Consider this: It’s Monday. You go into your favorite store after seeing an ad for a new brand of cookies that’s on sale Just Today! You get to that aisle, decide which flavor to buy, pay, and leave.
Now it’s Friday. You’re back in the store – you want more of that tasty new treat! You go to the shelf, but that new product ISN’T THERE. You’re confused. You leave without purchasing. What was different? The cookies were there, but their packaging changed, and you didn’t recognize them.
People are wired to make comparisons, to find the difference vs. what they just saw. But if you design your shopper research with this model, you risk getting answers as clear as mud.
What we’ve just described is what researchers call a “sequential monadic” methodology. It can be cheaper to execute than the industry-recommended “monadic” (meaning “only one”) approach, because the same respondents react to more than one concept, reducing sample cost. But for many retailers and manufacturers, lost sales as the cost of bad decisions is a far greater hit to the bottom line.
Let’s look at another simple example: A manufacturer is trying to determine if their exciting new brand will generate incremental sales, and, if so, whether its package should be green or gray. It will not get space on shelf unless it can provide incrementality.
In a sequential monadic design, the first shopper sees and shops Shelf A with the new brand in green, then shops Shelf B, which is identical, except now the new brand is in gray. Since Shelf A and B are randomly assigned, the next shopper will see Shelf B first followed by A. Still with me?
Let’s say the sales results from this test are as follows:
20% Bought Brand D in Shelf A when Shelf A is shown first
30% Bought Brand D in Shelf A when Shelf B is shown first
Combined: Brand D in green is 25%
12% Bought Brand D in Shelf B when Shelf B is shown first
18% Bought Brand D in Shelf B when Shelf A is shown first
Combined: Brand D in gray is 15%
Because the data is inconsistent for a given package depending on if they saw it first or second, it is hard to use the data except for when it was shown first. So the effective sample size is cut in half for this study and the estimate of brand D selling 20% in Shelf A is now not significantly higher than the 12% in Shelf B; mainly because of the lower sample size. Therefore, this tells us it should not be launched.
In a monadic design, each shopper sees only one shelf, either A or B above:
20% Bought Brand D in green
12% Bought Brand D in gray
In this scenario, the new brand will sell at 20% in Shelf A and 12% in Shelf B. With double the sample size it is now a significant difference and the green packaging should be launched.
At InContext, we realized the inconsistency of testing with sequential monadic methods. It increases the risk of bad decisions. Because VR testing costs less to implement and takes less time, we are easily able to offer monadic testing for our clients, which allows for an unbiased, more accurate read on whether or not a new product has the potential to increase incremental sales.