Data-Driven Presentations: How to Win the Room With Non-Technical Audiences
In a session at the Data Science Innovators Conference 2023, an analytics leader shared tips from the book Winning the Room for presenting results to non-technical audiences. The advice covers planning, slide design, rehearsal and tying findings to business impact.
By The AI Enterprise Magazine Team4 min read
Key takeaways
- Results are necessary, not sufficient. The speaker said at least half of success comes from how results are presented and received.
- Design before slides. Plan the story, order and actions first, like a film, and focus on how the audience will use the results.
- Show fewer numbers. Extract only the figures behind your headline and round them, with full tables in an appendix.
- Rehearse for time, not slide count. Practice out loud and highlight the number that matters.
- Tie facts to impact. Convert a metric change into dollars or efficiency and link it to strategic goals.
In a session at the Data Science Innovators Conference 2023, an analytics leader and author laid out how data professionals can present results to executives and other non-technical audiences. The 20-minute talk, "Winning The Room: Creating and Delivering an Effective Data-Driven Presentation," drew on three sections of the presenter's book of the same name: planning, building and delivering a live presentation. The video description does not name the speaker.
The advice applies to live rooms, conference sessions and webinars alike, the presenter said. It is not about data visualization or storytelling as such, but about influencing people who must act on your findings.
Why do good results still fail to drive action?
The goal of any live presentation, the speaker said, is to get an audience that is often far less technical than the presenter to understand the findings, grasp the implications and take action.
The presenter argued that accurate results are only part of the job. "Results are necessary, they're just not sufficient," the speaker said, estimating that at most half of success comes from the results themselves, with the rest coming from how they are presented and received.
Many teams spend months on analysis and leave only a few hours to build the presentation, according to the speaker. The key reminder: "you are there to tell a story, not to write a story."
How should you plan a data presentation?
Designing the talk comes before making slides. The speaker compared it to filmmaking: a director does not shoot random scenes and hope they assemble into a movie. Decide what to tell the audience, in what order, with how much detail, and what actions you want.
Focus on how the audience can use the results, not how they were computed. The speaker used a car dealership example. A buyer who asks how a backup camera works wants to hear that it shows what is behind the car in reverse, not a rundown of screens, cables and processors.
For a business sponsor asking about a response model, the useful answer is that every customer gets a probability of response and the team can supply a contact list of whatever size is wanted.
How do you build slides that keep attention on the speaker?
Keep slides brief so attention stays on the presenter. One example reduced a slide of four action-item boxes to a single slide, with animations revealing each point as the speaker reached it.
The speaker also recommended using PowerPoint's notes section for extra talking points. Printed notes pages can serve as a leave-behind for people who were not in the room, and as a quick review before presenting.
Other build tips included:
- Show the fewest numbers needed to make a point. A 31-number table was cut down to the organic honey row the headline was about, with the full table left for an appendix.
- Use lay terms. Instead of sensitivity and specificity, say true positive rate and true negative rate.
- Round aggressively. Showing dollars in millions and percentages to the nearest whole number is enough when the point is roughly five to one.
How should you rehearse and deliver?
Practice out loud, the speaker said. Before presentations the speaker walked around a hotel room giving a full dry run, which exposed rough spots and ordering problems.
The speaker's rule of thumb: focus on time, not slide counts. "I have slides I cover for 10 seconds, I have slides I can cover for 10 minutes." During rehearsal, add circles, arrows, bold or color to direct the eye to the number that matters.
For online talks, look at the camera and not at attendees' thumbnails. Looking at a face on screen reads to viewers as ignoring them.
How do you tie results to business impact?
Two closing tips focused on translating numbers into outcomes. A 10% change in complication rate means little until it is tied to dollars, efficiency or another metric the audience is judged by, the speaker said. Do the math for them, such as showing cost per complication down 20%.
Then connect a tactical result, like last month's marketing numbers, to the company's larger strategic goals so executives see how it advances their agenda.
Where to learn more about communicating data
Presentation skills are a recurring theme at AI Connect events. The 2027 webinar schedule offers 24 free live sessions across four series, including the AI × Enterprise Panel Series and AI Demo Days. See /webinars/; registration for 2027 events is coming soon.
Frequently asked questions
How do you present data to a non-technical audience?
The speaker advised focusing on how the audience can use the results, using lay terms such as true positive rate, and showing the fewest numbers needed to make each point.
How many slides should a presentation have?
The speaker said to focus on time, not slide counts: some slides take 10 seconds and others 10 minutes, and rehearsal shows whether it fits.
Why do data science results fail to drive action?
According to the speaker, accurate results are only half the job; if they are not put in context and the audience's language, listeners get confused and nothing happens.
How should you present on Zoom or a webinar?
Look at the camera rather than attendees' thumbnails, because looking at faces on screen reads as ignoring the audience.
This article is based on remarks by Presenter not named in the video description (author of the book Winning the Room) at Data Science Innovators Conference 2023 on April 12, 2023. More coverage: More AI Connect Events.
Filed under: People, Culture & Leadership
