Telling Stories with Data

  • Sharebar
Tim Mack

Tim Mack

A Review by Timothy C. Mack

From data analysis to storytelling in scenario building. A semiotic approach to purpose-dependent writing of stories” by Javier Carbonell, Antonio Sánchez-Esguevillas, and Belén Carro. Futures 88 (April 2017).

Published in Futures, the article “From data analysis to storytelling in scenario building” describes scenario building as an important method that companies use to understand and communicate strategies concerning the future. Unfortunately, the authors note, the transition from research and analysis (R&A) to the storytelling used for communication is a wide and deep gap and is very likely to result in “loss of consistency and information” that as a result can “jeopardize the whole process.” Fortunately, the authors note, a range of semiotic concepts and techniques can be brought to bear on the problem, using a structured format of their own design. The article’s authors appear to use the terms storytelling and scenario building interchangeably, and this review will do so as well.

Following a short history of the origins of scenario building, the authors determine that there is a lack of consensus about typologies and terminologies and question the utility of that foresight tool in its present state, arising in part because of the “under-theorized” nature of scenarios. Accordingly, they observe a clear difference between the analytic approach of research and the narrative approach of developing persuasive storytelling. In summary, by using two “different types of information,” there will be a resulting “loss of information” between the two stages (called herein a transformational information reduction).

Noting that a range of techniques and practices exist, where elements may be mixed or interact in varying ways along the process of scenario building, the authors helpfully propose that, in their article, the two phases of research and narrative will be artificially separated in order to “facilitate the analysis.” This seems somewhat of a contrast to a number of scenario-building strategies, where the research, analysis, and scenario design and writing remain interactive throughout the process, from beginning to end.

The article considers research and analysis to include both qualitative and quantitative data, and it organizes that data into four categories: scope of the study, objectives, internal aspects of the company (a private-sector client is assumed here), and the external context (also including assumptions and other existing forecasts). The authors note that selection of the information to be analyzed in R&A has previously been done in a wide range of manners and that the choice among approaches depends largely on the scenario’s purpose. The authors summarize this hodgepodge of approaches with an impressive table, outlining the major differences among them. This is followed by the conclusion that, in contrast to storytelling, R&A techniques require “high analytical mindset, mathematical background and high understanding of strategy concepts.”

Scenario building, in contrast, captures major stakeholders, basic trends, key uncertainties, themes, consistency and plausibility, quantitative models, and decision scenarios (Schoemaker 1995). The authors then float a deep concern about the ability of text alone to describe complex situations, especially relations between entities, numerical factors, and probabilities.

To state the problem that is being solved in this article more clearly, the authors opine, “Moving from the first phase rich in data and analysis-oriented to the second, which is more descriptive and oriented to the communication means that some information vanishes.” This disappearance is termed an Information Reduction and can “lead from good research into mediocre storyline.” [Reviewer’s emphasis.]

In fact, it would be difficult for anyone with experience in scenario building to disagree with the authors that the greatest beneficiaries in that process are those organization staff members who actually determine the present situation, future challenges, and likely new trends just over the horizon. And it also seems true that the organizational staff who were not involved in the process are not gifted with this deeper level of insight about the past, present, and future of their organization.

In summary, the article’s premise is that, during analysis, hundreds of variables are usually considered, but in storytelling only the most relevant are included. While this may seem like faulting a funnel for performing its function, the authors offer specifics about where losses of data might occur that could harm the final report:

* Levels of probability and uncertainty may be winnowed out by storytelling simplification.

* The narrowing down reduces the big-picture complexity available at the research stage because of the need to provide scenarios of readable length.

* The range of entities (stakeholders) is often reduced in the conversion to storytelling, including the relationships between these entities (system dynamics).

To counter these deficiencies, the authors offer a semiotic approach, which aims to link signification, representation, reference, and meaning. For those not familiar with semiotic terms such as sign vehicles, object, ground, effect, contour, interpretant, and representamen, some significant effort is required to understand the nature and usage of these elements, since they are not clarified in the article.

This article’s approach is to “help storytelling editors maintain the coherence and traceability the analysis carried out in the first phase, capturing the maximum information from it accordingly to their purpose.” To do so, these storytelling editors must determine what is going to be transmitted and how it is going to be transmitted to the intended audience.

The steps here are the analysis of the target audience, the purpose of the story, and the overall mission statement. This includes not only the R&A report, but any attendant charts, graphs diagrams, and tables that can assist in “making a difference” (relevant to the chosen purpose). Ideally, the elements chosen for inclusion should reflect critical uncertainties, entity relationships, and contour facts that have arisen in the R&A stage. The goals of the story can be emotional, logical, or pragmatic in nature, depending on desired effect on the audience. Finally, the purpose is considered in combination with the story format and length.

The authors strongly argue that it is mistake to tell the story in text only. They propose use of multimedia and “other resources such as colors, voice tone, images” to “cope with the difficulties to change from a data oriented mindset to a message oriented one.” In conclusion, the authors offer a standardized process, which “reduces subjectivity.”

I would contrast this to my understanding of the “early signals” approach, which often uses a wide range of data from a wide range of analysts, then focuses down on those that are most robust—in other words, those that appear the most often across that range of input. Previously, that was the hallmark of good foresight: initially spreading a wide net for relevant data and then narrowing, utilizing experience and expertise to guide that process.

In the article’s process model, the Research and Analysis data is provided by third-party consulting firms and the storytelling element is the responsibility of in-house “story editors,” whose focus is on the structure and impact of the message. This approach seems to leave little continuity between the first and second stages in that setting. Another option, where foresight analysts provide their R&A and also work with the storytelling staff to craft the nature of the story told in the scenario process does not appear to be considered here. A third option is that the R&A and the storytelling (scenario building) are all done by the same team. These other options are not mentioned in the article.

The authors’ focus is on a more intensive documentation of the process (through taskification and quantification) to allow transparent connecting links between the stages, which then provides greater consistency and traceability in the storytelling process. However, “one size fits all” is not always an ideal approach and suggests a distrust of classic (and time-tested) methods of scenario construction. It is also possible that these highly structured and uniform guidelines for converting R&A to story could, in fact, trim off important findings and trends that fell outside the semiotic structure. Ironically, that outcome sounds quite similar to the problem that the semiotic approach was supposed to solve.

In summary, the article’s focus is only on dissecting and restructuring that storytelling process, because of present scenario building’s “lack of maturity in the standardization of methods.” The authors seem to feel the semiotic framework they have designed will offer a greater ability to “set the tone of the story, to create numerous versions, and to ease the traceability between the analytic information of the scenario and the final story.” Accordingly, this “is a first step in order to incorporate a methodology to a process that tends to be artisanal” [reviewer's emphasis], and their next step would be the construction of a manual for teaching this process to foresight professionals.

I must admit that this approach brings to mind the “time and motion” studies of the 20th century, where productivity improvement was expected after a more uniform process was put in place. What seems to be missing from this process is any consideration of real-world situation analysis and testing of whether the introduction of semiotics into scenario building will, in fact, produce the results and improvements that have the authors have projected.

While this review has proceeded on a somewhat skeptical note, it is difficult to fault the authors for their efforts. They have thoroughly researched scenario building (as viewed by other foresight researchers) and have been tireless advocates of the value of semiotics in this new context. However, as a long-term believer in the artisanal value of experience and expertise in analyzing foresight research and then crafting appropriate scenarios, it will take some additional convincing that quantification methods can clearly transform a process where qualitative input has been valued into a “vastly improved” qualitative product.

And finally, in strange contrast to their consistently stated goal are the author’s “afterthought” suggestions that cobbling of additional techniques from science fiction, creative writing, and script writing for television and movies have appeal and could offer valuable cross-fertilization (which this reviewer seconds). As for the value of overlaying the semiotic approach on scenario planning, I would suggest that more research and analysis is needed to make a fully informed judgment.

Timothy C. Mack is managing principal of AAI Foresight Inc. Contact him at