Same thinking – same result. Fix it with Innovation Feedback Loops.
The dilemma of Innovation Feedback Loops
Innovation Feedback Loops are needed but not easy. The hardest part of working with innovation is reimagining an old problem or discovering a new one. This also makes Innovation Feedback Loops very difficult to create. Anthropological studies, business model canvases, value propositions, jobs to be done, the customer journey, design thinking, and many other techniques have been used for decades. Now Design Thinking has evolved to Design Thinking 2.0. All of these are great and fantastic frameworks, but the fundamental problem remains. How do you institutionalize the behavior to get the insights? From discovery to commercialization? You need feedback, but there is no accessible feedback. This is the dilemma.
Relying on super teams is not scalable; it is not sustainable, and it is very inefficient in terms of cost and scaling up. Building the capability and structure, while fostering a culture of constant innovation from improvement to rethinking is on the other hand less risky, requires less working capital, and increases the possibility that you will see the oncoming S-Curves before they hit you. Moreover, you will be able to launch a stream of great improvements in everything from processes and business models to services and products.
In our experience, firstly, feedback loops are at the core of the solution. And secondly, building up the funnel of ideas in the discovery phase. In this post, we will focus on innovation feedback loops.
Pull and push innovation
There are two fundamental aspects of innovation feedback: push and pull. A push is when you launch something and study the reaction. Pull innovation occurs when you ask for feedback, either through surveys, focus groups or through anthropological studies and analyze the data to uncover behavior.
Example of pull and Innovation Feedback Loops
Using big data from retail and eCommerce transactions to find common patterns, e.g. checking what items are bought together, is a simple and straightforward type of analysis, although one that is rarely done systematically.
Another form of pull is performing correlation analysis between Google Trends and buying patterns, e.g. against in a specific region by a specific demographic group at a specific time (evening, night, or day of the week).
Other straightforward analysis methods that can be linked to buying patterns are, face recognition, voice recognition, and text sentiment. But do not believe for a second that analyzing big data without understanding and being interested in human behavior will help. You will find meaningless patterns and potentially ignore many useful patterns that might not be strong (as right now, we are just in the discovery phase) but holds high potential.
You must think like an anthropologist (or if you prefer to call it “an ethnographer”) and use big data as a tool, not as something that will determine the future will of humanity.
Example of push and Innovation Feedback Loops
Great examples of push innovation are when you are doing something active, when you are the subject of the story. For instance, if you shuffle products around on a web page, change a process in real time, post messages on social media, or launch a test like a concept car or a haute couture dress and then measure and analyze the reactions. We can use cameras to capture data and then analyze it with AI or statistics, but here also, an understanding of true human nature is essential.
Iterating over gates and milestones
From our data, we know that most decisions are made without the right input. Instead, they rely on assumptions or old data. The problem with old data is not the quality, it is the relevance, or lack thereof. How can you learn from experiments that were never made? Therefore, it is essential to iterate back and forth over decision points. Decision points such as gates or milestones. Iterate until the data have been thoroughly validated. This is against our instinct to get things done, but that instinct can lead us astray and create false certainty despite the uncertainty.
When a spike occurs in agile development, you need to be able to escalate and go back. This is because you cannot be agile in an uncertain situation. You must first define the problem. Spikes are also highly relevant in innovation management, as feedback can actually bring a totally new perspective to a situation. You need a formal mechanism for handling spikes that complements the iteration over decision points. This allows you to fine-tune the lessons learned and reduce the risk when moving ahead.
Being agile is not the same as being adaptable
During COVID-19, it has been obvious that agility is not the same as adaptability. Agility means moving fast within a known scope. Whereas adaptability means being able to change to deal with a completely new situation. If you are working with feedback, you must be adaptable so that you can pick out the insights and make the right decisions. An adaptive leader must be able to reimagine a situation and collaborate across the board. You must encourage failure and demand learning. At a higher level, you need business litmus paper to test everything: Are we doing the right things? Do we need to change, and if so: how?
Working with feedback loops call for capabilities and structure
If you want to work with feedback loops, you need to have the right capabilities. In addition to being able to manage uncertainty, business anthropology, iterating over decision points, and escalating spikes, we discovered a few specific capabilities from our data based on the largest study of innovation management and capability, which included over 5000 organizations in 105 countries.
• Design capabilities
As an organization, when designing, you need to be able to do so digitally (using software), intellectually (e.g., with a business model) and physically (a product), and then test the product systematically. This system and anthropology are the two most important tools for understanding future innovation and how you can impact it.
• Experimenting with hypotheses
We all learned about hypothesis testing in school, but most people do not use this method, as it is perceived as being too difficult and too time inefficient. But nothing could be farther from the truth; most of the time, going ahead with an untested hypothesis can result in real issues during the commercializing phase. The old truth that $1 dollar spent on the blueprint is $1000 saved in production is way more relevant today than it ever was. To make hypothesis testing work, you need to introduce formal processes, coaching, and also training, and you need to build a database of the results from your experiments. This requires organizational commitment, as it is not just an isolated task.
• Systematic A/B testing
Just like hypothesis testing, most organizations don’t conduct A/B testing. This violation of basic early-stage innovation is not neglected for intellectual reasons, but rather from a fear of failure-culture and poor relationships with stakeholders. Organizations are simply afraid of testing because of the very results it is put in place to discover. But with strong stakeholder management, supportive leadership, contingency plans, and a disciplined process, A/B testing has been proven to be the most efficient tool for validating the insights gained from feedback loops.
• Selecting vendors and partners during the development phase
It can be really hard to get feedback in the development phase. But there are ways of doing this and building up the right capabilities and structure. It is often a mistake to innovate with big customers, suppliers, or partners, as more often than not, they will either insist on owning all the results or will not provide you any feedback that will be useful beyond the constraints of your arrangements. Better then to, for innovation projects, vet the partners and vendors you use and proceed with lead innovator partners that you have a natural advantage over (by size or other reasons) and are therefore willing to take risks for you as they can’t afford to lose you as a customer.
• Piloting the commercialization
Oftentimes, the commercialization phase consists of a big bang. Simply because you want to make a huge impact on the market. However, a big bang can create problems in learning quickly, iterating the feedback, and adapting to the lessons learned. By using a pilot, especially in a smaller market using a dedicated team, you can avoid such issues. The pilot will also allowing you to scale faster with fewer issues.
A hotbed of related innovations
Right from the beginning, you should plan for the end of the life of your innovation. Having a master plan can help you to maintain your momentum while letting you react to feedback. Feedback from the launch, post-launch, and post-mortem. To quote Sun Tzu. A general who has fought thousands of battles in his tent will also win on the battlefield. Simply put: the more you prepare, the higher the chances are that you win.