Know-How Exchange : Research/Metrics

Why not run an A/B split test on your landing page (assuming that's your likely sales point, instead of a chain store)?

Head to head seems a good place to start. Consider using Social media, PPC, and e-mail to drive traffic to a landing page. Offer discount coupons or rebates as incentives. Run the same tests on Facebook via PPC. Then regroup. You'd also do well to connect with MONMARK (Randall Montalbano on this forum). Randall has MASSES of retail knowledge. https://www.marketingprofs.com/ea/profile.asp?userid=279897

I've done tests like this by creating a simple concept statement (for the product/benefit) and then exposing the concept to 3 different groups in the target audience. The only difference across groups is the package shown along with the concept statement. One gets the current package, one gets new package #1, and the third group gets new package #2.

We ask a few questions about the appeal of the concept and then a purchase intent question. The idea is to see if one (or both) of the new package versions can outperform the old package (without specifically calling attention to the package depiction).

To see a statistically significant difference, if there is one, you'll probably need 100 respondents for each version. But if there is no significant difference, that would be good to know too.

Consumers are not very good at expressing the effect of a package change when asked directly, which is why we do this monadically and without direct reference to the package itself. It's not that they don't want to be helpful. It's that they really don't know until they are faced with making a real-world choice. Their expressed purchase intent is about as close as you'll get.

Of course, it's important to select the right respondents. It's possible your current customers will prefer the current/familiar package, while a non-customer will prefer one of the new ones. If that's the case, then the research results will be largely determined by the make-up of your respondent sample.

You might want to hire a market research professional for a brief consultation to help you get at the issue before you decide with the flip of a coin. There are a number of possibilities, each with pros and cons. It's probably worth a small fee to make sure you're considering the right options. (Let me know if you need a referral.)

As Jay recommends a split test ('cross-sectional design' in experiments') and Gary head-to-head ('within-subjects' design), why not do both? They each have different advantages:

1) in a split test, customers are exposed to just 1 package, as they will be after your decision. Package is just one of the many factors they use to come at their choice. This is more natural, i.e. similar to the actual purchase process
2) in a head-to-head test, customers see all proposed designs and thus (a) can compare and (b) won't decide against purchasing because of 1 design they hate.

Michael is 100% right that it is important to consider who the respondents should be. Yes, it is easier to ask past buyers (as you have their email addresses), but that is insufficient if you also want to reach new customers with the new package design.

Good luck!