You don't have to have a fancy dashboard to measure your performance and do better.
Every public sector agency is trying to make better decisions, based in evidence rather than anecdote. Where is our time really going, what does the public really want, how should we manage this application backlog?
It's 2022: "data-driven" is the new black.
But every public sector agency I speak with also struggles with their data:
"Our system is old. I have to get IT to pull a report."
"Our data isn't good -- every staff inputs things differently."
Or simply: "We don't have data."
What do we do when data is unavailable, inaccessible, and untrustworthy?
Option 1: We throw up our hands and continue to make decisions based on anecdote and intuition. ¯\_(ツ)_/¯
Option 2: We buy a new database, take on a year-long RFP followed by a year-long IT implementation. Best case, we get good data in two years -- but more often, we get the same lousy data with a prettier interface.
Let's talk about Option 3.
Lo-fi data is still data
Back in San Francisco, several now-PPI partners worked with the maintenance staff at SF Rec and Park: when something broke in a park, the maintenance staff would get a work order to make the repair. But when those incoming fix-it requests were vague, unclear, or duplicative, the result was extra site visits and longer time to make the repair.
They wanted to improve the percent of work orders coming in to the yard that were "complete and accurate." But how do you measure a complete and accurate work order?
First, we needed a simple, agreed upon definition. We called a work order request "complete and accurate" if the shop lead could receive it and immediately dispatch staff to work on it without the need for clarification.
Second, we needed a way to capture this data, even though it was inherently subjective. The work order database was never going to tell us this information.
Instead, we designed a very simple form for each of the trade shop leads. This is all we asked them to write down:
The work order number and date
Was it complete and accurate?
If not, what was it missing?
We didn't ask them to do this forever -- just for one week. That's all we needed to understand what was happening:
60% of incoming work orders were complete and accurate
Of the incomplete or inaccurate work orders, half of them were incomplete due to just two reasons: the location of the broken thing wasn't clear ("where is that bench?") or it wasn't clear what the broken thing was made of ("is the bench metal or wood?").
This data collection took maybe 20 minutes per shop lead per day for a week, and another hour or two of data entry and analysis on the back end. Mostly on paper, none of it in the database. It was more than enough to solve the problem.
What's data for anyway?
When we facilitate work with clients or coach Lean Leaders, we always ask some version of these questions:
What problem are you trying to solve?
What's happening with that problem right now? How do you know?
That second question is where data comes in: Just how big is your backlog? How many days does it take to get approved for food stamps? How often is a prescription error happening? How many of your customers are frustrated, and why?
You don't need perfect data. You just need enough to tell you what is happening.
Here's the thing: We just need enough data to solve the problem. The data doesn't need to be perfect. It doesn't need to be complete. It just needs to be good enough to tell us what is happening:
What are people complaining about? Gather a week's worth of emails or phone calls to your generic inbox and categorize them!
How often do people request a service? Ask your front desk staff to count them for a week!
How long does the process take? For something short and frequent (e.g., applicants at a window), observe folks with a stopwatch for a few hours. For a long process (e.g., hiring, contracting), ask staff to estimate in days how long each phase takes.
Our Tally Sheet reference guide has more guidelines for how to get data that is quick-but-not-dirty.
Data indigestion? We can help.
At PPI, we love data: If you need help thinking about how to measure your performance or tackle an issue, reach out to us. We can coach you through fast, lo-fi solutions to your data problems, and when you're ready to build a dashboard or run an engagement survey, we've got you covered there too.
コメント