This message was deleted.

Such a good discussion topic–data visualization/storytelling can make or break all the hard work you put into an analysis or dashboard. Here are some rules I try to live by: 1. Boycott pie/donut charts! Okay, maybe this is a bit too extreme, but pie charts are really only good for a few use cases, and are often used incorrectly. Some common mistakes: Don't make a pie chart with more than 5 slices (it makes the visual comparison between slices of similar size too difficult), and make sure you're including all segments in your pie chart, otherwise you're misrepresenting the data. Read more about why pie charts often suck here. 2. Use colour sparingly: Don't add colour to a chart if it doesn't convey additional meaning. And think about accessibility when choosing your colours (choose palettes that are colour-blind friend). Here are some more great colour tips from Sigma. 3. Know your audience and design your story to cater to them. Is your dashboard for a C or board-level audience, or an operational team tasked with tactical optimization? Something you build for one audience is not going to suit the other, and these groups are going to care about really different things. Don't waste time over-engineering dashboard interactivity if the end-user is not interested in this type of UX. Gather requirements before you start building, and if you have time, design dashboard mock-ups and get feedback before you start development. This will save you time and rounds of revisions.

I feel like we’ve got some experts here. @fancy-megabyte-9510, @mysterious-scientist-27120, @sparse-car-47488, @bright-magazine-45146, @purple-sugar-23657, @agreeable-fish-22746, @busy-fireman-86564 care to weigh in?

Definitely with you @prehistoric-vase-57094 on boycotting the 🥧 and being intentional with color. Along these lines, having the right tool for building out a dashboard/visual is almost as critical as making sure you set your visual up right. Highly recommend checking out Equals.com for this — familiar interface, every graph is built on live data, and we have a bunch of customization options (without overwhelming). Graphs are almost always used (in my experience) as a conversation starter, with the expectation of diving deeper into a certain anomaly or point of interest. Being able to quickly show your work (from how you’re pulling your underlying data -> the analysis -> chart) can make metrics meetings 10x more effective.

I agree with you guys and Equals seem really great! As a marketer. i use Google Data Studio (now Looker studio) to create dashboards for all my clients. Here’s the most important elements for me: 1 - Less is more. My clients are not data or marketing experts, they are business owner or marketing director. They need a dashboard with only the most valuable metrics. I try to have less than 10 metrics by dashboard. Curation is key. 2 - Time is precious, make it clearer. They need to understand what going on in 30 seconds. I love Looker Studio comparaison date option so you understand very quickly the performance of the previous period. 3 - Make the dashboard for your client, not for you. The goal is not to impress your data friends by making a CrAzY dashboard but to make a tool used by decisions makers to improve their business.

@broad-agency-96135 all great points! How do you think about managing alignment and knowledge of your KPI definitions? (do you keep definitions on the dashboard somewhere, or is there separate documentation for this)...that one has always been a hard problem to solve. It's been a while since I've used Data Studio, but Looker itself doesn't do a great job of exposing the KPI definitions or calculations at the dashboard/Look level, only at the Explore level.

Very true @prehistoric-vase-57094! My dashboards usually have link to an external document with all the KPI definition.