Data-Driven Execution: Infrastructure

Building dashboards to monitor performance and make decisions sounds like an intuitive, non-controversial approach – not a lot to quibble with there. In this quote, Rabois is likely referring to business performance metrics (external) – this data is typically easier to come by even in the most dynamic environments. But building dashboards to manage operational execution has a different set of requirements and processes.

First, you’ll need an aligned work culture – more on that here. Second, you’ll need a data-focused tech infrastructure. For this discussion, I’m defining infrastructure as a mix of point solutions that capture internal workflow and data technology for storage and presentation.

Workflow & Work Quantification

Measuring execution starts with designing workflow processes so they can be captured inside function-specific software (aka point solutions). With the explosion of SaaS companies over the last 10 years, there’s now a software solution for almost every type of work activity and industry. In situations where they don’t exist (yet), there are dozens of generic task management tools that will capture progress metrics (Monday, Jira, Asana etc). These structured work applications are cheap and easy to use and create the data sets needed to start analyzing operational performance. 

In addition to their data capture capabilities, these tools create transparency for managers and their teams. Everyone has insight into workflows so they can contribute ideas on how to improve or identify friction points. They also help managers identify where value is being created inside their departments by phasing processes into smaller tasks and thereby isolating costs needed to execute.

Some Points about Point Solutions 

As I mentioned, there is productivity enhancing software for all types of work processes across all departments. The goal is to organize the work inside the software so it can export metrics (ie how long, how often, by who etc) for presentation and analysis. Couple of other factors to consider: 

1. The upside of using these systems is you are able to choose the best of breed solutions because you’re not forced to use only one vendor/platform – flexibility FTW! The downside is you now have dozens of different applications running inside your company – complexity FTL.
2. As the company matures, it’s possible to centralize around 1-2 platforms but you will be sacrificing nimbleness and flexibility; slowing teams down for implementation; and, you may find yourself struggling to quickly adapt in faster moving environments. ERP software is clunky and expensive to modify.
3. These tools typically have low-quality analytics and reporting capabilities that fall short of being useful. Even if they’re half-decent (rare), centralizing all operational data into one presentation layer (or dashboard) reduces friction and increases the likelihood that managers and team members will review the results. 
4. Due to #3, they must have data exporting capability. The functionality is quickly commoditized so data export becomes the killer feature for our needs.

Presentation Layer aka Dashboards

Once you have your teams up and running inside point solutions and workflow management systems, the next step is moving all the data created inside these tools into a central data repository. This is a basic task  but can take time to automate for a data team but once its built the maintenance time is typically minimal. 

There are many good options for the presentation layer and your choice is dependent on your specific needs, team capabilities and budget. In general, go with a tool that allows for customization and the ability for non-technical users to modify data sets easily – usually done via filters. There are new tools launched all time (Analytics/Data startups have been over funded in this last cycle IMO) so scrutinize any tool migration requests your data team pushes for. It’s a messy job, doesn’t always add business value and will frustrate end users. 

Once the team/department level dashboards are in place, your data team is in position to build higher-level dashboards for the C-team that summarizes the team/department level into a few key metrics. When that is done, you’ve taken a big step towards truly data-driven execution.