So let's talk a little bit about organizing for innovation. Let's compare and contrast two leading innovative companies, Apple and Google. Their innovation processes are actually quite different. Apple is all about design, and at least the way it's often portrayed, they'll have a small design team working together to come out with new products and services. Google is perhaps the ultimate what we call skunk works here. They encourage diverse innovation, lots of effort in a lot of different directions, which is not necessarily very controlled and more free formed perhaps. They're famous, for example, of giving their employees one day a week to work on any project they choose that might be a future innovative opportunity for the company. So what's way, which way is the best? Well, there is not single best way to organize. Each way has its own pros and cons. And really the best way might depend on both the competition, the market you're in, and the underlying capabilities that you have as an organization. Whether you want to be quick to market with the complexity of the products that you're trying to devise. All of these might impact the way that you best organize. To help us think through this challenge, I want to talk about four different kind of generic ways we can think about organizing for innovation. So the first is what we might call a chaotic or a unidivisional structure here. This is in essence the idea that there isn't much structure to innovation. You generate an idea, you act, this is the skunk works idea, characterized by disorder and conflict and lots of things kind of bubbling up within the organization. When we think of the pros and cons of such a kind of loosely structured approach, it could be fast in some instances because there's little bureaucracy slowing down innovation. It could also be slow in the sense you never know when the new big idea's going to emerge. For many startups, this is really the only choice for you. You don't have the large set of organizational structures that would allow you to do maybe a more sequential process that we'll talk about in a second here. You're just a small team, you get together, you figure things out. Arguably, such a less structured role leads to maybe more creative solutions, perhaps even more breakthrough products. The downside, of course, is it's very difficult to control, and at the end of the day can be relatively inefficient and might lead to much conflict when you're trying to decide all right, what are the bets we ultimately wanna place? We have so many different ideas being bandied about. On kind of the opposite end of the spectrum, we have very sequential processes that some organizations pursue. One could imagine a functional organization where initial ideas are then passed on to an engineering design team, who are eventually given to manufacturing, and then eventually distribution and sales, and so forth. At each of these different stages, you could have a go/no-go decision, you could have processes in place to approve. And it doesn't necessarily have to be completely sequential, there could be possibilities for iteration and looping as you go through such a process. Now, the pros of such an approach is you have some structure, you ensure that each step is well executed, you can manage such a process. When high uncertainty is about a project, you can see this in a lot of science-based organizations, this is helpful to make sure you're making progress, and making sure that if things aren't progressing, that you can cut it off. Good when lots of time is available. So when you see things like the pharmaceutical industry and the process they go through in drug discovery and development, they'll have such a stage gate process. And it's also good when the market environment is relatively stable, because then you have time to get to the market opportunity that might present itself. The downside is, it might be bad for speed and responsiveness if the market is continually in flux, such a sequential process might delay your entry with innovation. And it's also potentially gonna be bad for accumulated learning potentially here because you have such a siloed process from one function to the other. So as a improvement on these two types of organizational structures we can think of a concurrent design, where we actually have cross-functional teams that overlap one another. So now we have our manufacturing team working with our design team here. All these functioning, working together over the entirety of the project here. Now, this helps with learning, with improving the team, and the firm coming together to discuss information as they go. There are classic examples of sequential processes in which a decision made by an early phase has huge implications downstream that weren't considered. Perhaps it was an auto company designing a car not realizing that for the manufactures, they were gonna raise the prices significantly to be able to manufacture a special tail fin or a special feature of the car. This type of process, where it's concurrent and cross-functional, tends to be best for the evolving kind of incremental products, where you have products that you're constantly updating. It does require a big time commitment from the team. It's very front-loaded, so it's very resource intensive, cuz you have everyone working together, and some might argue it's not a way for a home run strategy here that you might get out of a more chaotic process if you're looking for a really radical innovation. Last, but not least, we can think of another kind of hybrid here which is an overlapping matrix structure, so you still have functions, they're still developing products and services in a stage gate process, however, there's overlap between them. So they don't have this what we might call in engineering a kind of throw it over the wall mentality, that each division finishes their part and just throws it over the wall to the next. This can be very motivating to the team, it limits some of these hand off problems, but once again it requires a large amount of effort to implement more management, more time by everyone having meetings and the like, to do such a structure. So as a summary here, I would just summarize by saying firstly, there seems to be a natural tendency towards the chaotic, especially if there are no functional areas. Again, if you're a smaller organization, innovation often has that type of feel. And as you grow and become perhaps more bureaucratic, a sequential process often takes over when it comes to innovation. Overlapping and concurrent are basically two solutions that are limits of these other two approaches, but recognize they require more management attention and can in some ways be more costly. In general, one might expect that as the more complex the product, perhaps more technologically grounded in science, or basic science it is, the more it's gonna cost to fix downstream mistakes, and therefore the more emphasis on organizing a process that's gonna minimize those downstream mistakes by addressing them earlier in the design process.
