It’s always a good laugh to joke about some of the crazy requests we get in regards to requirements, but it’s also good to have a plan when these “jokes” arise. Here are a few things I’ve learned over the years.
We’ve all been there! You are sitting at your desk innocently working away when all of a sudden a meeting invite arrives with the subject line “Dashboard Requirements Meeting”. Your first instinct may be “awesome, a new project AND they have requirements, or at least want to talk about them! Better than the last time when they just gave me a dataset with no requirements and say ‘make a dashboard’ right?”
You arrive at the meeting and the discussion begins… oh no… now what?
Can you make this look like Excel?
Over the summer I wrote an extensive blog post on this very topic mainly because it is the most common request I’ve seen. The TL/DR of that post is that you have multiple strategies you can employ to avoid this request which includes:
- Find someone above the requestor in the hierarchy who receive this data and see what they want and try and convince them to go visual
- Find an evangelist who loves data visualization and will help encourage their peers to move away from the tabular approach
- Present an alternative view – create their tabular request and an actionable dashboard, and present the dashboard first.
- Subtly add some preattentive attributes to their tabular report so you ease them into data visualization
This is hard do to the fact that people hate change and Excel is the warm comfy blanket that comes with hot chocolate and marshmallows… and no one wants to have their blanket and hot chocolate taken away from them! It’s like dealing with a small child, encouraging them they don’t need a night light, and then they don’t need you to tuck them in. Do it gradually and eventually, they will forget all about tabular reports… maybe.
What I want is a pie chart (or whatever chart is the flavor of the month)… OR This visual has to be this chart type! (It’s completely the wrong chart type)
When stakeholders are asking for specific chart types, or the wrong chart types it can be frustrating. This is all about education and again, comfort with what they may know. Pie charts are as common as dirt, and generally, they are as replaceable with something better. Educating your stakeholders on “Right Chart, Right Time” can be a huge game-changer. My favorite tool for this has become the Visual Vocabulary created by Andy Kriebel (adapted from the Financial Times). Set it up as a lunch and learn, based around data literacy, for anyone who wishes to attend is possible. Talk about chart selection, with a good dose of “it depends”. Stakeholders are demanding, but the more knowledge they have, the better chance you have. Hopefully, you get to a point where they just ask the question of the data, and leave the design to you.
As a side note, if they are asking for a new chart type, and it makes sense for the data question being asked, make sure you include somewhere on the dashboard a “rosetta stone” on how to interpret the chart so it’s consumed correctly.
Please color each category (there are 20+) on the dashboard.
Overuse of color is such a common issue, mainly because it so so easy to do in Tableau! You can drop any field on the Color of the Marks Card and “BAM!” you’ve got color. If everything is colored, nothing is colored. My advice, if coloring a value is required give the user a mechanism to highlight one or two categories at a time. Using a parameter in this situation allows color to be applied with care, and adds a nice level of interactivity. If they have particular values they always look at, encourage them to set up custom views on Server for quick and easy access.
I want all these metrics on a single dashboard (there are 20+)
Here is a very “Zen Moment” question. If all your metrics are on one dashboard, but you have to scroll horizontally and vertically to see the content, is it really on “one dashboard”? Do you scroll on your car’s dashboard to see all the metrics? NO! Best practices suggest to us that we put a maximum of 6 metrics on a dashboard (as always, it depends) so that the user doesn’t get analysis paralysis. If the metrics are not related, and the potential for false correlation is present, those items should be separated on to different dashboards. With the addition of navigation buttons in Tableau, moving between those dashboards is a breeze! You can even use actions to move between dashboards if you need to drill down or filter subsequent dashboards. You are only limited by your creativity, so use it!
How do I export to Excel?
Why do you want to export it to Excel? That’s the question. If they want the ability to build their own analysis based on the data, maybe set them up with an Explorer license and some Web Authoring training! Keeping the data in Tableau is the best way to manage your data governance because once it’s in the wild you have no idea what’s happening to it. Giving the users the ability to self-serve in a governed area is an amazing option over exporting it to Excel. If they say they don’t have time for that, ask them how long they spend playing with the data in Excel. If the answer is “I copy and paste it into a template and I have this set of visuals simple beat them to death with a blunt object ask them to see if that can be created in Tableau so you can save them some time.
How do I export to PowerPoint?
This is inevitable, and I’d be lying if I said I have a good answer for this. Tableau makes it easy to export all your dashboards to PowerPoint with the simple click of a button. With that said, do they carry an iPad or another device? Maybe creating a mobile device layout will entice them to use that device in the meeting to display their metrics. By giving our users options like this, we can help combat the desert of static dashboards. Encourage them to present the data live, so that if questions occur then they can filter and manipulate the data live. If they insist on PowerPoint, while it’s not the most elegant of solutions, you can embed Tableau dashboards directly into PowerPoint and have them be interactive… check out this blog post on Clearly and Simply
I want to be able to filter by Customer (there are thousands!)
One of the things that can kill Tableau dashboard performance is high cardinality quick filters. Allowing for the filtering at extremely low levels of granularity is not ideal, so you need to recommend alternatives. One alternative is allowing the user to drill down to a set of customers, either by region, segment, account manager… literally anything that will allow a pseudo-hierarchy to be utilized. You can also create sets that contain certain customers… maybe it is a “VIP list”, “watch list” or “at-risk” list that you can ad as a filter option. Again, creativity is key.
I need a dashboard that has all our data points so I can answer all my questions…
This is yet another tell-tale sign that they need Explorer access to Tableau Server and a few lessons on Web Authoring. It is impossible to create the “Be All End All” dashboard. Add to that, the performance of a dashboard built in this manner is probably not going to perform well at all. Setting up data sources for this stakeholder that can answer sets of like questions will go a long way in satisfying their needs.
I need a “Stop Light” scorecard/dashboard (Red/Yellow/Green)
Ah yes… stoplights. Building your vizzes to be as accessible as possible should be a goal we all strive for; this is one of the most common pitfalls. The only executives who don’t ask for this type of visual are those who are color deficient. And sometimes, even those who are color deficient ask for them. I had a former boss, who was, do this to me. After I delivered it, he mentioned he could differentiate what the dashboard was telling him because he was color deficient…
So, I wrote an earlier blog post about allowing for the highlight of the good, or the bad, for added flexibility. But if they really need to see the red and the green together, make sure and include an indicator to show whether or not it is good or bad. Unicode is a popular option, or you can use custom shapes inside Tableau (or add your own). There are also shades of red and green that are more “color deficiency friendly” than others. It just takes a bit of research and testing.
When all else fails…
It’s easy to get frustrated, I know because I have in the past. The key is to get creative. You are a professional, you know what’s “best” in data viz and you want to share those skills with those who need it most. Fight the good fight, share your knowledge, educate the masses and move forward. One great advantage we have in the Tableau Community is truthfully each other… when stumped, ask the question!
Special thanks to all the #datafam for indulging my thread on Twitter on 2/7/2020, you added a couple of great examples to this blog post!