At Versta Research we pride ourselves on rigorous design, flawless execution, and smart statistical analysis. But none of that matters unless the audience for the research is able to connect it immediately to the people, places, and events they care about. Turning data into stories does that.
In this newsletter we offer a how-to guide: Versta Research’s 7 Steps to Turning Data into Stories. It is the process we use every time we write a report.
Other items of interest in this newsletter include:
- Bon Appétit’s Big Lie Survey
- Newsrooms Flooded with Silly Surveys
- Rules of Thumb for Survey Length
- Do You Have a “Metrics” Fetish?
- Finding Insights in Virgin’s Data
- Opinion Polling Basics for News and PR Professionals
- Just Published: Handbook of Web Surveys
- Focus Groups Save Spider-Man!
- A Path to Better Research with Geo-Maps
- Your Margin of Error Is Probably Wrong
- The Night Before Christmas: A Research Approach
- 42 Smart Applications of Marketing ResearchWe are also delighted to share with you:
The Versta Team
Turning Data into Stories: A How-To Guide
“I find that most people know what a story is until they sit down to write one.”
If there is one element of research that stymies research professionals the most, it is the point at which they must write, or at least try to write, a report. That is the point at which good writers stare at their screens, struggling to communicate what the research means and why it matters. It is the point at which poor writers offer convoluted sentences sounding like math formulas, unfortunately dooming their reports to dusty shelves. It is the point at which clients wish there were an easy way to turn all those vendor-supplied charts and tables brimming with data into something their management will listen to and get excited about.
The problem, in our view, is that too many in our industry write reports about research, when the most useful reports are not about research at all. Research is the means to an end; it is simply a way of getting information to answer questions. So the goal in writing a research report should be to have readers “see through” the research and the data to the people, places, issues and business questions they care about. In short, a good research report should turn data into a story.
In a previous edition of this newsletter we offered a perspective on why turning data into stories is important and how it helps improve research. In this edition we offer a nuts-and-bolts description of how to do it. The process is not easy, but neither is it mysterious. It is not about focusing on story lines, plots, character development, and dramatic resolution. It is simply about translating statistics and data into meaningful statements about what people think, feel and do in the world.
Here, then, are…
Versta Research’s 7 Steps to Turning Data into Stories:
1. Review the questions that need to be answered. The path from the beginning of research to the point at which you write the story is rarely direct. Nor should it be, as good research involves focusing intently on methodological details that support the rigor of the story and that ensure meaningful data, even if they are not central to the story itself.
But this means it is critical to go back to the RFP and to the research proposal. Review notes from the first meetings where stakeholders discussed (or where you artfully elicited) their questions and helped define what the research needed to answer (see The Art of Asking Questions). Remind yourself of the core business questions at stake, and remember that the story you need to write is not about the research you conducted, but about the answers to those questions.
2. Assess how the data are calculated. An important step in analysis is simply determining the base against which percentages are calculated. There are always choices to be made based on the story you want to tell, and unfortunately many analysts tie their stories in knots by relying on automated tabulations.
Suppose you asked your survey respondents whether they increased their savings compared to last year, and for the 40% who did, you asked a follow up question and learned that 50% did so because of economic uncertainty. Now you have a choice: do you report “half who increased their savings did so because of economic uncertainty,” or do you report “20% increased their savings because of economic uncertainty?” It depends upon the story you want to tell. Unless you have your calculations laid out correctly, your story will get muddled.
3. Lay out simple data points. Showing data for all response options to a survey question is usually too much and it detracts from the numbers you really want to show. So think ahead in terms of which data points will make it into the CEO’s rationale for a strategy or into a press release highlighting the research findings.
For example, your CEO will never say: “Our new strategy is based on the fact that 45% of customers strongly agree that speed of delivery is their first priority, and another 35% agree, while just 15% disagree and another 5% strongly disagree.” She is going to say: “80% tell us that speed of delivery is their first priority.” Period. So highlight the eighty percent and forget about the rest. You can always add detail, and it is easy to consult the tabulations or posted questionnaire if needed.
One benefit of showing less is that it makes you think more about your story. You need to make smart choices about which points on measurement scales to emphasize (should you combine the agree with strongly agree groups; should you focus on 80% who agree or 20% who disagree?) These are questions that can only be answered in terms of the story that needs to be told.
4. Organize the data thematically. A mistake that analysts often make is laying out a research report in the order that survey questions were asked, such that the report becomes a blow-by-blow account of how people responded, rather than being organized around the central questions the audience cares about. Typically, you will need to assemble data from disparate sections of the questionnaire to write a strong report.
Think about the themes of the story and the questions that need to be answered, not about the logic and themes of the questionnaire. Go through the data point by point, and start inserting each summary statistic into a section that helps answer a question. If, in each section, you find yourself listing data points in the same way as they are laid out in the questionnaire, go back to step 1.
5. Put the data into statements as bullet points. Nearly every statistic in your report should point to something that real people do in the real world. So the goal is to write statements about what the statistics point to, rather than writing statements about the data or about how respondents answered survey questions.
Here is a strong statement that tells us what people care about in the world:
- A large majority of customers (80%) focus on speed of delivery as their top priority
Compare that to these weaker statements that emphasize how respondents answered survey questions—both of which require the reader to make the implicit connection to what the question actually measures:
- 80% selected “speed of delivery” as their top priority from among eight service areas tested
- Most customers (80%) agree with the statement that speed of delivery is their first priority
6. Summarize the statements in headlines. When you have a document with every survey item summarized into a simplified data point, organized into themes that answer the key business questions, and with meaningful statements that reach beyond the data itself, write a one- or two-sentence headline that captures the theme of each section. Similar to the statements in each bullet point, these headlines should clearly state what the data point to, and they should emphasize what people think, do, or feel rather than simply labeling the topic of each section.
In writing headlines, avoid using numbers, relying instead on quantifying word such as many, most, few, etc. Moreover, look for opportunities to emphasize contradiction, contrast, logical flow, evolution, or development. Good stories nearly always have these elements, and conceptually they will help you tie disparate data points, sections, or themes into a whole that will help readers see connections.
As examples, consider these headlines from a recent study we conducted about saving for retirement:
When it comes to managing their retirement assets, most employees—even those closest to retirement—feel like beginners, and few have formal strategies in place.
Retirement is their number one savings goal, but fewer than half feel confident they are saving enough for what they will need.
7. Revise and strengthen the words. The headlines, strung together, should tell a clear, multi-dimensional story in one paragraph. Each headline should have several supporting statements to back it up. And each statement should have one or more statistics offered as proof. As all of this starts coming together, re-vision it, re-organize it, and re-write it.
For an outstanding guide on how to revise and strengthen the writing itself, we enthusiastically recommend Richard Lanham’s Revising Business Prose. Just as our approach outlined here will help you find and extract the story in your data, Lanham’s approach in Revising Business Prose will help you find and extract the story from a muddle of words.
The Story Is What the Data Prove
Ultimately, the key to turning data into stories is to keep research data in the background while keeping the issues that the data point to in the foreground. Each step we have outlined above, from reviewing the motivating questions, to editing mountains of detailed data, to summarizing statements about the data with headlines, will help you do that.
The next time you sit down to draft a research report, here’s our advice in a nutshell: Do not write about the data, and do not write about the research itself. Write about what the data prove, because that is the story that your internal clients and managers really care about.
Stories from the Versta Blog
Here are several recent posts from the Versta Research Blog. Click on any headline to read more.
One way to get customer data is to fool customers into taking an “opinion” survey. But the truth is a better approach, and helps ensure future goodwill.
If you need a survey for PR, it is possible to cut through the clutter of silly surveys with research that is well done and that focuses on what readers care about.
Ideal survey length depends on several factors, but here we outline our industry rules-of-thumb for maximum survey length, which vary by survey mode.
Research often focuses too much on measurement, losing sight of what is being measured. The solution? Formulate better questions that the research must answer.
As marketing professionals face increasingly immense volumes of data, the need for smart research is greater than ever. Our advice: focus on the questions, the tools, and the story.
The Poynter Institute and AAPOR have teamed up for a webinar on the basics of opinion polling and survey research, useful for PR and marketing professionals.
Looking for advice on how to design, field, and analyze web surveys for rigorous research? This just-published handbook covers the essential topics you need to know.
The producers of the Broadway musical Spider-Man turned to market research and rescued their show from disaster. Steve Jobs is not always right!
Geographic mapping is a compelling way to show customer data and it is now easier to create maps with data and tools available (for free) on the Internet.
A recently compiled database of polls from the past 40 years shows that two-thirds of the election outcomes fell outside the margin of error. What’s going on?
Here is a humorous survey written by a professional telephone interviewer about what is going in peoples’ homes on the night before Christmas.
Versta Research reviews and highly recommends the book Consumer Insights: Findings from Behavioral Research from the Marketing Science Institute.
Versta Research in the News
Fidelity Investments released findings from its national survey of active traders that focuses on motivations to trade and keys to success.
Versta Research has entered into a two-year agreement with one of the nations top 100 pension funds to monitor and understand employee health.
Public Relations Tactics has published its October 2011 issue with a full page article from Versta Research entitled “Telling the whole story: When numbers aren’t enough.” The article describes the difficulties of using traditional numeric scales, common in survey research, for data that needs to tell a story.
MORE VERSTA NEWSLETTERS