How to develop billion-dollar product opportunities

Internet of Things and Big Data have all changed manufacturing and marketing functions. In manufacturing, we talk a lot about industry 4.0, predictive maintenance and a lot of predictions and prescriptive functions. But the question I asked myself about two years ago was: What about research and development in the engineering function? This is the I will be exploring in this webinar.

The webinar is based on a new Masterclass called Leading Intelligent Engineering, which demonstrates how to reduce cost, market faster and increase productivity in product development using new technologies and has already helped over 100 heads of R&D departments and CEOs of major Fortune 500 companies.

In product development, we always follow similar process and stages to identify opportunities for innovation followed by ideation and concept. Normally we have a big steering committee that is electing a concept that you then put into design and development. Across this process we had different understandings how to do these and different stages more efficiently. Normally we do it in a kind of stage gate logic, so one big stream of research and also practice in the last decade or so was open innovation.  We said ‘if you want to be successful in R&D and engineering you don't do it alone’. We integrated external partners, employees, customers, suppliers, research institutes.  But we realised that we now have the ability to really look into data sources, not just a few scientific applications, scanning millions of documents, corporate publications, research or social networks. There’s been a lot of discussion recently about how we now have higher processing power to run these algorithms but the real core trend behind all of this is that we have lots of connected data that was not accessible before and we can make connections between those

A successful example of this data-driven innovation comes from Nivea, one of the most known consumer brands on earth, and their aim to innovate in the mature, deodorant category.  Traditionally, we would carry out market research, ask questions to a selected audience and try to draw conclusions. But Nivea wanted to know what customers are really saying about deodorants and what are their real needs so we turned to blogs, forums and social media. We looked into a vast amount of data, not just a selected number of people in a focus group.

Nivea data-driven product development

When we really looked into these communications, conversations between consumers, we found out what people were actually talking about, and it was nothing like what was presented as a result of a market research carried out prior. One problem that was discovered as a result of our data processing was an issue of deodorant residue that’s difficult to remove. Someone on the forum said they are creating their own special washing powder to get rid of the residue. We saw a big open market opportunity. We actually found 99 self-made recipes by professional dancers, athletes and others. The resulting product was the most successful product introduction in the history of the brand and the most successful launch in the deodorant market. This is just the starting point of how digitalisation is changing the R&D function.

To get a framework allowing you to leverage digitalisation, big data and AI/ML to change and empower the research & development and engineering function, watch Frank’s Piller webinar.

The foundation of the framework consists of the innovation process, supported by the so-called digital twin or digital shadow. If you know anything about Industry 4.0 you will be familiar with those terms. You’ll often find them in modern PLM systems.  The former is an as exact as possible representation of the assets and infrastructure of the factory, see it as a planning and simulation environment that is equal to the current state of your factory. For operational decision making, however, this data is too large, even with today’s technologies. This is where the concept of the digital shadow comes in, a reduced, task-specific data set for real-time decision making (predictions or prescriptions).

Think of Google Maps as an example. Google Maps is a simulation of reality in the form of a map, much more precise than a paper-based map. Google Maps also has live traffic information showing you the utilisation of the road system. In Germany, data is gathered every five minutes, using the location data of cell phones to calculate the movements on the street. Processing this data in real time would be incredibly difficult, so we get a digital shadow of the utilization of the road infrastructure, reducing data to what we really can process.

The framework identified consists of four areas, structured along the new product development process, where digitalisation and machine learning are changing the R&D function. The first we call predictive innovation, with regards to analysing big data and opportunity recognition. At the concept ideation stage, we can turn to collaborative frontend, having algorithms that come up with new designs at the design & development stage that have very different approach to technical problem solving. Finally if we produce connected product we also have to think about smart product ecosystems.

A great example of the first stage in the framework, predictive innovation, is a company called Choosy, a fashion brand that creates outfits discovered by an algorithm finding the top trending fashion on Instagram. It delivers algorithmically informed fashion in as little as two weeks. The team creates small batches of the first crop of styles in-house. If an item proves popular, the manufacturing is outsourced to another nearby clothing factory in order to meet demand.

The AI predictions in this case are all based on connectivity. We get these big data sources like Instagram and then we have artificial intelligence as a means to explore the automation of these data sources in making recommendations. For more information on this approach, you can refer to “Identification of the to-be-improved product features based on online reviews for product redesign” and “Mining and Summarising Customer Reviews” articles. Crowdfunding is another great source of predicting the success of product concepts. To summarise, predicitive innovation is about finding innovative patterns in connected sources of big data using advanced analytics with AI/ML.

Second stage of the aforementioned framework, collaborative frontend, relies on the ability of online platforms to connect different owners of knowledge. This has been done in the consumer market under the names of crowdsourcing and co-creation. In manufacturing, it’s about open innovation platforms. Innovation is really a disciplined problem solving activity. First, we search based on prior experience and then experiment for the purpose of trial and error learning, using knowledge and creativity in the process. However, local search bias and past dependencies come in the way. Platforms such as Experfy help you find algorithms for all kind of business applications. You can also broadcast your problem on a crowdsourcing platform such as NineSigma to a large network of solvers who then submit their solution proposals.

Open innovation process

We can define open innovation then as a utilisation of external concepts and technologies from unobvious others as an input for the innovation process. It involves utilising the opportunities of digital connectivity and collaboration to connect with unobvious others in the form of customers, users and technology providers.

The third pillar, digital (AI) design and engineering, relies on generative design algorithms, virtual prototyping and model-based simulations along with robotic experimentation to enhance experimentation capacities.  One example of the generative design system comes from aerospace industry, where a plane manufacturer was able to design a new partition for planes quicker than normally and reduce weight of the partition by 45% while still fulfilling structural property constraints. In this scenario, AI augmented human designers by calculating ideal solutions for given design goals and constraints. Another example on utilising the benefits of 3D printing can be found in another article on The Leadership Network’s blog. Digital design and engineering, essentially complement human problem solving skills with machine intelligence and automation capabilities. 

The final pillar, smart product ecosystems, involves smart products that lead to digital experiences.  It involves value creation and capture from digital services and experiences, based on connected products and delivered via open platforms. A simple example comes from Tasty, which distils complex recipes into bite-size video tutorials. I explore this topic in more detail in my second webinar, which you can watch here.

The product design really has changed. The usual innovation process for traditional products would stop upon launch. For smart and connected products, upon launch is where it really begins, with services making sense of the data generated through product, enabling the solution ecosystem to create sustainable innovation.

digital business model

A new innovation system brings innovation in the usage cycle. Such process calls for a digital business model, which involves the following: smart products and services (value creation, but also maintenance of the digital twin/shadow) and smart experiences, enabled by data insights generated.

The question nowadays is not about value creation but value capture. How do you position your product in this ecosystem to make sure you capture the value, not Facebook or Google?

Share this article on your TLN® Profile
Share this article