On connecting data to successful business decisions

By Grant Goad

Let’s say you work as a marketing manager. Maybe you’re lucky enough to have access to a dedicated analytics team.

You request specific data, get reports in return, then use that data to hopefully make better decisions on AdWords spend, content marketing, email marketing, and related channels and tactics.

But take a step back and ask yourself—Are you consistently happy with the outcomes? Do the data reports you create really help improve business results?

And even if they do, couldn’t those results be even better?

I recently sat down with business analyst and data visualization expert Charles Sutton to discuss how we can improve the data reporting process.

Charles’s company, Impacsis (a West South analytics partner) specializes in creating reports that connect relevant data to successful business decisions. His goal is to make his clients happy by providing them with reports that truly meet their needs.

The company’s tagline sums it up:

Building reports people love.

Charles has honed his analytics expertise over the last ten years with work for companies in a range of businesses, including Petco, Intuit, and life sciences firm Organogenesis.

His approach is based on the three “A”s of good reporting: accessibility, accuracy, and actionability.

We’ll discuss this approach in detail, and also cover topics such as asking the right questions, choosing a data visualization platform, information overload, and more, in the interview below.

Matt Shoemaker, left, and Charles Sutton, right, at the inaugural Diego Data Discovery workgroup meeting.
Charles Sutton, right, with Matt Shoemaker at the first Diego Data Discovery workgroup meeting.

The background


Hi Charles. Thanks for joining me today. Tell us about Impacsis and what you do.


Hi Grant. Pleased to be here. Impacsis is a consulting firm that helps people build reports that they love. By that, I mean we enable clients to overcome the many obstacles that can come up with data reporting.

What I’ve observed in the business world is that decision makers are frustrated by their interaction or experience with reports. Reports are slow loading. The data’s not always accurate. Or they’re not getting the right data—the data they actually need to make a business decision.


Is this in relation to any particular type of report? Marketing, for instance?


I’m referring to reports in general. But let’s take it from the marketing perspective.

My first question is who’s looking at the report. Is it the director of marketing? The chief marketing officer? The marketing manager? Business analysts that tag the site, capture data, follow the customer through the lifecycle?

I ask because each one of them is going to want different levels of detail. The CMO will want to look at higher-level details, a broader perspective. They want to see how the totality of their business is running, and how the different components are working together.

The CMO will look at this higher-level data, see something there, and reach out to the marketing manager—“Hey I’ve noticed that our new customer volume in our Google channel is lower than we’d expected. Can you tell me some of the reasons why?”

And now the marketing manager, who has a narrower, deeper set of data, will dive into that data for an answer.

You have to know who your audience is and then create a dataset that fulfills their needs. That’s where the conversation—communication—needs to happen.


So, you’ve got to think about your customer’s needs first, and then choose the data to collect and sources to collect it from. Instead of pulling down all the data you can from the get-go.


Yes, because as big as servers get, they still have limitations. The more data you throw on a server, the slower it’s going to run. Depending on how well your data architecture is put together, performance has to be taken into consideration.

If you can narrow down the amount of data you need in a report, it’ll help reduce performance issues. With tools like Tableau, Excel, and Power BI, you can’t just throw everything in there. They won’t perform well.

You really have to refine the dataset in the beginning, which means you want to pull the least amount of data that can answer your questions.

For example, you can’t just go in and say, “Hey, I’ve got data source A, B, C, D, and E, and they’ve got 150 rows or 120 columns,” and then throw that data onto a server and expect it to be responsive.

And you definitely can’t take that approach with a data visualization tool. If you tried to throw every single slice of data you had into a tool, it would take an hour to load and you’d never get the information out.

Start the conversation first

In order to understand your customer’s needs, you need to have conversations. I usually start off by getting together a key group of analytics users within a department. I want key people from each level to be part of this work group, or task force, if you will, and just tell me every single business question that comes to mind.

Anything they can think of. Because each of them is going to have a different perspective. I then take all the questions and put them into categories—organization-wide, director-specific, manager-specific, etc.

I then build a framework for each of the reports I’ll be creating. I determine the data sources we’re going to need, and I work with the engineers to build the data pipeline.

So you ask them questions to start out, create a Google doc. Ask them to say anything that comes to mind. I want to know all their questions so I can build a data source that’s comprehensive enough to answer them.

Data report created by Charles Sutton on the state of American well-being.
Data report created by Charles Sutton on the state of American well-being. View the report on Charles’s public Tableau profile. >

Data sources


And what might some of your data sources be?


Right now, I’m working on call center reporting. This is performance reporting. The data usually comes from five different sources.

Performance data on how well each call center agent answered calls comes from Amazon Web Services. Data on whether or not they’re logging cases about who they’re talking to and how they’re interacting comes from Salesforce.

Information on how much experience each agent has, how long they’ve been with the company, and how much training they have comes from a QuickBase data source.

Their hours, when they worked and how long they worked, come from Workday.

The fifth source is survey data, which comes from Qualtrics.

So now you have to bring survey data, agent data, performance data, operational data, and human resources data into one comprehensive data environment.

Let’s say we’re talking about our net promoter score (NPS), which is all about customer satisfaction. This is a critical key performance indicator for a chief marketing officer.

If the NPS is below expectations, the CMO can reach out to the call center manager and ask them to look into possible reasons for this. The manager can find out which agents are underperforming, and which agents take the most calls, and then work with the underperforming agents to improve.

Keep in mind that when I start a data project, I don’t ask the stakeholders, “What columns of data would you like to include in your reports?” My job as a business analyst is to make sure that decision makers have the right data to make the decisions they need to make.

And I deliver that information through an interface called the report.

The job of the report is to deliver the right information to the right people at the right time.

It’s customer-centric design. The reason Impacsis has the tagline “building reports people love” is because I’m really trying to concentrate on that people part. Yes, I build reports. But who do I build reports for?

I build them for people. I want my customers to really enjoy interacting with my reports.

It’s not as easy as just throwing some data into an interface and saying go. When you do that, you end up with information overload. If a client wants a high-level overview of their business, you can’t bombard them with 400 filters and graphs that they can’t easily understand.

Data report created by Charles Sutton to help users select lodgings in Berlin using Airbnb.
Charles’s location selector for Airbnb rentals in Berlin. View the selector on Charles’s public Tableau profile. >

Data visualization tools


What’s your favorite tool for data visualization?


I really try to be tool agnostic. My message is that it’s the communication, the thought process, the collaboration, and the trust we build together, that make it possible to get the right information to the right person at the right time.

Now I will say that the tool I use most is Tableau. It’s an easy-to-use tool. For me, it’s the tool that makes it easiest to create a report that’s accessible, accurate, and actionable.

These are the three pillars of good reporting:

Accessibility—you have to be able to get to the information easily.

Accuracy—you have to be able to trust the information.

Actionability—you have to be able to use the information to solve business problems.

With a data visualization tool like Tableau, you get a Tableau server, where you can store all your reports. It gives you a centralized repository to access information.

Imagine if you went to the San Diego Public Library and there were no signs, no computers, no librarians, and you wanted to find a book on Golf. Maybe you’d find the book five days later, if ever.

Analytics for enterprise and small business


Does your approach to creating data reports vary for enterprise clients as opposed to small business clients?


The nice thing about our approach is that it works for both small and large businesses because it’s all about the process.

Whether you’re a one-person hotdog stand or a hundred-person software company, a report is meant to answer questions. You have to define those questions.

You have to dig deep and ask yourself, “What do I want to learn about my business?”

Determine your business goals and then ask what kind of information can help you achieve those goals.

The next step is to find the data sources to help answer those questions. You don’t just get Salesforce and expect it to provide you with all the answers.

As a business analyst, I’m here to help you reach your goals by answering questions with data. I’ll tell you where to get the data, give you options on which tools to use, and provide recommendations on what to do with that data afterwards.

Data reporting recommendations for business owners


Do you have any recommendations for business owners when it comes to analytics and data?


Start with questions. Work with an expert. Business analysts are out there because we understand business and data. IT people understand data. Businesspeople understand how to run a business.

You need a bridge between those two parties to understand what information the business needs and define the data pipelines for the IT people to build. Without this bridge, the businesspeople don’t get the information they need, and IT gets overwhelmed with requests that they can’t handle.

And thus, trust is eroded, products aren’t built the way people want them, and you end up with tension between the two groups.


Any other closing thoughts?


It’s all about the process. The takeaway from our discussion is that communication has to be improved between the people who want reports and the people who build them. And the people who build reports need to understand what the report users do.

And the way I achieve this is to ask questions.

It’s all about the conversation.

See more of Charles’s data visualization work on Tableau. >