How organisations use Swift workshops to understand the future clearly

Can forecasting help organisations think?

How organisations use Swift workshops to understand the future clearly
Illustration by Laurel Molly, © All Rights Reserved

Forecasting can provide critical insights to guide organisational strategy, and our collaborative workshops are one way Swift Centre allows institutions and businesses to utilise it.

The Swift Centre has expertise with a range of forecasting methodologies, a global team of forecasters ready to both advise and support client decisions, and an easy-to-use forecasting application to streamline the process. At the Swift Centre, our forecasting is built on the foundation of an open-ended, collaborative model.

We ask our community of expert forecasters important questions but, uniquely, we enable the forecasters to investigate new questions according to what they see as most important. We believe – and founder Michael Story’s persistent wins at the US Intelligence Service IARPA’s annual competition supports this view – that drawing on collective wisdom in an open-ended way increases the chance of arriving at accurate probabilities and predictions.

We structure workshops using the Delphi technique, where individuals make predictions, results are shared, and individual forecasts are refined. We teach participants how to avoid bias and prevent any one voice from dominating.

💬
"We worked with the team from Swift Centre on a project to forecast risks to one of our biggest revenue drivers.

It was an absolute pleasure to work with them: they’re professional, responsive, and practical, working hard to understand not just the percentage probabilities of key events, but what this would actually mean for our business.

Not only will the output of the process be a very useful adjunct to our future decision making, but the process itself was excellent training to improve our own calibration and forecasting skills.

We’ll definitely look forward to working with the team again and would recommend them to anyone interested in how world-class forecasting can contribute to their work."

— Client testimonial, Metrea, August 2023

EA Global workshops

While most of our client work deals with private and sensitive information, we can publicly share the forecasting output of two workshops the Swift Centre hosted in October 2023, where we introduced the fundamentals of forecasting at Effective Altruism Global conferences in Boston and the Philippines.

A guiding principle for those affiliated with effective altruism is to use reason and evidence to figure out how to help humanity in the most impactful ways. Forecasting, when practised well, can be an effective tool for decision-making, as it keeps you squarely focused on the true consequences of changes in the world. Additionally, improving decision making in important institutions is considered to be a key effective altruism cause area. Many view the incorporation of forecasting best practices as part of the answer.

Michael Story at EA Global: Boston 2023

The workshops involved working with a lay audience for an hour and a half. Because of the limited time, we selected a topic that most attendees were likely to have pre-existing knowledge about: how the future of Twitter will unfold and its consequences for the 2024 US election. This served as an introductory example for teaching conditional forecasting.

Participants used our intuitive Swift forecasting app showing all forecasters’ judgments and comments via data visualisations. We also teach a Swift Forecasting Checklist:

  • Base rate: what typically happens in similar situations.
  • Inside view: the specifics of the situation.
  • Scope and scale: quantify the questions: how many, and by what date?
  • Secondary sources: what other data can you use to check the reliability of your view?
  • Pre-mortem: interrogate findings, identify gaps, identify more questions and research.

As with all of our workshops, participants thought of how events could unfold and suggested new question ideas. Participants discussed the effects of Twitter going bankrupt on political discourse and the correlations between broader economic shifts and voter turnout. We observed in the discussion that participants naturally came up with relevant and important areas of enquiry. We did not have time to incorporate these into the forecasting exercise, but this generative approach is what enables us to adapt forecasting workshops towards questions that are most relevant for our attendees.

Of course, the workshops' audiences were not professional forecasters. However, learning about and trying forecasting is not only good for creativity, it's also an effective way to understand cognitive patterns, how to seek the truth as a team, and determine the real consequences of different events.

So, what did the audience come up with? 

Does Twitter (or ‘X’) matter more than we think?

Boston had an older, mixed audience; in the Philippines, the audience was mostly college students. We expect this difference in demographics drove the differences in aggregate forecasts between the two cohorts, with the Philippines participants being more clustered in age.

We asked the top-level question, "Will X, formerly known as Twitter, declare bankruptcy before Aug 22, 2024? ". We recorded both the range of answers and the average result.

A graph with text overlay

Description automatically generated

We then asked a series of conditional forecasting questions, which are about the likelihood of certain events happening if Twitter does or doesn't declare bankruptcy. 

When asked, for example, whether Trump would win the 2024 US Presidential Election, both audiences scored similarly – 38% for the Philippines and 39% for Boston. 

But how did the prospect of Twitter going bankrupt or not affect those scores? 

A graph with text and numbers

Description automatically generated with medium confidence

Both groups – to different degrees – forecast that if Twitter declares bankruptcy, Trump would be less likely (Boston reduced to 32% and Philippines to 29%) to win the election. If Twitter did not declare bankruptcy, Trump’s chances of being re-elected would increase by nearly 1%. So, if Twitter goes out of business, it would – for these amateur forecasters – harm Trump’s chances of winning the election. 

We asked other conditional questions on US election spending levels, whether turnout would increase from 2020, and whether Twitter would have a new owner by 2025. You can see the full interactive results for all the conditional questions we asked on the links at the end of this article. 

A view of the future

Using collective wisdom to predict outcomes of events – like the probability of Twitter going bankrupt or the likelihood of the Israel/Hamas war spreading – is critical to governments, organisations and businesses alike. No one prediction is ever 100% accurate – it is a calculation of probabilities not certainties.

These political and social events impact organisational strategy, and so our ongoing mission at the Swift Centre is to refine and evolve to enhance the accuracy of our approach. By doing so, we can use forecasting to assist in forward planning, and in helping predict what the most impactful changes or policies might be.

The Swift Centre offers a host of services to provide insights that inform risk management strategies, help organisations grow staff forecasting capabilities, and generally enhance strategic planning:

  • One-off workshops to introduce a team to thinking like a forecaster.
  • Collaborative workshops where Swift analysts and clients work together to make estimates.
  • Written reports, monitoring and analysis by top forecasters.
  • Ongoing collaboration to support business and institutional forecasting.

We are highly responsive and happy to answer any questions you may have. Please get in touch to find out how forecasting could impact your business or organisation.

Boston: https://viz.swiftcentre.org/results/O9KeUv_DaD0

Philippines: https://viz.swiftcentre.org/results/hImrzyNozGw