Dashboard Design Part 3 of 3: An Iterative Tale
On Monday, we covered the first chapter of this bedtime tale of dashboard creation: a cutesy approach that made the dashboard into a straight-up reflection of our sales funnel. Last night, we followed that up with the next performance management tracking beast — a scorecard that had lots (too much) detail and too much equality across the various metrics. Tonight’s tale is where we find a happy ending, so snuggle in, kids, and I’ll tell you about…
Version 3 – Hey…Windows Was a Total POS until 3.1…So I’m Not Feeling Too Bad!
(What’s “POS?” Um…go ask your mother. But don’t tell her you heard the term from me!)
As it turned out, versions 1 and 2, combined with some of the process evolution the business had undergone, combined with some data visualization research and experimentation, meant that I was a week’s worth of evenings and a decent chunk of one weekend away from something that actually works:
Some of the keys that make this work:
- Heavy focus on Few’s Tufte-derived “data-pixel ratio” –- asking the question for everything on the dashboard: “If it’s not white space, does it have a real purpose for being on the dashboard?” And, only including elements where the answer is, “Yes.”
- Recognition that all metrics aren’t equal –- I seriously beefed up the most critical, end-of-the-day metrics (almost too much – there’s a plan for the one bar chart to be scaled down in the future once a couple other metrics are available)
- The exact number of what we did six months ago isn’t important -– I added sparklines (with targets when available) so that the only specific number shown is the month-to-date value for the metric; the sparkline shows how the metric has been trending relative to target
- Pro-rating the targets -– it made for formulas that were a bit hairier, but each target line now assumes a linear growth over the course of the month; the target on Day 5 of a 30-day month is 1/6 of the total target for the month
- Simplification of alerts -– instead of red/yellow/green…we went to red/not red; this really makes the trouble spots jump out
Even as I was developing the dashboard, a couple of things clued me in that I was on a good track:
- I saw data that was important…but that was out of whack or out of date; this spawned some investigations that yielded good results
- As I circulated the approach for feedback, I started getting questions about specific peaks/valleys/alerts on the dashboard – people wound up skipping the feedback about the dashboard design itself and jumping right to using the data
It took a couple of weeks to get all of the details ironed out, and I took the opportunity to start a new Access database. The one I had been building on for the past year still works and I still use it, but I’d inadvertently built in clunkiness and overhead along the way. Starting “from scratch” was essentially a minor re-architecting of the platform…but in a way that was quick, clean and manageable.
Looking back, and telling you this story, has given me a chance to reflect on what the key learnings are from this experience. In some cases, the learning has been a reinforcement what I already knew. In others, they were new (to me) ideas:
- Don’t Stop after Version 1 — obviously, this is a key takeaway from this story, but it’s worth noting. In college, I studied to be an architect, and a problem that I always had over the course of a semester-long design project was that, while some of my peers (many of whom are now successful practicing architects) wound up with designs in the final review that looked radically different from what they started with, I spent most of the semester simply tweaking and tuning whatever I’d come up with in the first version of my design. At the same time, these peers could demonstrate that their core vision for their projects was apparent in all designs, even if it manifested itself very differently from start to finish. This is a useful analogy for dashboard design — don’t treat the dashboard as “done” just because it’s produced and automated, and don’t consider a “win” simply because it delivered value. It’s got to deliver the value you intended, and deliver it well to truly be finished…and then the business can and will evolve, which will drive further modifications.
- Democratizing Data Visualization Is a “Punt” — in both of the first two dashboards, I had a single visualization approach and I applied that to all of the data. This meant that the data was shoe-horned into whatever that paradigm was, regardless of whether it was data that mattered more as a trend vs. data that mattered more as a snapshot, whether it was data that was a leading indicator vs. data that was a direct reflection of this month’s results, or whether the data was a metric that tied directly to the business plan vs. data that was “interesting” but not necessarily core to our planning. The third iteration finally broke out of this framework, and the results were startlingly positive.
- Be Selective about Detailed Data — especially in the second version of the scorecard, we included too much granularity, which made the report overwhelming. To make it useful, the consumers of the dashboard needed to actually take the data and chart it. One of the worst things a data analyst can do is provide a report that requires additional manipulation to draw any conclusions.
- Targets Matter(!!!) — I’ve mounted various targets-oriented soapboxes in the past, but this experience did nothing if it didn’t shore up that soapbox. The second and third iterations of the dashboard/scorecard included targets for many of the metrics, and this was useful. In some cases, we missed the targets so badly that we had to go back and re-set them. That’s okay. It forced a discussion about whether our assumptions about our business model were valid. We didn’t simply adjust the targets to make them easier to hit — we revisited the underlying business plan based on the realities of our business. This spawned a number of real and needed initiatives.
Will There Be Another Book in the Series?
Even though I am pleased with where the dashboard is today, the story is not finished. Specifically:
- As I’ve alluded to, there is some missing data here, and there are some process changes in our business that, once completed, will drive some changes to the dashboard; overall, they will make the dashboard more useful
- As much of a fan as I am of our Excel/Access solution…it has its limitations. I’ve said from the beginning that I was doing functional prototyping. It’s built well enough with Access as a poor man’s operational data store and Excel as the data visualization engine that we can use this for a while…but I also view it as being the basis of requirements for an enterprise BI tool (in this regard, it jibes with a parallel initiative that is client-facing for us). Currently, the dashboard gets updated with current data when either the Director of Finance or I check it out of Sharepoint and click a button. It’s not really a web-based dashboard, it doesn’t allow drilling down to detailed data, and it doesn’t have automated “push” capabilities. These are all improvements that I can’t deliver with the current platform.
- I don’t know what I don’t know. Do you see any areas of concern or flaws with the iteration described in this post? Have you seen something like this fail…or can you identify why it would fail in your organization?
I don’t know when this next book will be written, but you’ll read it here first!
I hope you’ve enjoyed this tale. Or, if nothing else, it’s done that which is critical for any good bedtime story: it’s put you to sleep! 🙂