How do you cope with uncertainty and ambiguity in data analytics innovation?

A few years ago, former GE Vice Chair Beth Comstock joined me on She Said/She Said Podcast to talk about her book "Imagine It Forward." One of the tactics she used for encouraging innovation on her teams was to give them permission slips to fail. Understanding that big ideas rarely if ever grow out of thinking "inside the box," she instead worked to create an environment where pushing the boundary in ways that would lead to failure before they led to success was a way to create this culture of innovation. Her perspective applies across the board. When we are too afraid of failure, we are simply ill prepared to take the risks needed to actually succeed in a bigger way.

Experimentation is a practice. Having a growth practice is how the best performing Tech businesses realize their big goals. In my study of 120 startups over the past year, only 1% of small and midsized organizations have a structured practice for cross functional experimentation, with product and marketing resources meeting every week to test, measure and iterate. Developing leadership skills, tools and mindset is how you solve this problem.

As a leader, it is important to keep repeating - Let’s try it and see what it leads to. We might fail but we will learn in the process. By doing so, you instantly provide psychological safety to your team to experiment and try how to interpret and leverage the directional data.

Totally adopt a growth mindset and be willing to question and analyse what is working. Give a new project some space to breathe before spiking it - but do be willing to get rid if it's really not working out.

Communicating technical processes is all well and good, but businesses will only have their apprehensions removed when they see data analytics platforms deliver actionable, deliverable insights. Organisations will only trust or use a solution they have invested in if they know they will deliver results. Of course, along with the solution, there must be a clear, concise explanation of how the result was achieved. In my experience, all uncertainty about a solution is removed when an organisation gives clear, concise explanations on how the result is obtained. Hence, it is best to perform operations within a solution scoping exercise to ensure that data analysts can give clear, but concise explanations on big data applications

😊😊😊 Managing expectations when it comes to DAI (Data Analytics Innovations) can be a challenging task for businesses. Not all startups that want to launch a DA platform understand how to do so effectively. Therefore, it's crucial to be aware of one of the fundamental challenges in DAI practices, which is the Alignment Problem. Brian Christian has produced an excellent book that details the complexities of using AI to navigate human values, which is a crucial aspect of building any DA platform. Moreover, to date, there is only one innovation stack that attempts to address building incorrect data models and provides recommendations for ethical AI and DAI performance.

Managing uncertainty in data analytics innovation involves embracing a culture of experimentation and accepting failure as a path to learning. Aligning stakeholder expectations, employing decision-making frameworks, and validating data are crucial steps. Constant learning and adaptability to new information are key. Guidance from mentors and insights from networks can provide valuable perspectives. Remember, innovation arises from understanding user needs. Trust your intuition when data is insufficient.

The biggest inherent risk in big data is over promising what your data can achieve. Data analytics is exceptionally powerful, and if used correctly it’s the perfect variable for decision making purposes. The most important thing to focus on is how powerful your data is, but also knowing the pitfalls and drawbacks of your data.

under promise and over deliver on every expectation. Set those expectations low and then deliver some stuff that they can have a feast on.

There have already been some good points on the importance of managing expectations, so to that I’ll only 1+ them. This can make or break you. What I’d add that hasn’t been said already is that part of managing expectations is explaining why the expectations are what they are. We often assume far too much about people’s understanding of data and analytics. Many think that because there’s data it should tell a comprehensive and conclusive story, which just isn’t the case. As such, when we set realistic expectations there is almost always immediate pushback. A way to mitigate that is to ensure you’re educating through stories that help illustrate the inherent variability and uncertainty.

The framework needs to be set upfront. It needs to be part of the overall picture for everything you do. You must set the tools that you are going to use and what you are going to look at in the end when all finished.

The VRIO analysis can also provide deep insight into whether your approach / product development tool for data analytics is Valuable, Rare, Inimitable and Organization (is it owned by the organization)? VRIO is great for validating new IP and ideas prior to launching to the market.

Leverage frameworks like SWOT analysis and tools such as Google Analytics for data-driven decision-making in digital marketing.

In addition to leveraging data and evidence, you should embrace a mindset of experimentation and learning. In data analytics innovation, uncertainty and ambiguity are often the norm rather than the exception. Therefore, always adopt a continuous improvement approach, where you test and iterate your ideas, products, and services based on feedback and metrics.

Evidence helps validate your data, which is extremely important. Validating a hypothesis gives customers the trust that your data is valuable and worthy of investment. I’ve learnt that not all data is what it seems, and with so much variability and ability to manipulate and skew data, trusting the source and validity of the data is critical. Data comes down to ethics. Ethically handling data and ensuring the data source is trustworthy and viable is essential in ensuring integrity of information between the customer and data provider.

Make informed decisions in digital marketing by leveraging data and evidence. Analyze metrics, customer feedback, and market trends to guide strategies.

In my experience, learning and adapting are important as the technology change has accelerated with a new game changing technology coming in every few months. This is due to the law of accelerated returns. But at the same time, the North Star should be kept in mind to stay focused.

Aprender y adaptarse a qué? A que el cambio es constante. A que la tolerancia a la incertidumbre se practica y se tolera bajo y con un liderazgo que de cuenta de las variables que puede y no manejar Aprender a comunicar lo incierto y adaptarse a la volatilidad no es cambiar todo el tiempo. Es planificar varios escenarios posibles para estar preparados a esos cambios.

Continuously learn and adapt in digital marketing. Stay updated on industry trends, analyze campaign performance, and adjust strategies accordingly.

Innovation is deeply contingent on user personas for data analytics tools. Often developers build for their own use, not actual users. As a former application architect this was one of my biggest lessons learned - create data tools for the user cohorts, not for myself!

Always begin with the why! What is "the thing?" What are you trying to solve. Treat the experimentation as if you would a project or an open innovation project. You have 6 weeks to get to an MVP to prove it out. After that, you may move on if there is value in doing so. Then evaluate if 1+1=3 Say your MVP did work, can you utilize what you have built to contribute data to another initiative? Did you prove you can capture weather & sensor data? Can this then be utilized for customer experience or an operating model change? Lastly, failed experiments are still so valuable. Often people may not know what they want, but they can tell you what they don't want. The failed experience produced a user design experience outcome.

According to Mark Twain there are three types of lies. Lies, Damned Lies, and Statistics. Consciously or unconsciously data can become biased or just plain wrong. Sometimes you really need to go with your intuition. In 1992 I was a speaker at the Earth Summit in Rio. One of the things that I've learned in following Al Gore, is that most people leave and you get to hang out and talk to Al Gore after talking to a relatively small audience. Not my best speech, but I still believe in the premise. Sometimes you can't gather all the information you need to make decisions. In the case of the environment, some cycles are thousands of year. You can't always wait for a pattern to take action. Trust your intuition and not just the data.

Data Science is science. Advancements and innovations are incremental. Hypothesis are unknowns you are trying to turn into knowns. Don't get too frustrated if your first few attempts don't resolve in groundbreaking findings. It takes a lot of experimenting to define new knowledge, but when you do -- it's a great accomplishment. Keep at it.

Define the problem you’re trying to solve or an opportunity you’re trying to unlock. Ambiguity and uncertainty in data analytics innovation appear more pronounced when a clear goal for its purpose hasn’t been established. Here are a few things I’ve found helpful: - identify if the purpose you’re innovating with data analytics is to solve a problem (need-based) or to unlock an opportunity (seed-based). - define success metrics based on the above. - explore the literature to understand what ambiguity and uncertainty mean in the context of your specific need (or seed) based innovation. Often, this will help resolve or provide further insight into the issue. -per the above insights, seek out real world data & experiment towards your goal.

As with any form of innovation, the focus must always be, and start from, the business value - what is the underlying problem we are trying to solve? If you keep this in mind, and look to make incremental progress that shows value quickly, you will get positive support from your business sponsors and it will enable you to keep aligned with what is needed. Building new data analytics platforms and innovations will only get buy in from the business / budget holder, if you can show value delivery quickly. I saw a talk by Microsoft about how they migrated there internal financial management reporting into real-time, this took time and require both process change and cultural change in addition to pure technical / data analytics innovation.