Data from Analyzing 600M Hours of Work

Posted by | April 21, 2019 | Thoughts on Enterprise Software | No Comments

Part of my job blends the qualitative and quantitative analysis of customer behavior. Either route can kick start a hypothesis that sends me down a rabbit hole. I’m a data nerd and love diving deep into the numbers so this most recent endeavor has been fascinating to me.

It all started with a theory that my colleague and I had: people don’t manage projects, they manage process. It’s a simple statement but is hugely profound when it comes to organizational change. Managing projects implies managing the activities within a specific timeframe. Managing process implies managing the deliverables and outcomes that run through the process, unblocking efforts to produce outcomes.

We spent a lot of time onsite with some of our largest customers and started to hear similar ways of them bending our system. You know you’re onto a fundamental shift in behavior when your users rigorously use your product for something that you didn’t intend for them to do. These customers use highly repeatable processes to get a significant amount of their work done. They leverage process because it helps with optimization since you have clear stage gates where blockers can be tackled.

We also heard this from industry analysts too. The largest companies in the world are shifting away from a typical PMO model to a more agile (little “a”) process driven model where the key output is deliverables that they measure a change in outcome against (eg. Did this deliverable drive a difference in outcomes). These enterprises are having to get really good at this because they are getting disrupted by smaller companies who can blitz their strategy better than they can.

We found this discovery fascinating and decided to dive deep into the data. Our system has tracked over 600 million hours worth of work, so we have a good sized data set that we can rely on for significance. Here’s what we found.

75% of all the work in our system comes from repeatable processes. Specifically, for our enterprise customers, that number is higher at around 80% on average. There are certain industries (eg. creative, finance, legal) that have even higher numbers at 90%+ of their work being process driven. The average number of unique processes an enterprise has is ~100 (ones with quantifiable usage).

The average process has 12 steps. Even more than that is that over 77% of the work performed in these enterprises comes from just 10 processes or less. I’ll note that these aren’t just small companies either. These are your Fortune 500s with thousands and thousands of users in our system on a daily and weekly basis. This last fact just blew my mind. You’re telling me that 77% of the work being performed (on paper) comes from just 10 or fewer processes? Crazy.

Some other interesting heuristics we found were around working behaviors. SMBs typically have activity lasting between 7am and 4pm. Enterprises between 8am and 6pm. Government between 9am and 4pm. The highest levels of activity are Friday (closing work out) and Sunday (prepping work for the week).

These were fun tidbits of data that we discovered through the process. For those in the software world, it would make sense that project management is dying. Our world doesn’t really recognize that method of work. However, for the rest of the world, this is a huge shift that has been caused by the disruption of software. They’re taking a queue from the software world on speed and agility, trying to inject that into their own organization.

This is an active journey so I suspect that I’ll have many more data points that will come to play. I’ll make sure to update them here as we dive deeper.