Resources / Blogs
Resources / Articles

Blogs

Service Delivery Review: Predictability

Service Delivery Review: Predictability

Welcome to our series looking at using Flomatika to run a Service Delivery Review. Today’s topic is going to focus on ‘Predictability’. Within Flomatika, we show your predictability across two performance indicators: Lead Time and Throughput. Today we’re going to focus on Lead Time Predictability.

Lead Time Predictability

When talking about Lead Time predictability, we’re essentially looking to see if there's a high or low variance in how long it’ll take for a delivery system to deliver on a piece of work.

Value Stream Health Dashboard

We can see within this Program, they have a low Predictability around their Lead Time. 

Flow Lens

Let's deep dive in more detail and see what we can discover. We can do this by navigating to the Delivery Management page. Lead time is based upon work that has been completed in the past, so we will be looking in the Completed Work perspective. We can conduct more detailed analysis on the individual work items by clicking on the Flow Items Types tab.

Product Management Flow Analysis Dashboard

Bug Analysis

We can now conduct more detailed analysis on the individual work items. 

Firstly let’s look at Bug. We can see for the time period selected: 

  • 46 bugs were completed with the service level expectation of 3 days or less
  • Target was met just under half the time (48%)
  • They are trending upwards
  • However, they have a Low predictability 

We can see their lead time distribution in the columns on the right. 

  • We can see the medium (50th percentile) of 4 days. So half the time they're completing bugs in four days or less.
  • On average they are completing them in six days or less.
  • The majority at a time (85th percentile) it's completed in 9 days or less.
  • With a tail (98th percentile) of 30 days or less.

What does this tell us in terms of Predictability? When a Bug is started by this team, half of them are completed in 4 days or less; on average it can take up to 6 days; the majority of the time they are completed in 9 days or less; but can take all the way to 30 days to resolve. That variance of 4 days (50th percentile) vs 30 days (98th percentile) is what categorises Bug as having Low Lead Time Predictability.

Enhancements & Optimisation Analysis

Let's run the same analysis, this time with Enhancements & Optimisations. Let's try and keep the analysis from a customer perspective. Firstly, the expectation is when a team starts on an Enhancement & Optimisation item, it should be completed in 7 days or less.

For this team, they are meeting this expectation 64% of the time, almost two in three. Not great, with plenty of room for improvement, but they are trending upwards. The majority of the time, they complete them in 17 days or less. However, these can take up to 95 days! So if the work is not done within 17 days, it could take anywhere up to 3 months to get done! That's a big variance. Which is why Lead Time predictability is considered low.

Lead Time Distribution

For further analysis, you can see two different visualisations of Lead Time on the Histogram and Scatter Plot located in the Lead Time distribution accordion. 

Lead Time Value Stream Management Metric

Let's take a closer look at the histogram, which shows the number of occurrences for your Lead Time (in days).

Lead Time Value Stream Management Metric

In terms of predictability, we typically look for the shape of the ‘tail’ on the Histogram. You can read more about this in this blog, but in summary:

  • Fat-tailed distribution results in high variance in your lead time, which we have translated as low predictability. 
  • Thin-tailed distribution results in low variance in your lead time, which we have translated as high predictability.

To assist teams in working towards shifting their Lead Time predictability from low to high, we show a ‘Predictability Target’ line. The items to the right (Histogram) or left (Scatter Plot) of the red ‘Predictability Target' line is what is causing your team to be rated as having Low Lead Time predictability. ‘Trimming the tail’ is something the teams can work towards to shift themselves towards High Lead Time predictability

Don’t forget to leverage the range of filters available on Flomatika. By filtering by work item type, we can see the histogram for just Bug, or just Enhancements & Optimisations.]

Lead Time Value Stream Management Metric
                                                                                                           Bug
Lead time Value Stream Management Metric
                                                                          Enhancements & Optimisations

As you can see they are very different shapes. The nature of how these work items are worked on the team’s context may require different approaches. This is more obvious when conducting analysis on Portfolio level work items such as Epics:

Lead Time - Value Stream Management Metric
                                                                                                  Epics  


By their nature, they operate on a different time scale to the Team level work items. Once again, based upon your context, may require different approaches. In this case, the Epics are classified as having High Lead Time predictability, as the ratio between the Median (39 days) and Tail (143 days) is considered not too wide.

Once an Epic is started, half of them are done in 39 days or less. Most are done in 100 days or less. If they fall outside this range, it can take up to 143 days. So your longest time is just over three and half times the time it takes for half of them.

Learn More

You can learn more by visiting these related links:

Want to see your own data in Flomatika? Arrange a Free Trial.

Contact Us

More

Blogs