Process mining is being used to automate multiple processes in finance and supply chain and driving real bottom line value, but Bloomfilter is betting that the technology can revamp the software development lifecycle.
In an interview with DisrupTV, Andrew Wolfe, co-CEO of Bloomfilter, and Erik Severinghaus, co-CEO & co-founder of Bloomfilter, outlined how process mining can be used for the software development lifecycle. Bloomfilter recently raised $7 million in seed funding.
Process mining has been the domain of CFOs and CEOs and continuous improvement teams. Process mining takes logs from enterprise systems to show how work is actually done (compared to how it works on a whiteboard), and companies can make processes such as accounts payable, accounts receivable, order-to-cash, hire-to-retire and logistics more efficient. Since many of these processes revolve around real cash savings enterprises can drive top line and bottom line returns.
Software development has been a different animal, said Severinghaus. "Anybody that oversees and manages a complicated business process knows that there's a ton of waste and inefficiency," he said. "The hard part is always figuring out where that is so that you can root it out and make it more efficient."
Severinghaus added:
"The product development side has always been too hard of a challenge since you can't just go take a typical process mining tool and connect it into the various different systems. There's are a dozen different systems along the product development roadmap. You have everything from Figma up front where you're doing design all the way to Jira. There are different tools across the entire lifecycle to stitch together."
The good news, said Severinghaus, is there is a lot of data to analyze, but it has to be normalized and focused on product development. Bloomfilter has a few patents with more in the works for product development process mining.
Wolfe said there's a big opportunity to automate software development processes. In software development, the typical analytics are lagging indicators such as velocity. "What's missing is why you got there. You may know velocities but don't know where you're going or how you got there," said Wolfe, who added that process mining can enable you to have conversations to better manage sprints. "What we find is a lot of people we talk to are tracking processes in Excel. Some are using PowerPoint. I've seen some amazing Mona Lisa level spreadsheets."
Software development has had a limited picture across the entire process across Figma, product board, Jira, GitHub to AWS. "The indicators are too lagging to make any kind of strategic decisions," said Wolfe, a Business Transformation 150 alum.
Ray Wang, CEO of Constellation Research, asked Severinghaus about tips when considering process mining for software development. He outlined the following:
- Drop the fixation on developers. Severinghaus said the first question Bloomfilter gets revolves around engineers and whether they're working multiple jobs. "There's an obsession that these really expensive developers are just screwing around and that's the reason software isn't getting built," he said.
- Focus on the overall software development life cycle. "We show our clients over and over again that developers are a small piece of the software lifecycle," said Severinghaus. "You may have amazing developers sitting downstream from terrible requirements. It doesn't matter how great and fast developers code or how much AI you apply if the requirements aren't right."
- Know the processes before hiring more engineers. Hiring 10 more engineers won't do you much if your software development process is hung up in QA. "Let's say you create 100 applications a day, but you still have to test them to make sure they're valid and you still have to deploy them," said Severinghaus.
- Beware scaling bad code. AI can speed up the software development lifecycle and deployments, but without a holistic process approach enterprises may just scale bad code at a speed humans can't comprehend, said Wolfe. You have to attack the bottlenecks but keep a holistic view.
Constellation Research analyst Liz Miller asked Bloomfilter's co-CEOs whether they were going to apply process mining to other art-meets-process disciplines like marketing. The short answer is no, but clearly there's a market to apply process mining beyond the usual finance processes.
Here's the full episode with the Bloomfilter interview starting at the 22 minute mark.