Persuading with Data: Mistakes to Avoid, Tips for Success

Each conference I attend brings me more insight about the analytics field. Sometimes it's something practical such as how to make my code faster. Many times it's how to be a better analyst.

One common issue analysts face is how to present data that changes mindsets, as the late Hans Rosling would say. To a data analyst, there is nothing more exciting than the data revealing insights about real business issues. But it is completely deflating to present the insights and have nothing happen. At this past SAS Global Forum, I found a few insights of my own on how to overcome this situation.

Don't Over Dazzle Them

One attendee revealed that he wasn't persuasive until he understood how to talk to the audience in front of him. This user had a Ph.D. in statistical analysis so it's fair to say that he knew a little about the subject. During his first job, he wasn't having much success when presenting his analysis to the management staff. Finally a manager said that the presentations had such advanced analytics he felt stupid because he frequently didn't understand what was being presented and was afraid to ask questions in front of others.

Ouch! In his effort to dazzle with this brilliant insight, he had failed to consider the audience. When presenting to business leaders, consider how much analysis is necessary to make your case. Perhaps talking about standard deviations and t-tests are too much detail or detail at the wrong level. It depends on the company culture. Many times managers want a higher level of detail. If you are worried that the detailed analysis is required, consider adding some backup slides. In this case, the attendee learned that his ability to persuade with data analysis had to match his audience's level of understanding.

Make the Issue Smaller

At breakfast an attendee commented that he knew his school could rise from a third position to the top position in the state. When thinking about all the work involved, it felt like a big challenge to me. After discussing the issue, I advised him to consider adding the human to the data since the audience may be more drawn to helping individual students become better than the district as a whole. People like to help other people and educators like to help students, right? I walked away feeling pretty smug and enjoying some fresh coffee.

Later when I was talking to another attendee, he offered some additional insight. Sometimes the number one position for the school may be as close as just 50 students showing improvement. This thought reduced the size of the issue. Instead of "how do we bring the entire school district to number one?" it could be "how do help 50 students excel in their academic careers?" Then you can present specific instances of these students and learn what was needed in each case. Perhaps the team would be able to find simple and cost effective solutions to move the school district forward.