3 Comments

  • Elizabeth McDonald says:

    This is great information …and thanks for such a handy diagram!

    Do you have any thoughts on guidance for data viz designers working for very large companies with the goal of driving self-service? I ask because I work for a huge global company composed of hundreds of sub-businesses and literally countless stakeholder needs (in terms of filtering requirements to arrive at the population of interest, questions being asked, metrics of interest, etc.).

    The Center of Excellence I’m a member of aims to drive self-service for virtually all business needs …and so I find myself almost always designing interactive dashboards to meet the diverse needs of our end users (the dashboards have different layers – a high level summary, a middle level, and then details through tooltip pages and/or drill through pages).

    My biggest struggle is how to meet this self-service need, when (a) the requirements are largely undefined (because we can’t possibly gather requirements for every single permutation of end user groups) and (b) a huge percentage of our end users are both non-technical and extremely busy, and so unlikely to have the time and/or ability to use the tools we offer. I’m doing my best to have clean, simple designs with built in help, easy access to metric definitions, and a training portal, but I worry that self-service is a pipe-dream and I am set up for failure.

    Do you have any advice?

    • Ann K. Emery says:

      Hi Elizabeth, This is definitely a common challenge!

      Here’s what I’ve seen in practice — let me know if your org is similar.

      What usually happens is that the people creating dashboards (like us) get a LOT of requests for different variables, datasets, charts, slices of data, etc. That’s great at first. People are interested in the data; we want to be responsive and helpful to their requests.

      A year goes by, and then another. And the dashboard(s) have ballooned and ballooned and ballooned. There are so many variables, datasets, charts, and slices of data. It often becomes unusable.

      If this happens in your org, don’t forget to pare down again. DON’T ask people, “Which parts of these dashboards should we delete?” That causes unnecessary stress. (“What if we need that chart/dashboard/slice at some point…???”)

      INSTEAD, try asking, “What’s the most valuable part of this dashboard(s)?” You’re listening to see what’s *not* mentioned. Did users mention A, B, and D? Great! Then remove C.

  • Elizabeth McDonald says:

    That is helpful – thank you!

  • Leave a Reply to Elizabeth McDonald Cancel reply

    Your email address will not be published. Required fields are marked *

    What Type of Dashboard Do We Need? 4 Types to Consider + Diagram to Download

    Updated on: Feb 7th, 2022
    Dashboards
    , ,

    What type of dashboard do we need for our project?

    I want to talk about something that’s a little controversial in the dashboard space: There are 4 types of dashboards, all of which are correct.

    You might need one type.

    Or, you might need all four.

    Every audience and every project is a little bit different.

    Our goal is to deliver the right data, to the right audience, at the right time.

    Watch the Video

    Here’s a 12-minute video to help you narrow down which type(s) of dashboard you need for your next project.

    First, you’ll think about your audience. Are they technical or non-technical? Leaders or doers?

    Then, you’ll choose a software program. Sometimes you’ll need spreadsheets (like Excel), and other times you’ll need dashboard programs (like Tableau or PowerBI).

    Here’s a summary of what’s inside the video.

    Dashboard Mismatches to Avoid

    What if the audience has one idea in their mind… and we design something completely different??

    Sometimes we get lucky, and we make exactly what they want.

    Other times, I see mismatches.

    Common Dashboard Mismatch: Static vs. Interactive

    Sometimes the audience really wants an interactive dashboard, but we make a static dashboard.

    In real projects, that almost never happens. I typically see the opposite issue: We design an interactive dashboard, but they really needed a static dashboard.

    Common Dashboard Mismatch: Single or Series

    Here’s another dashboard mismatch I see a lot:

    Sometimes the audience really wants a single dashboard, but we design a series of matching dashboards.

    Or, I see the opposite:

    The audience really wants a series of matching dashboards (e.g., one for Project A, one for Project B, and one for Project C), but we give them a single overview showing all three projects combined.

    The Four Types of Dashboards

    Let’s be proactive and avoid these dashboard mismatches altogether.

    Here are some factors to consider at the very beginning of your next dashboard project.

    Technical or Non-Technical?

    Is your audience technical or non-technical?

    Technical audiences love data, details, and decimal places.

    Non-technical audiences would rather be doing something else, like leading the project, developing new policies, or managing the team.

    Time is another factor: Are they busy? Or, do they have time to explore a dashboard on their own?

    Non-technical or busy audiences tend to prefer static dashboards. These short PDFs can be shared as email attachments or as printed meeting handouts.

    Technical audiences (or those with plenty of time available) tend to prefer interactive dashboards. They love exploring these clickable, dynamic dashboards and coming up with their own insights.

    Leaders or Doers?

    Next, figure out whether your audience is mostly leaders or doers.

    The leaders need an aggregated overview of the work, e.g., one dashboard for the state as a whole.

    The doers need individualized, disaggregated data, e.g., one dashboard for their charter school x dozens of charter schools in the project.

    If we give the leaders the disaggregated dashboards, we risk that they’ll get lost in the weeds.

    And if we give the doers the aggregated dashboards, we risk that it’s not actionable enough for them to do anything about the data.

    Audience First, Software Second

    AFTER we narrow down our audience, then we can choose a software program.

    Single static dashboards can be made in spreadsheet programs, like Excel, Sheets, or Numbers. We can sorta make them in infographics programs like Canva or Piktochart; those templates are meh but they’re getting better all the time.

    Need a series of matching dashboards? Spreadsheet programs can handle those, too. I make one template in Excel and then automatically populate it with all the dozens of dashboards’ data. You can write VBA code, connect everything with drop-downs and lookup formulas, or use Slicers.

    Have a technical audience? Interactive dashboards are possible in Excel (via Excel Tables, pivot tables, pivot charts, and slicers). Or, you can make them in dashboard programs like Tableau or PowerBI. Or, you can learn coding (e.g., R).

    In the video, you’ll see real-life examples of these dashboards, too.

    Download the Diagram

    Want to download this diagram? Share it with your team, and discuss it together at the beginning of your next dashboard project. Let’s avoid those mismatches altogether.

    Download the Diagram

    Your Turn

    Do you currently have any dashboards?

    Who are they for? Non-technical or technical audiences? Busy people, or those with time available? Leaders or doers?

    More about Ann K. Emery
    Ann K. Emery is a sought-after speaker who is determined to get your data out of spreadsheets and into stakeholders’ hands. Each year, she leads more than 100 workshops, webinars, and keynotes for thousands of people around the globe. Her design consultancy also overhauls graphs, publications, and slideshows with the goal of making technical information easier to understand for non-technical audiences.

    3 Comments

  • Elizabeth McDonald says:

    This is great information …and thanks for such a handy diagram!

    Do you have any thoughts on guidance for data viz designers working for very large companies with the goal of driving self-service? I ask because I work for a huge global company composed of hundreds of sub-businesses and literally countless stakeholder needs (in terms of filtering requirements to arrive at the population of interest, questions being asked, metrics of interest, etc.).

    The Center of Excellence I’m a member of aims to drive self-service for virtually all business needs …and so I find myself almost always designing interactive dashboards to meet the diverse needs of our end users (the dashboards have different layers – a high level summary, a middle level, and then details through tooltip pages and/or drill through pages).

    My biggest struggle is how to meet this self-service need, when (a) the requirements are largely undefined (because we can’t possibly gather requirements for every single permutation of end user groups) and (b) a huge percentage of our end users are both non-technical and extremely busy, and so unlikely to have the time and/or ability to use the tools we offer. I’m doing my best to have clean, simple designs with built in help, easy access to metric definitions, and a training portal, but I worry that self-service is a pipe-dream and I am set up for failure.

    Do you have any advice?

    • Ann K. Emery says:

      Hi Elizabeth, This is definitely a common challenge!

      Here’s what I’ve seen in practice — let me know if your org is similar.

      What usually happens is that the people creating dashboards (like us) get a LOT of requests for different variables, datasets, charts, slices of data, etc. That’s great at first. People are interested in the data; we want to be responsive and helpful to their requests.

      A year goes by, and then another. And the dashboard(s) have ballooned and ballooned and ballooned. There are so many variables, datasets, charts, and slices of data. It often becomes unusable.

      If this happens in your org, don’t forget to pare down again. DON’T ask people, “Which parts of these dashboards should we delete?” That causes unnecessary stress. (“What if we need that chart/dashboard/slice at some point…???”)

      INSTEAD, try asking, “What’s the most valuable part of this dashboard(s)?” You’re listening to see what’s *not* mentioned. Did users mention A, B, and D? Great! Then remove C.

  • Elizabeth McDonald says:

    That is helpful – thank you!

  • Leave a Reply to Elizabeth McDonald Cancel reply

    Your email address will not be published. Required fields are marked *

    You Might Like

    Our complimentary mini course for beginners to dataviz. Takes 45 minutes to complete.

    Enroll

    Redesigning a Thesis Chapter

    Farihah Malik had the opportunity to work with a public health agency, which she was really excited about. Until she had to present the research to a group of policy makers… Condensing two full chapters—73 pages of Farihah’s thesis—into a short report for the policy making group seemed like an impossible task.

    More »

    Want to wow your boss with a dynamic dashboard? Or, how about a one-pager of key findings? Our 4-course bundle provides all the how-to’s.

    Enroll

    Subscribe

    Not another fluffy newsletter. Get actionable tips, videos and strategies from Ann in your inbox.