It's Wednesday which means day 3 of dashboard week! The excitement in these blog posts is waning, I can only apologise. It's tricky getting so many blog posts out in one week and making sure the quality is still of at least somewhat decent standard. Anyway, excuses aside, I think today's final product was probably the one I was most proud of so far this week. Formatting makes the world of difference which could be why the other two didn't feel as good but this one I spent time to make look good (hopefully it shows in the feature image!). There were some minor oversights but in general, I'm pleased with how it turned out. My expectations may also have lowered since day 1 which shows progress too, at least that's what I tell myself. So let's get on with talking about the day!
The Task
We began the day with a blog post from Serena explaining the task at hand. You can find all the details here. The key information is that we needed to focus on:
- Communicating campaign success clearly and efficiently
- Highlighting key performance indicators (KPIs) relevant to marketing performance
- Allowing the stakeholder to quickly assess what is working well and where performance may be lagging
Planning
The phrase KPI dashboard is used somewhere in our instructional blog post which I'll admit, threw me for a bit of a loop. I know what KPIs are but what exactly is a KPI dashboard? Surely not a dashboard purely composed of KPIs? Well, google images seemed to think not thankfully so I decided to basically go with a KPI heavy dashboard with charts at the campaign level to supplement my vision. I drew it out (by hand, sorry there's no excalidraw evidence) with filters, KPIs and charts and got down to business.
Charting
Then my plan went sideways. After putting my vision into charts, I realised that I wasn't really showing anything. So I thought more deeply about how to translate this information into insights that the stakeholder could quickly assess. I didn't want the charts to be too interactivity heavy as otherwise that would rely on the stakeholder selecting the right filter, I wanted the information to jump out. That's when I went to my new favourite chart - the trusty dot plot. The dot plot is the upgraded version of the bar chart (in my opinion, may not be factually correct although it was once recommended by Carl [the global head coach at the Data School if you didn't know who that was] to show distribution more clearly so the idea holds weight!). A dot plot is great because it can show the distribution of a value without the need for a drilldown. Thus, in my charts, I could see how each campaign varied across each platform for a specific value (e.g. engagement rate), rather than showing engagement rate overall and then drilling down to the platform level. How much more helpful is that?! So, I created two dot plots for each value I felt most clearly assessed campaign performance, one for target audience x platform and the other for campaign x platform. But then I came to the sticking point of not being able to show how campaign interacted with the target audience so I decided to break things down even further, demonstrating how the campaigns performed at the platform level specifically for each target audience.
Dashboarding
Thus I ended up with 5 charts per measure which was just too many charts to fit into my original plan of one giant KPI dashboard. So I decided to go with something different, that I didn't know all that much about, and use navigation buttons with each page showing a separate measure as well as having an overview page with the KPIs. I kept on with the pink and purple theme from yesterday but made it prettier! I also ensured the formatting across the pages was the same, made the chart backgrounds more transparent and just generally tried to clean it all up. As I've said many times before, it's much easier to sell a good-looking product because it practically sells itself.
Reflections
- I should do more research into the topic before creating a dashboard. I think my understanding of what a marketing executive might want to see in a dashboard to make it as effective as possible is somewhat limited. I think it would have been beneficial to have done some research into what key calculations are used in marketing to make it as useful as possible.
- Power Query profiling does not show the whole picture. I know this and yet, I ignored it anyway and made some oversights in showing my yearly data.
- Formatting does make the world of difference. I also knew this but think it deserves reinforcing because my dashboard was not a pretty sight before and now I think it looks at least somewhat decent.
That's all from me for the day. Just one more dashboard left - a 2 day project so should be exciting (I know, I'm so cool). Until then, happy dashboarding!