JWorks Tech Blog

October marked the 2020 edition of the BA & Beyond conference. As with all conferences being held in these COVID-ridden times, it was a virtual gathering and there were a variety of people from across the globe that came to share their insights, experiences and opinions on topics from their respective fields of knowledge. This ranged from agile practices to data science to process analysis and management. For me the workshop that struck a chord was given by Lori L. Silverman (Partners for Progress). She is a leading authority on decision management and its underlying data analysis techniques and an advocate for business storytelling. This workshop was titled “Facts Tell, Stories Sell”.

A New Approach

A typical data analysis project starts with a business sponsor handing a large blob of data to a team of data scientists, of whom he asks to make heads and tails about the numbers in front of them. The sponsor wants actionable insight on which he can act to enhance whatever goals and objectives the sponsor is chasing. What comes naturally to most data scientists is to employ the OSEMN framework. This framework consists of the following five steps:

• Obtain: Gather all accessible data from relevant data sources.
• Scrub: Clean up and formalize the retrieved data into a format that can be interpreted by a machine or data analyst.
• Explore: Detect significant patterns and trends within the available data sets.
• Model: Construct models to predict and forecast future data entries.
• iNterpret: Utilize the models to gain actionable insight. In essence we apply the models to detect good and bad cause-and-effect sets and use them to duplicate these results.

Lori has a SMARTER approach for this type of undertaking. This approach expands on the OSEMN steps. Where OSEMN just stops after handing over the conclusions of the analysis and hopefully providing the much-needed actionable insight, the SMARTER approach actually takes these insights and executes actions based on decisions based on these insights. This is very similar to the realization that DevOps brought to the development world. A project for the development of a solution doesn’t simply stop when the results are delivered but goes further with actually following up on the solution when it is running in production.

The steps that make up the SMARTER approach are the following:

• Seek Context: Detail the context in which the data science analysis originated.
• Manage the Data: Collect all relevant data and organize it in a useful structure.
• Assure Confidence: Clean up the data to increase the level of trust that can be had in said data.
• Reveal Insights: Determine the insights that can be extracted from the data at hand.
• Take a Stand: Formulate decisions on how to proceed with these insights.
• Execute Decision: Act on the decisions that have been agreed upon.
• Relay Results: Verify and report on the results of the actions to the different stakeholders.

Not only does the SMARTER approach have a higher scrabble score, these additional steps in the process tackle some of the common issues with data science projects and why they tend to fall short of what is needed. The first additional step is pivotal to tackling these issues. More often than not it will transpire that data analysis start off on the wrong foot. They get asked a quandary by a sponsor, but the analysts will not have a proper frame in which to place this. The first step of SMARTER determines the context. Why is the sponsor asking the questions he/she is asking? But learning the context and the business value attached to the questions, a more critical view can be taken on the questions themselves. Are we looking at the right things, and are we asked the correct questions to come to the insights needed to improve on the associated business value? Once the big picture is clear, the plan of attack presents itself much more clearly. Or if you would like to hear it in Lori’s own words, check out her YouTube presence.

After performing the necessary steps already formulated by OSEMN, we should not only have results of our data analysis. Data in any form or structure isn’t insight. And we need actionable insights if the business value is going to increase. Actionable insight works on three layers: knowledge (taking stock of what we know at present), the current state (what we need to tackle today), and the future state (what we might innovate to improve future dealings). A plan of action should be formulated to act on the insights we have gathered, and decisions should be taken on how to proceed (step T). Once the plan is clear and everyone is aligned, the team should execute the actions dictated by the decisions that have been made (step E), and the results of these actions should be communicated to the different stakeholders (final step R).

Business Storytelling

These last steps are where business storytelling comes in. Although it is certainly already useful to set the context, convincing stakeholders about what the most important actionable insights are and what decisions to take and how to proceed, becomes easier and more relatable in the form of a story. The infographic below taken from the Staying Alive UK website shows to power of a good story, and how it is processed by the brain.

Constructing such a story is a skill of itself. Whereas it used to be a common pastime for us as a species, with hunters telling tales of their hunts around the campfire, or priests telling entire myths populated by a pantheon of gods and heroes, nowadays most of us have delegated this to a subsection of society: the writers of novels, movies, and music. So, we might have forgotten how to go about crafting a proper story. In essence, every story consists of 5 elements:

1. The Setting: This is the framework in which the project will take place. It details the planning, the budget associated with the project, the locality of where the project will take place, as well as any other important factors that will have an effect on it.
2. The Characters: These are the stakeholders that will be participating in the gathering of insight and the making of decisions afterwards. The story should indicate their involvement and what they expect from the conclusion of the story.
3. The Plot: The plot strings together the events that happen in a story. It paints the roadmap of how the story will progress. This is a listing of all actions the project will undertake to get to actionable insight as well as the actions needed to be executed once decisions have been made.
4. The Conflict: During the rollout of the project, there is always conflict that takes the center stage. For the story to have a happy ending, we need to outline the obstacles we will face when trying to get to the needed insight.
5. The Theme: Where the plot lists and strings together the different actions that need to be taken, the theme gives these actions their why. This is the opportunity or problem that is the initial trigger for starting the project. It is the origin or intro for the story that determines how we go about realizing it.

Similarities with this way of thinking can be found in the most data driven world we know: the stock market. Where we have the champion of the efficient markets, Eugene Fama, Nobel prize winner for economics, stating that information gets absorbed and reflected by the market instantly, we also have his co-Nobel prize winner, Robert Shiller, sterling professor of Economics at Yale University painting a different picture. Professor Shiller hearkens back to the days of the first illustrious economist, Adam Smith. In his book “Theory of the Moral Sentiment”, Adam Smith expresses that companies are not solely driven by a need to maximize their profit line, but also be the need to be praiseworthy. Not to get praise but be worthy of it. Professor Shiller also elaborates on his beliefs that narratives can help us understand and predict evolutions that will take place, and in doing so help us to better prepare (or make decisions about) for what is to come. This is further detailed in his book “Narrative Economics”.

When telling such stories, it is important to keep a positive tone. Research shows that negative news heavily influences the decision-making process. More specifically, it impacts the willingness of individuals to shift away from their respective opinions towards a more fitting decision. A study by Bradley R. Staats, Diwas S. KC, and Francesca Gino titled “Maintaining Beliefs in the Face of Negative News: The Moderating Role of Experience. Management Science” (2017) published several findings on this topic:

• Negative news makes people change their views after hearing it.
• People who have a great deal of experience on the topic will be less likely to change their decision when confronted with bad news.
• Similarly, people who are surrounded with more experienced peers are also less likely to change their opinions in light of such news.
• Negative news gets dismissed more quickly when presented to more experienced individuals.

Conclusion

Stories are powerful tools to guide projects that need to gather actionable insight and form the decisions that are needed to move forward and address the initial requirement be it an opportunity or a problem. They give a sense of familiarity to the different participants. They provide a form of abstraction on the complexities of the project that can serve as a reduction of said complexities when communicating with those stakeholders that don’t need to give into the nitty-gritty details. And last but not least, they present a unified way of thinking about the project that helps with the onboarding of the stakeholders as well as the marketing towards external parties.