Four Ways To Improve Predictability (Tip #2)

This is a continuation of a four-part series.

• Part #1: Narrow your planning horizon
• You are here --> Part #2: Use empirical data
• Part #3: Create smaller teams
• Part #4: Eliminate the fear of failure

Tip #2: Use empirical data

Building strong plans based on solid data seems like common sense, right? Unfortunately not so much. How many times have you had a leader dictate the project deadline and scope and budget, only to then ask for a bottoms-up plan to achieve said goals? Or, how many times have you witnessed the smartest guy in the room commit to work estimates without consulting the people actually doing the work? All too often we shoot ourselves in the foot by creating plans based on hopes and guesses, rather than data.

Measure what you want to improve

Sustained team predictability results in overall organizational predictability. As such, this article only focuses on improving team predictability. That said, if you’re seeking a business outcome then ask for it, measure it, and frequently publish metrics. Here are some other examples of business outcomes and related measurements:

Outcome: Improve predictability from 40% to 80%. Measurement: Do/Say ratio; work forecast vs delivered

Outcome: Maximize business value delivered. Measurement: Average priority of completed work (assumes 1-n ordered backlog of business priorities)

Outcome: Increase projects completed. Measurement: Work in progress ratio; backlog items complete per cycle / backlog items in progress per cycle

Source: QuickScrum.com

The predictability equation

Predictability = story points + Fibonacci sequence + relative sizing + consensus-based estimation + velocity-based planning

Using story points for sizing can vastly improve team predictability. Story points are a measure of effort, volume, risk, uncertainty and complexity of the work. Ask two people how long a task will take and you may get very different answers. Ask them how much effort is involved, and you’re more likely to get similar answers. Story pointing is more accurate because it takes into account so many more variables than a traditional time-based estimate.

When story pointing, a best practice is to use the Fibonacci sequence (0,1,2,3,5,8,13,21…) for sizing. The distance between Fibonacci numbers are orders of magnitude, which simplifies the estimation process by eliminating pointless debate between units that are less than an order of magnitude. If you want to nerd-out, see Weber’s Law, which describes how humans perceive physical magnitudes.

Another best practice is relative sizing. In relative sizing a reference story is selected by which all other stories are sized against, leading to more consistent estimations. Relative sizing is best used in combination with the Fibonacci sequence.

Source: Amazon.com

Leveraging a consensus-based estimation technique such as Planning Poker will lead to even more accurate estimations. Using this method, the whole team participates in estimation, resulting in active dialog across differing points of view. Furthermore, estimates are chosen privately by participants and then revealed at the same time, eliminating bias.

Team velocity is the average number of story points the team can complete in a cycle. It take a few cycles of estimating with story points before teams will establish a velocity. Once a team establishes a reliable velocity, it becomes very easy to create an accurate plan each cycle. A stable velocity is the holy grail of predictability.

Benefits

1. More accurate work estimates
2. More consistent work estimates
3. More accurate plans
4. Improved predictability

How to get started

1. Engage your teams in a dialog about these estimation practices. Start by communicating the business outcomes you’re seeking. Ensure that the team understands why these methods could help achieve the desired outcomes
2. Provide training so the team knows how to apply the practices
3. Leverage planning poker, relative sizing using the Fibonacci sequence, and consensus-based estimation
4. Ensure that the team doing the work is the team that estimates the work
5. Ask teams to establish a velocity and leverage it in planning

Pitfalls

1. While it’s perfectly fine to expect teams to have stable or ever-increasing velocity, never over-inspect story points. Story points and velocity information are intended to help the team in planning; any use beyond that should be scrutinized
2. Never compare story points between teams. Story points are subjective by design; what is sized as an 8 for one team, may be sized as a 3 for another team
3. Never align story points to work-hours. If you do so, you’re unlikely to reap the full benefits of story pointing

Let’s discuss

What types of empirical data does your organization use for planning? What has your experience been with story pointing?