Few companies have as strong a track record when it comes to open innovation and pilot testing as Minneapolis-based General Mills. The $17.6 billion food giant sometimes puts packages on a shelf without actual product inside, as a way of learning whether consumers will pick it up and take it to the cash register. Executives there talk about creating “lemonade stands” as a quick and cheap way to test new product concepts, instead of setting up a “big bang” product launch and hoping for the best.
And the company expects failure and learning to be part of the journey of any new product wending its way to a consumer’s kitchen.
“In the past,” says Jim Kirkwood, Chief Science and Technology Development Officer, “if you had a failure, it was over. But now, what we’ve done is actually built it into our process.” If you don’t fail at least once before a finished product winds up on store shelves, he adds, “you probably didn’t test enough.”
You can listen to the complete call, moderated by InnoLead Editor Scott Kirsner, or read an edited transcript. (We’ve also posted a two-minute excerpt in which Kirkwood discusses how General Mills has adapted the “lean startup” approach to its business.)
…
InnoLead: General Mills is obviously a company that has had an R&D function, and probably a very big R&D team, for a long time.
How do you think about innovation? Is R&D supposed to be the fountain of innovation? Is innovation something that sits alongside R&D? Talk a little bit about the relationship between those two things.
Jim Kirkwood: R&D in the past used to be where innovation came from. But we’ve learned that innovation is actually a team sport, and that it needs to be not only R&D, who has a great understanding of some of the technical needs, but it needs to be a broad connection with consumer insights, with marketing, with finance for business models.
As an example, we knew that there were some big problems with tortillas. When people eat tortillas, if you put lots of tomatoes and wet stuff in it, and you pick up a nice, rolled tortilla, and you try to take a bite, you often mess up your shirt.
We knew that there was a consumer job that needed to be done to try and make tortillas a little bit neater to eat. There’s a huge business model issue there, because tortillas are extremely cheap. There was a technical issue from an engineering and operation standpoint, on how we could change the way tortillas are made. We came together as a team and created the Stand ‘N Stuff Tortilla. It’s basically in the shape of a boat, which allows you to put everything in, and eat them without having it spill all over your shirt.
A very basic consumer need was identified through some of our marketing consumer insights, combined with the technology capability that was invented at R&D, combined with the ability to make it practically and economically through our engineering and operations world.
We ended up with a pretty huge success, especially in Europe. It’s really taking off right now.
Organizational Structure
InnoLead: That’s a great example. Can you talk about the organizational structure there, and where you sit? We’ve talked to Peter Erickson, who I think is SVP of innovation and quality for General Mills. Do you report up to Peter?
Kirkwood: Yes. I report to Peter. Our whole technology creation and connected innovation worlds report in to him. …
We also have what we call our “Strategy and Innovation Team”, which is a very small team that’s focused on helping all of us become consumer-first. [They do] a lot of work to ensure that we have interactions directly with consumers, no matter where you are in the organization. Whether you are a technician who’s working on the floor, whether you are a PhD scientist in fats and oils, or whether you are someone who’s working on line extensions. Having that opportunity to have direct connections with and conversation with consumers is extremely important.
Connecting with Consumers
InnoLead: Can you talk a little bit about what’s new on the consumer insights and the consumer research front? I imagine General Mills is one of these companies that probably was among the best in the world at doing consumer research with surveys or with focus groups, or maybe with going into the kitchen to watch how people prepare meals.
But what’s new in terms of trying to connect with consumers and understand them?
Kirkwood: You mentioned a little bit about going into kitchens. We’ve taken that true anthropological approach, where we actually spend time with consumers in their life. We follow them as they go shopping.
I had a really unique opportunity to spend time in the home of a consumer who makes less than $30,000 a year for the household. I went shopping with her. She had $20 to spend for a week, for a family of four. Watching … how she did that was extremely insightful for me, and taught me a lot about some of the products that we sometimes say are the “value products,” and so forth. We need to be very, very respectful of those products … because they are playing a very important role in those people’s lives.
We’ve [also] tried to connect consumers with some of our deeper scientists. We had an opportunity to bring in a group of people who are in that world of, “I hate grain. I don’t like grain. Wheat is bad. Wheat belly.” [There are all these] books that are out there today talking about how grain has been demonized. We brought a group of those people in to put them in direct connection to some of our deep carbohydrate and milling scientists so they could hear, know, and feel what people are saying.
We made sure that they didn’t try and refute them and say, “You’re wrong.” It was to listen so we could understand what the problems are that these people see with grains, and what can we do to change that conversation about grains.
That’s at the deep scientists’ level, not just the folks that are doing consumer work in marketing, consumer research, those kind of things.
Making the World Our Lab
InnoLead: I feel like you’ve been in the vanguard in terms of this idea of opening the doors and opening the windows of the company to ideas from outside. I don’t know if it’s Peter’s term or if he borrowed it from somewhere else — Peter Erickson — but he talks about, “At General Mills the lab used to be our world, and now the world is our lab.” I’ve heard that quote echoed by people at other
Kirkwood: Yes, I think it was probably about seven or eight years ago, there was a real revolution in General Mills where we went from being insular to that exact statement, which resonated with everyone.
Initially we had what we called internal entrepreneurs whose job was for each of the platforms, for each of the functional groups like the G-Tech organization, their job was to go out and connect. They had a second job, which was to teach everyone else in the organization how to connect as well.
Today, we don’t have any internal entrepreneurs anymore because we have made it everyone’s responsibility to get out and connect. We still have a small connected innovation group, because there’s some areas where we really need to do some advanced connecting, and I’ll talk about that in a second. (For more on connected innovation at General Mills, see this downloadable presentation from our Resource Center, or our 2014 interview with Mike Helser, who runs the General Mills Worldwide Innovation Network, G-Win.)
We need to have everyone, no matter what their position is, being willing to pick up the phone first and call to see if somebody else has solved the problem before we spend the money to go out and solve the problem.
An example is we had a big issue in soup, where people say it’s too salty. The problem is [that] if you take the salt out they won’t eat it, because it tastes really bad.
We had a big problem we needed to solve. That team connected with a three-party external ecosystem to borrow some discovery technology from another industry, which allowed us to be able to screen through millions, literally millions, of natural compounds to find compounds that might actually taste like salt but weren’t sodium.
We identified flavors that truly enhance salty taste and we’re … commercializing those to … significantly reduce the sodium in our soup products.
Collecting Dots and Connecting Dots
InnoLead: Can you talk about how that co-creation partnership came together? Are these small, medium, or large companies you are working with?
Kirkwood: They’re actually smaller companies that are in totally different industries. These folks were discovered on … tours that we go on. I call them dot collection tours and dot connection tours.
The collection tours are where you are specifically meeting with people who might be very, very different than who you are. They might be in a industry that’s completely different. They might be in the pharmaceutical industry, if you’re in food. We actually have talked to people at NASA.
All of a sudden, you identify something in that discussion that says, “Hey, we could work together on this using what you’ve traditionally used in your industry to solve a problem that’s in my industry.”
I want to own what we learn for certain aspects, but I can open that up for the other parties to be able to own it in other ways, so that we can create a win-win-win for everyone in those kind of situations.
InnoLead: Talk about what dot connection is.
Kirkwood: Dot collection [I sometimes call] kissing a few frogs, because you have no idea how [the ideas you find are] going to relate to your business.
Then, you might go on dot connection tours, where you know what you’re looking for, you know the people that are out there, and you’re saying, “I need to connect this dot, which I know you can do, to this dot, which I know [another company] can do, which will then bring value to all three of us, or all four of us.
You had mentioned big companies, too, which is interesting. We’ve done some work, even with other consumer packaged goods companies, where we in some areas compete, but [where we] are able to carve out and build a joint project where we develop something that’s meaningful to both of us.
An example is where we’ve worked with another CPG company to build the technology for a new-to-the-world package that is going to change the way many of our categories will be looked at, once we get [it commercialized.]
What it allows us to do is be able to invest in things that neither of us could have done alone. It also allows us to be able to leverage what we know about our categories, what they know about their categories, to be able to build something that’s better for both of us. One plus one really does equal three in these kind of situations.
Making Co-Creation Work for Everyone
InnoLead: One of the things, Jim, when you talk about collaborations and co-creation, whether it’s with startups or big companies, I’ve heard that a lot of people say, “Look, it’s great in concept, but in reality, you get the attorneys involved. You get compliance people involved. You get so many cooks in the kitchen that it can actually slow down innovation rather than speeding it up.”
Are there things that you do or things you’ve learned so it doesn’t turn into a five-year co-creation project, when it should be a five-month co-creation project?
Kirkwood: I think probably the first thing I would say is you have to have a trust relationship. The partners you work with have to be people whose values align with yours. If they don’t, you’ll be constantly fighting.
…We had one partnership where we were working with another large company. It was an extremely complex situation where they had a number of manufacturing facilities that could benefit from this technology. We had manufacturing that could benefit. There was a potential for some competition in a given area and so forth.
We sat down with the group, and we said, “We can either fight through the legal [aspects of] every eventuality of this over the next six to nine months and try and get to an agreement, or we can say we’re going to agree to agree, and we’re going to look at each other and trust each other to say we’re going to get to an agreement on the business model here, but we’re not going to be able to define that business model until we have defined the technology.”
“Let’s allow the technology teams to begin to collaborate. In parallel, as the technology teams learn, we will have a commercial team that will be looking at what is coming in and saying, “OK, how do we modify? How do we change? How do we structure an agreement between us from a business perspective to make it work?
“It is really essential that…you have values that align, but you agree, and you’ve built those relationships of trust so that you can do that. If you try to cross every T and dot every I before you do the work, you’ll never do the work.
‘Getting to the First Dollar Fast’
InnoLead: I have one more question, then I’d love to start working in some listener questions. My question is about what you guys call getting to the first dollar fast, and this idea of putting a product in front of consumers to see if they’ll buy it, versus testing things in labs or in a focus group room.
How do you do that? You guys are an industry where, obviously, there’s regulation. Obviously there’s quality and safety that’s important.
Kirkwood: We have really tried to put together some streamlined processes within our organization to be able to get to small [tests] quickly. We have what we call minimal viable product, the MVP, that we go for in this instance.
What do we need to put in this product to be able to get it out there as fast as we can and learn about it? That can be everything from a product that’s 80 percent, 90 percent of the way there, but we’re not quite right on the package.
We’ve actually gone into tests with things in a baggie — things that are just built upstairs in a grocery store, brought down, and brought to people. They would say, “Would you buy this? How much would you pay for it?” Those kind of things.
This minimal viable product can [be] everywhere from close to completely done to very, very far away from being done. As a matter of fact, in one situation we wanted to learn very quickly whether people would buy a given soup if it was in a box. We filled the boxes with water and put them on a shelf, and then watched people go and purchase them.
Then, obviously, we didn’t let them go home and eat the soup. We pulled them aside and said, “Why did you purchase this? What are the things that really drove your decision? What things on the package told you this is what you were interested in? Is it meeting the job you want?” We learned from that, before we had to invest a lot of money in getting the product where it needed to be.
Constructing Hypotheses and Opportunity Maps
InnoLead: That’s a great example. Our first listener question is a little bit more about consumer insight. How does General Mills connect the consumer insight marketing and technical teams? Is there a formal process? Informal? Do you have cross-functional teams?
Kirkwood: That’s a very good question, too. We have platforms.
InnoLead: What’s an example of a platform?
Kirkwood: We call it One Global Cereal, so anything and everything that has to do with cereal around the world.
InnoLead: Okay.
Kirkwood: We have teams that are made up of [multiple] functions, and we don’t necessarily do this work against a specific product. We do it against a specific set of jobs or requirements that we want to understand from our consumers.
…We first immerse in everything that we know from a literature standpoint about our category.
We’ve got years and years of cereal information. We do an exercise called “Positive Pickup,” where we all highlight information. We actually cut the pieces of highlighted paper out and then come together as a team and then group these, to see where all of the insights are that might come out of our past work.
Then we create some hypotheses [out of that about] the key major opportunities. What are the key main jobs that are not being done today by our products? We’ll take those jobs and we will create hypotheses around them and then go out and actually go into consumers’ homes together — marketing people, consumer insights people, R&D people.
It doesn’t matter what function you’re from. We’re one team. You go in and you really listen to what these consumers say. Are we right in those hypotheses? Are we wrong? What should we adjust? We come back and only then, after we have all that information, do we ideate around where are the opportunities for us.
And then we create what we call “opportunity maps,” which bring together the core key jobs that need to be done. Underneath those, [we put] the hypothesis on how we could deliver those jobs in products. Those maps are much broader than,” I need to put such-and-such a product in the marketplace today.”
We can draw from [the opportunity maps] over a period of time, and prioritize which ones make sense as we go forward.
Again, when these teams are working, you can’t tell whether they’re an R&D person, whether they are an insights person or whether they’re a marketer, because they’re working together to truly understand what these needs are.
Innovating the Business Model
InnoLead: The next listener question builds on that. Are there some examples of General Mills developing new business models or services, versus product-based innovation?
Kirkwood: Yeah. I think there’s a lot of opportunity to look at business models and how we do things differently. I’ll give an example that we did a few years ago, actually, but it gives a picture of it.
We recognized that Fruit Roll-Ups could be done differently than we do it today. We decided that we could actually print pictures, we could print words, we could print whatever on top of a Fruit Roll-Up and actually make it like a canvas.
We built a completely new business model, where we didn’t sell it from the grocery store. People would go online, [and] choose a message that they would want to have printed on their Roll-Ups, or a picture of their soccer logo or whatever for the game that they were going to have.
They could order a case of product. That case of product will be shipped to their home…
We tested that model and it was a very interesting model. It was a little bit before its time, because we did this back in 2005. I’m sure we’ll revisit the model. At that time, Amazon and some of the ways of delivering were not in the stage that they are today.
We developed the up-front part of this to work really, really well. The back part of it was a little bit more difficult to deal with, but as we look at how the world has changed today, that business model I’m sure will again become something that we would want to look at.
There’s two points here. One is the fact that we’ve looked at a different business model. The second is we know that there’s going to be times when things are going to be before their time. You don’t want to just throw them away. …Every time we go back and do one of these immersions, we look at some of those things to say, “Is the time right now?”
InnoLead: That’s an interesting point. I feel like a lot of times, there have been so many cycles of different innovation efforts and initiatives. The next one starts and everybody forgets the tests and learning and experiences of the last one. Big companies can be bad historians sometimes, when it comes to knowing what’s been tried before.
Kirkwood: Actually, to your point, one of the things that R&D is within General Mills is we are the historians, because many of the R&D folks have a long history in their product categories. People in consumer insights and in marketing often rotate more frequently than R&D. R&D is often really respected for being the ones that truly can bring that historical perspective.
The ‘Lemonade Stand’ Approach
InnoLead: Jim, I’m curious. You talked about getting to the market even with products that may not exist, but you’re testing packaging or you’re testing things in Ziploc bags. Can you talk about iteration and how you think about iterating? I feel like there are still a lot of companies where they feel like if a first test doesn’t work, it can often get killed and “Let’s move on to the next thing.”
Kirkwood: We do have a specific model we use. We recognize that with the “big bang approach” that we’ve used in the past, you have a hypothesis on what the product need is and you develop it and you build it and you test it in a standard consumer test. Then you launch it big, [and] you miss.
[We now test with what we] call “lemonade stands,” where we go out and we produce the first prototypes. We go and try and sell them. We peel people off who are willing to buy it, [and we] find what we’ve done wrong, just like entrepreneurs do.
Entrepreneurs have to do this because they are in their garage, but we try and put ourselves in those shoes. We learn from that first lemonade stand that, “Oh my gosh! The flavor of the granola bar should not be this way. This is not working.”
Then we can quickly make a turn, and we pivot, and we create, and we do another lemonade stand.
We might do five, six, seven lemonade stands to adjust the product offering and the concept all together while people are still putting dollars on the barrelhead. They’ll tell you in a focus group what they think they want. Then when it comes time to purchase something, their behavior does not match what they say. The closer you can get to the grocery store or the place where these people are purchasing it, the better information you’re going to get.
…In the past, if you had a failure, it was over. But now, what we’ve done is actually built it into our process. That iteration is something you need to do on the product path. You don’t do it on everything because in some of your categories, you know. If you know enough to be able to know, you can go out with a big launch and you can make sure that it’s not going to be that big of a risk from a dollar standpoint. If you have enough history, go for it.
Where your risk is high, we’ve learned and really created this process that says we are continually iterating on the small scale until we get large enough that we have most of these questions answered.
Failure and Metrics
InnoLead: So failure is going to be part of that process?
Kirkwood: Absolutely. You will fail. As a matter of fact, if you don’t fail, you probably didn’t test enough.
InnoLead: Let’s see if we can get one last listener question here. Can you talk about metrics you use for innovation, aside from revenue being generated by newly-launched products? What do your senior executives view as important?
Kirkwood: One of them that is most important to us is what we call “turns.” If you’re not in the food industry or retail industry, you might not know what that is, but that is the number of units that come off your shelf per week. It’s basically a rate of sale. We know that that is one of the top indicators of the sustainability of a new product in the marketplace. If your turns are low, if you’re not on top third when you start, it’s likely that you will not last more than a year or two at the most.
We look at turns as a leading indicator. We look at sustainability as the lagging indicator. How long is that product in the marketplace? Volume, of course, is important. Revenue is important, but if you have a sustaining, high-turning product, chances are over time, that will become a high-volume, high-revenue product as well.
Small Successes as a Tool for Culture Change
InnoLead: I’ll save the hardest question for last here. “Our company is just at the beginning of open innovation and thinking about how we scout for interesting startups and outside parties to collaborate with. Our management is not yet entirely on board. Any advice for us?”
Kirkwood: Pile up small wins — experiments where you can show success. First of all, define what your needs are. Be very focused on what your needs are. If you look for everything, you’ll find nothing. Go explore. Make as many connections as you can around that small, focused area and then prove by connecting with these folks and making something happen that it worked.
It’s amazing that within companies, success begets success. As you start to see positives happen, tell the story. Celebrate it. Get it in front of everyone.
That’s basically how it started at General Mills. A few people started to experiment with it. When those success stories became something that was highlighted in the quarterly meeting in front of all R&D, then everybody wanted to do it. That’s how you change the culture.
