The world’s leading publication for data science, AI, and ML professionals.

How to Tell a Story With Data

Tips to present data results effectively

Photo by Suzy Hazelwood from Pexels
Photo by Suzy Hazelwood from Pexels

I’ve written about the importance of soft skills to become a successful data analyst. One main soft skill is to learn how to tell stories with Data. When I first became a data analyst I didn’t know that meant. Years later I can finally answer the question "how to tell a story with data" and how you can develop this key skill to present data results effectively.


Description

I describe "telling a story with data" as the ability to answer a business question and explain why it matters to the audience. The answer becomes a "story" when you’re able to "tell" how your results link to company KPIs and how changes in KPI trends impact the business.

Storytelling using an acquisition funnel

The best way I’ve found to create a data story is identifying the KPIs impacted in the customer acquisition funnel. An acquisition funnel differs by business model but two main types are B2C ( business to consumer ) and B2B ( business to business ). B2C sell products or services for individual use while B2C sells to businesses. Companies can also have a combination of business models. Find out which business model or models your department supports to identify the relevant KPIs for your data stories.

Let’s assume you support the marketing department in an ecommerce company with a B2C model. A typical acquisition funnel will have these stages:

  1. Awareness – These are new visitors to the website that never purchased before. Visitors can come from a variety of channels such as organic search, paid search, email, referral, and so on.
  2. Consideration -This involves any visitor behavior on the website such as browsing products, adding items to cart, and checkout.
  3. Conversion – This is when the visitor successfully pays and makes a purchase.

Funnel KPIs

Each department may have KPIs more important than others because they are responsible for a portion of the acquisition funnel. Marketing is normally responsible for top of funnel to drive new visitors but may also be responsible for customer loyalty which is further down the funnel after conversion to encourage customers to make repeat purchases. These are a few common KPIs for each part of the funnel.

  1. Awareness – Unique visitors to the website is the main KPI. Common ways to group the visitors are by geography or by channel. Since visitors may come back multiple times before making a purchase, another common segment is new versus returning visitors.
  2. Consideration – Pageviews per visit or average session duration.
  3. Conversion – Sales and average order value.

There are KPIs across the funnel such as conversion rate which is calculated using order count divided by unique visitors.

Story type

Another component to presenting your story effectively is determining your story type. There are two main types:

  1. Informational – This type is intended to provide new information to the audience that was previously not known. This may be in response to a trend the stakeholder has noticed and would like researched. Alternatively this can be showing data with new segments to identify opportunities for future growth.
  2. Actionable – This type of story shows results that can be used to make a change that impacts the business. A common example are results from an A/B test. Knowing the variant winner will help the stakeholder decide if they should keep the existing control version or implement the variant that had better results.

Knowing the story type helps you tailor your story for the audience. If the story is informational, you want to include new information as it relates to the aggregate KPI. For example if website visitors has always been aggregated you can first show the trend overall and then broken down by state. Marketing can use this information to run campaigns in states with low visitor counts to increase awareness and convert them to new customers. For actionable stories you want to include relevant data to support the recommended action. Note a story can be both informational and actionable. New information can be presented that leads to actionable recommendations to improve the business.

Use case

Let’s walk through a fictional use case at the ecommerce company. Marketing has noticed a recent drop in website visitors and would like you to research a possible cause. Website visitors ( awareness ) is at the top of the customer acquisition funnel and a common KPI to look at is website visitors by channel . You discover the visitor drop is from organic search and upon further investigation you discover there were website changes that occurred on the day the drop started. On the surface this may not seem to be a huge concern for marketing and senior leadership to take notice and have fixed immediately. This is where your knowledge of the customer acquisition funnel will help build your data story.

In addition to researching the cause of the visitor drop you also segment visitors and sales by channel. It turns out visitors from organic search make up 20% of all visitors to the website but they account for 50% of total sales.

With this information you’ll be able to tell a compelling story to show stakeholders and senior leadership why the website should be fixed immediately. These are the key points to get across in your story to marketing’s question "why did website visitors drop recently":

  • Answer – Recent visitor drop is from organic search because of website changes that caused a drop in Google search rankings.
  • Impact to downstream funnel – Organic search visitors that purchase make up 50% of total sales.
  • Why this is important – Failure to fix this issue will lead to a decrease in visitors translating to less customers making a purchase and declining revenue that may put the revenue targets at risk.

A visitor drop may only cause marketing to notice but any impact to revenue will make senior leadership pay attention because it’s an important KPI that affects the company. In addition to answering the question, you showed the downstream impact, why it was important to fix immediately, and the cause of the problem. These are all elements of telling a story with data that ensures your results are presented successfully. Now that you’ve seen an example of how to tell a story with data I hope this helps you craft your own data story.


You might also like…

How to Create a Successful Data Presentation

How I Used a Machine Learning Model to Generate Actionable Insights

How to Translate Data Into Actionable Insights


Related Articles