Shaping Pack Redesign with AI Insights: Kellanova Case
Editor's note: This article is an automated speech-to-text transcription, edited lightly for clarity.
EyeSee was one of the organizations that sponsored a session during the Consumer Reset series part of Quirk’s Virtual.
Nicole Tudosie, business development director, EyeSee and Simon Pollock, senior director insights and analytics, Kellanova teamed up to share the process of Cheez-It's pack redesign for its U.K. product launch.
Learn how predictive eye-tracking made all the difference in this session.
Session transcript
Joe Rydholm
Hi everybody, and welcome to our session ‘Shaping Pack Redesign with AI Insights: A Kellanova Case.’
I'm Quirks Editor, Joe Rydholm. And before we get started, let's quickly go over the ways you can participate in today's discussion. You can use a chat tab to interact with other attendees during the session. There won't be a live Q&A today, but you can use the Q&A tab to submit questions for the presenters and they'll be answered later via e-mail.
Our session today is presented by EyeSee. Enjoy the presentation!
Nicole Tudosie
Hello everyone and thank you for joining us. My name is Nicole Tudosie, and today we'll be discussing the many challenges of pack redesign and innovation.
Before we dive in, I'd like to give a quick shout out to Quirk’s for hosting this event. If you have any questions along the way, feel free to put them in the chat. Our organizers will collect them, and we'll be sure to address them during the follow-up.
But first things first, it's my pleasure to introduce today's co-host, Simon Pollock, who is senior insights director at Kellanova. Simon, could you share a bit about yourself?
Simon Pollock
Yeah, Nicole, thanks so much for inviting me to talk today about packaging. I really believe packaging is the first point of interaction many consumers have with our products, and because of that, it's a critically important touchpoint.
I've worked in the packaged food industry for most of my career, working across lots of geographies and almost all food categories in the supermarket. And with that I've learned that that interaction and packaging, that first point of touch is critically important.
One of the key learnings I've had throughout this time is the importance of simplicity and knowing that consumers, particularly at the point of purchase when they are in their shopper mindset, only really see very few items on pack. So, getting clarity and consistency on pack is really important.
It would be wrong of me not to mention packaging like Kellogg's Packaging in store. I didn't work on that, but I really, really liked the way that team have taken simplicity and really focused on the core distinctive assets of a brand to communicate their message.
Nicole, I'd love to learn a little bit about you as well.
Nicole Tudosie
Sure. I am the business development director at EyeSee and I specialize in helping FMCG and CPG brands like Kellanova, leverage behavioral insights that will improve the in-store and online engagement and visibility.
I specifically focus on understanding the client needs and seeing things from their perspective, which is why I want to start the webinar with asking you, Simon, what do you think are the biggest challenges in the pack redesign process?
Simon Pollock
Packaging is such an important touchpoint. It is usually the single touchpoint that consumers get to our brand and to our products.
As a result of that, it's critically important that we know that we're creating both the visibility on shelf, the findability in store, but also the usability when they get it home. And because of that and because of the importance of that, your traditional research and learning and investment made in research and learning can often take time and cost significant money. It's really important that we invest to do the learning, but what it means is often we only learn at certain points.
So, one of the challenges or one of the biggest challenges is ensuring that we have iterative learning or ongoing learning. Ensuring that we're really focusing on the behavioral aspect of packaging, what it functionally needs to do both in store and in home. And allowing us to do that at speed and pace so that we can get the products to market more quickly.
So, that for me would be the biggest challenge, the agility to be able to move quicker throughout the packaging development process.
Nicole Tudosie
And I couldn't agree more with that, Simon. The lack of agility is the most common frustration I keep hearing from consumer insights teams.
The market research industry is catching up where traditional pack validation tests now have complimentary testing solutions. These are tech enabled to meet the pressure of innovation.
For example, with the launch of our predictive eye-tracking, this has allowed us to offer a pack test and protocol that leverages AI to get quick cost efficient and practical optimizations before and after you validate the pack designs on shelf.
This is an overview of our pack testing protocol where we start with the pack screening element. This is where you have multiple designs that you need to screen out, but also what we do is we help with pinpointing the areas of optimization.
In the middle, there's pack validation. This is probably the most traditional phase that you are familiar with. This is on shelf, your testing, two or three winning designs against the current one on the competitive shelf.
Finally, once you've had your recommendations on how to further optimize the pack designs, we go into the last phase, which is a pack check. This is a very quick and efficient behavioral check that the changes that you've made are doing the job that they were meant to do.
Simon, can you tell us a bit about why the Cheez-It redesign happened and what were the goals for the brand globally?
Simon Pollock
Yeah, it's such a fun brand to be working on.
Cheez-It is a U.S. cracker brand that has existed for many, many years in the U.S. and is highly, highly established and has multiple product forms and ranges. We were looking to launch that brand into the UK. And by the way, if you're in the UK or Ireland, you go into store, you will see it in there now and you'll see the packaging.
We were looking to launch it in the UK, but obviously the challenge there is the brand was not known. It was only one format, and we really had to establish the brand identity and presence by using those distinctive assets. That was the story behind the redesign, taking the best of what we had in the U.S. and finding an appropriate solution for the UK market.
That's when we partnered with EyeSee to work with them on helping us get to that best design.
Nicole Tudosie
We were absolutely thrilled to support the Kellanova teams with this strategic redesign.
Where we started this process or phase one was actually the pack validation test, the middle pillar that I showed earlier. And this is because we already had from previous testing some very strong winning designs.
So, the two winning designs were what we actually wanted to put on shelf in a competitive environment and test the visibility and the purchase from there.
Simon Pollock
We weren't starting from zero on this design. We actually had the foundation, we knew the branding, we knew the distinctive assets. And so that allowed us to explore different ways to bring that to life, to deliver against the objective of this UK challenge of launching a brand that wasn't known by consumers.
And the great thing about partnering with EyeSee is they had a suite of tools that helped us do that. And pack validation was the one that we started with on this particular journey.
Nicole Tudosie
So, how did we do it?
Well, we used real eye-tracking to determine whether the new pack designs have visibility and if they hold attention on shelf. We also used real eye-tracking on the key elements or the key areas on the pack design itself and how the eye visually reads them. We used virtual shopping or behavioral buy-in to determine if consumers would actually purchase it from the shelf.
So, beyond percentage bought, we also look at penetration share, market share and value share. And that's because we typically use a higher sample size of around 300 respondents per sale. This is validated within market performance.
We use interaction tracking. We use that for both quantifying likes, dislikes and open-ended answers around those likes and dislikes.
We also use it for findability on shelf. So, this is time to findability and it's another layer to tell us whether the new design stands out. So, how quickly can consumers find it on shelf, how successfully, but also what they actually confuse it with and get it wrong.
Finally, RTM and survey. RTM stands for Reaction Time Measurement. You might also know it as implicit associations. Essentially it monitors how quickly the respondent will agree or disagree with the benefit or claim, which constitutes for an implicit response.
Obviously, survey where we will track the typical survey brand fair like ability, uniqueness and so on, which was also fully customizable.
Simon Pollock
And it was because of the combination of methods that allowed us to understand how we could get to a design that would work most effectively, both on shelf with the greater visibility, but also the messaging relating to the brand.
As I mentioned, we had an established brand. We were looking at what were the distinctive assets that we could bring to market in the UK. The other thing I'd point out at this point is the brand also existed in a couple other markets. It's in Brazil and Canada. And so, we were also wanting to make sure through this test that we were developing something that would work in those other markets as well.
We got some really clear guidance on how to adjust some of the packaging and product. And also, were able to incorporate that learning into the final design. And through that, we were able to make the optimizations.
What we did, which was a great benefit, is we started to leverage the AI to retest.
Instead of running a full test again in market, we're able to use the learning from the first test, use the AI method that Nicole talked about, to actually validate if we successfully delivered the changes and did we improve the packaging.
I've worked on many projects where you make changes to packaging after you do the learning. And what ends up happening is sometimes you make it go backwards, not forwards. So, the great thing about this is we had a method that allowed us to validate or double check that we still wanted to take the design forward.
Nicole Tudosie
Echoing what Simon was saying, once the pack design is improved, it is essential to verify if those iterations are effective and they're doing the job that they need to do. However, often time and budget constraints make it difficult to conduct another large-scale research like pack validation.
To address this challenge, the pack check solution was created.
This is leveraging AI-powered predictive eye-tracking in combination with other behavioral and explicit methods to deliver fast and efficient insights to make sure your final pack is fully perfected.
In terms of the method used, we had predictive eye-tracking, interaction tracking to capture likes, dislikes and the reasons behind that. And then also reaction time measurement and survey.
Simon Pollock
Yeah, so as Nicole mentioned, we used the predictive eye-tracking element as a way of checking our packs and the changes in the optimizations we made based on the learning in the first round.
Predictive eye-tracking itself, AI eye-tracking is not new, but what was new here, and what EyeSee was able to provide, was linking it back to their benchmarks but also linking it back to the previous tests so that we knew whether we had made improvements on the visibility on shelf.
Alongside that, because we'd also combined that with a consumer survey element, we were able to understand ‘have we made optimizations on the likability,’ ‘have we made improvements on the attributes that we were looking at improving?’ Also,have we been able to flow that through into how the product performed not only on shelf, but when they took the product home?’
Finally, so you'll see if you go into a store in the UK right now, you will see that Cheez-It is in store, and you will see the pack. I can't reveal any of the changes or tweaks we are making obviously, but as new products come to market, you'll see some of those new images and some of that learning coming to market in the coming months.
Nicole Tudosie
As Simon mentioned, predictive eye-tracking is not new and has been around in the market for a while. What is new with EyeSee’s predictive eye-tracking that goes beyond saliency.
So, where traditional models will look at the contrast, the shape and the color, and it will assume that that is where naturally the eye will go. What our predictive eye-tracking does, is it uses advanced AI algorithm that is based on real eye-tracking gazes of respondents that we've been capturing for the last decade or so.
This AI powered solution requires no respondents and it's highly accurate as you can see from the two images on screen where we are comparing real eye-tracking with predictive eye-tracking. In terms of outputs, we will get AI visibility, attention and heat maps of the key areas on the pack.
And what's even more exciting is that we'll also unveil a brand new additional output that is showing the order in which each pack element is seen, which is designed ultimately to help shape the pack architecture.
However, AI is not a substitute for real eye-tracking. It is complimentary because they give us different levels of depth. While real eye-tracking works best when combined with virtual shopping, providing the full shelf context and even deeper insights.
With over 80 pre-made virtual environments in both 2.5 D and 3D options available, the possibilities are countless. However, if you are looking to make quick and robust decisions, either pre-validation, so where you've got multitude of pack designs that you need to optimize and narrow down, or if you're looking to do a final post validation check, then predictive eye-tracking lends itself well.
Simon Pollock
Thanks Nicole. It has been a pleasure working with the EyeSee team on this project that we talked about. A couple of key highlights and learnings from us.
As I mentioned, one of the biggest challenges in packaging research is the fact that it is highly involved, and particularly as you do in context learning in a shelf context, in a home context, it can be quite expensive and slow but it's really important. So, we should continue to do that, but what that means is we often don't have the agility to retest or recheck things.
What was great about this project is we were able to use the AI and the predictive tracking to go back and retest post-op optimization to see if we'd actually made improvements. The good news is we had made improvements, and we believe we have a better pack in store as a result of that.
So, that's really fantastic. I'm really excited about this capability because it not only gives us the ability to do predictive eye-tracking on packs, but it also provides us with the benchmarks and then link back to the original test, which I thought was really exciting and that's been one of the gaps in this space for quite some time.
The one thing I'm really excited about is using this type of capability pre-consumer learning, so using it in the development process and understanding how we can improve the designs before we even get to a consumer test. We're currently working on one of those now and I'm excited to see how that continues to track.
So, for me, I just want to say a big thanks to EyeSee for having me today and also to the Quirk’s Media team for hosting the webinar. I really appreciate it and happy to chat about this with anybody at any time. Thanks so much.
Nicole Tudosie
That's fantastic. And with that we'll wrap things up.
Thank you again, Simon, for being here and for sharing your expertise with us. And a big thank you to Quirk’s and to everyone who joined us today. Your time and questions are greatly appreciated. Be sure to follow Simon, EyeSee and myself on LinkedIn to stay up to date with the latest trends in market research. Until next time, take care.