Objective
RockWare was growing very quickly. At the point at which we started working with them, they were hiring a new developer every week. RockWare looked at where their trajectory was headed, and wanted to implement a solution to avoid hiring one person for every person’s worth of productivity they had.
The interesting challenge was that RockWare wanted higher productivity, but not at a cost of quality. Churning out more software full of bugs was not the goal here!
Challenges (Current Reality)
RockWare, being a software product company, experienced the industry-typical, highly-variable environment. Managers reported about 5-6 hours’ worth of actual productive time per day from their developers. That was a fairly big opportunity for productivity gain!
Most of the remaining time was burned through the management of work — handovers, planning, teamwork, etc.
The biggest observable loss was the amount of waiting time between resources, i.e. when one team member finished a piece of work and handed it off to the next resource, they’d have to wait for that next resource to be available. This was exacerbated by the fact that the experts in the team were in highly valuable job roles, and presented as a bottleneck for the team; with handovers to or from these experts being critical time-loss opportunities. As a result, RockWare’s work in progress was higher than they would have liked.
People felt the pressure to be productive, so they’d go pick up a new piece of work to start working on, while waiting on the bottleneck. In turn this increased multitasking and decreased the flow of the team’s work.
Another challenge RockWare experienced was in relation to their review work. In general, the quality of the review decreases as the time between when the work is coded and when it gets reviewed increases, and the situation at RockWare was no different. The issue was that the time between coding and having the code reviewed was too large, and that time period needed to be shortened because it was critical to quality—the shorter the time, the higher the quality!
Process (Solution)
Traditional project management techniques had long been avoided at RockWare due to project management planning being out of touch with software development reality.
In software, estimating into the future becomes less and less reliable to the point of guessing. And committing to those guesses leads many software teams to be compelled by customer due dates and project manager techniques that enforce the wrong kinds of behaviours (e.g. rushing, ignoring quality steps, and cutting back on long term scalable structures).
Essentially, this equates to prioritising speed over quality, which was a strategic deal breaker for RockWare.
The solution implemented at RockWare involved:
- Design and implementation of a visible workflow board (i.e. a schematic of major development milestones)
- Managing work based on true time estimates, rather than an arbitrary week-based Sprint estimates
- Work managers to maximise the value added by expert developers (the team constraints)
- Time buffers to manage risk, rather than internal due dates
- Stand-ups focused on risk management, rather than trivial ‘who dunnit’ updates
Outcomes/Results
Managers reported that work flowed more smoothly, stress levels went down, and the team leaders and managers could focus on planning work more effectively, as well as getting more of their own development work done.
Upcoming work was visible, so as team members completed tasks, they were able to work on the next item in sequence—this smooth work system created a sense of calm among teams and managers alike.
The amount of unnecessary intra-company emails decreased, because the necessary notes were on the tasks, the relevant tasks were on the board, and the documentation made for clear trust and communication amongst staff and managers alike. Managers no longer needed to chase progress of tasks or projects, as all was visible on the newly implemented dashboards.
Using this system helped RockWare clearly know their own pace, what the team could accomplish, and how long it would take. This was fantastic for team members, project managers, and the C-levels alike.
Management noticed that team members quit panicking over lack of resources because they now had clear visibility of the organisation of the workflow, and clear visibility of the available capacity —that used to be hidden by high WIP.
RockWare went from hiring roughly a person a week down to a few developers over the entire year. The project paid for itself in manpower immediately. Once the initial team saw success with this process, it was rolled out to the rest of the company.
RockWare is committed to ongoing process improvements and continuously tweaks their standard operating procedures, and these changes are immediately implemented across the whole company, as all teams run the same system.