“Eating your own dog food,” “ice creaming,” “drinking your own champagne” — whatever you want to call it — refers to a company using its own product, either to test drive it or to show confidence in it. Well, here at Tasktop we have another reason for using our own products: We actually really like them and what they do for our business. In this blog series, we’ll be sharing our stories on why we’re finding our dog food so tasty.
For the first installment in the Tasty Dog Food Series, I spoke to Rebecca Dobbin, our firecracker Product Manager based out of Austin, about how she uses Planview Viz™ to maximize the capacity of her engineering team for the best possible business outcomes. Planview Viz is a value stream management (VSM) solution for software delivery that measures and optimizes the flow of business value across end-to-end collaboration networks.
“As a Product Manager, I want to be able to focus as much capacity as I can towards delivering new product capabilities that meet customer needs,” Rebecca told me. “One of the ways that I measure how much we’re delivering for the business is through Flow Velocity.”
Viz lets you see Flow Velocity split into four types of business value:
- Merkmale: New capabilities that drive business growth
- Defects: Fixes for quality problems that affect the customer experience
- Die Risiken: Work to address security, privacy, and compliance exposure
- Debts: Work to improve future delivery flow by addressing process, skill or technical debt
As a product manager for a growth-stage product, Rebecca is trying to maximize Feature velocity, while reserving adequate capacity for the necessary Defect, Risk and Debt work that ensures a product’s longevity. To help keep Rebecca’s eyes on the prize, Viz measures the value stream’s flow side-by-side with the business results she aims to achieve.
“A central aspect of value stream management is looking at your Flow Metrics in conjunction with your business results, tying your team’s delivery to the outcomes that matter to the company,” Rebecca says.
As with any SaaS product, a key business result is uptime, and the team is very conscious of keeping the number of production incidents low and resolving them quickly within their SLA.
Back in April, the Viz dashboard was showing Rebecca that Feature Flow Velocity was declining in proportion to the number of production incidents per month. Turns out that production incidents were having an outsized impact on Feature Flow Velocity.
“I noticed something interesting: Every time we had a production incident, Flow Velocity would decline by an average of five Feature stories.” While it’s natural to take a velocity hit when facing an incident, five Feature stories felt like too much. “Everyone was dropping everything and swarming the issue. Their heart was in the right place, but was it really efficient?” Rebecca asked. “There had to be a way to reduce the blast radius of the incident response such that we minimized the impact on our flow as a whole.”
“I never would have been able to quantify the impact of a production incident had I not been measuring our flow and tracking the business result,” Rebecca told me. “Viz helped me identify the problem. The next step was to brainstorm with the team and experiment with different ways of handling incident response. We decided that the assigned engineer would try to work on it in isolation as much as possible, to minimize the context switching for the rest of the team.”
To track whether the process change had the desired impact, Rebecca added timeline events to her Flow Metrics to indicate when the experiment began and monitor whether she sees an improvement in Feature Flow Velocity going forward.
“That was step one. The next thing we wanted to tackle was what we could do to resolve incidents faster. Because any time saved on incident resolution can be devoted to Features. That’s where we began focusing on Defect Flow Time — that’s the total time it takes to resolve a production incident start to finish.”
Rebecca and the team discussed this in one of their retros, using Viz to drill into the incidents with the longest Flow Times to understand what they had in common.
“We noticed that when an incident involved infrastructure changes, resolution times were much longer. There was a clear business case here for more self-service when it comes to simple things, like increasing memory.” They worked with their colleagues from the Cloud product value stream to get this set up, and once again, they monitored the impact of the change on their metrics through timeline event tagging.
Through these two initiatives, Rebecca’s product team was able to mitigate the impact of incidents on Feature Flow Velocity while also improving resolution times for incidents.
“We saw great results after running these two experiments,” Rebecca said. “The Flow Velocity impact of an incident has been reduced from five flow items to just one, which is natural and I can live with”, Rebecca says with a smile, “and the Flow Time for incidents has improved by 50 %.”
“Not only did we find a way to protect our Feature capacity, but we also found a way to improve customer satisfaction. And all this was made possible by measuring Flow Metrics side by side with business results, monitoring both, and actively looking for opportunities to optimize.”
More in the Tasty Dog Food Series
- Improving Delivery Predictability for Features and Defects and Boosting Team Engagement
- How We Built the Case for Application Replatforming Using Flow Metrics
- Fighting pandemic fatigue by safely experimenting with engineering team structures
Take a Course in Flow Metrics
The Flow Institute offers a range of courses on the Flow Framework® and Value Stream Management.
In this on-demand course by Dominica DeGrandis (bestselling author of Making Work Visible) introduces the Flow Metrics, providing a deeper dive into what they are and why you need them. The course explains the theory behind Flow Time, Flow Velocity, Flow Efficiency, Flow Load and Flow Distribution.
This course is provided on-demand and includes video lessons by Flow Experts and in-course quizzes. Login details will be provided within 24 hours of registration.
