample size to play around in.
petercooper 17y ago
What an attitude the author of this piece has. He keeps referring to "pseudo-science" and uses scare quotes all over. Campbell's chooses "neuromarketing" over consumer feedback in rebranding its iconic soup cans. Of course it does. It's trying to sell as many cans of soup as possible, not please a focus group. Implement what ultimately works, not just what customers request. Of course, this smacks of the same pseudo-scientific marketing bullshit Is this guy totally unaware that there is an art to persuasion (some of it grounded in results from scientific studies, though not all) and packaging? It's not like astrology. There are psychological tricks and shortcuts that are taken in most high end products and advertising to make them perform better.
- mbreese 17y ago
  They used a sample size of 40 to determine the packaging for a multi-million (billion?) dollar company. Even if they used every possible technique to capture the "emotional attachment" to a label, it wouldn't be a large enough sample size to be statistically significant. If you're going to go to the trouble to run the tests, at least use a large sample size. There is no way they could have gauged the entire soup buying population with just 40 people. Like what was mentioned elsewhere in this thread: they would have been better off shipping 50 redesigns to various places and seen how they did in the marketplace.

15 points by alexandros 17y ago

I know nothing about shipping physical products, but how difficult would it be to make 50 different kinds of labels, ship each kind to a different set of stores, and measure the sales of each store? That seems like it would be more reliable than trying to detect differences in people's "emotional responses".

clistctrl 17y ago
maybe, but I think there are forces involved besides the label which could alter the results. For instance, what if the weather is colder in Test Space 1 so soup sales increase not because of label, but because of desire for some warm soup. Maybe soup sales are down at test space 2 because that particular location had a super sale last week so a large portion of the local consumers are no longer in the soup market this week.
- tel 17y ago
  You'd use the statistics from the likely huge amount of sales data that Campbell almost certainly already has to adjust for things like that. You'd also have a magnificent sample size to play around in.

petercooper 17y ago

What an attitude the author of this piece has. He keeps referring to "pseudo-science" and uses scare quotes all over. Campbell's chooses "neuromarketing" over consumer feedback in rebranding its iconic soup cans. Of course it does. It's trying to sell as many cans of soup as possible, not please a focus group. Implement what ultimately works, not just what customers request. Of course, this smacks of the same pseudo-scientific marketing bullshit Is this guy totally unaware that there is an art to persuasion (some of it grounded in results from scientific studies, though not all) and packaging? It's not like astrology. There are psychological tricks and shortcuts that are taken in most high end products and advertising to make them perform better.

mbreese 17y ago
They used a sample size of 40 to determine the packaging for a multi-million (billion?) dollar company. Even if they used every possible technique to capture the "emotional attachment" to a label, it wouldn't be a large enough sample size to be statistically significant. If you're going to go to the trouble to run the tests, at least use a large sample size. There is no way they could have gauged the entire soup buying population with just 40 people. Like what was mentioned elsewhere in this thread: they would have been better off shipping 50 redesigns to various places and seen how they did in the marketplace.