Agustín Benito Bethencourt
You can go fast by going
together: software delivery
process performance metrics
Who am I?
● Experienced manager in open/remote environments.
● Consultancy/management on creation of complex Linux-based
systems/platforms/distros/products.
● Contributor in several FLOSS communities (KDE, openSUSE…).
● Worked at known FLOSS orgs. like Linaro, SUSE, ASOLIF,
Codethink…
● Currently at Daimler Group Services Madrid for MBition.
● Metrics lover.
● More about Agustín at his site.
Problem statement
If you want to go fast, go alone. If you want to go far, go
together.- African proverb
Problem statement
OR relation
The challenge
Can we go fast by going together?
The challenge. Can we go fast by going together?
Could an increase in number and frequency of interactions among developers
(collaboration) improve efficiency? Could an increase of developers improve
effectiveness? Is is worth the additional cost?
● For many senior managers the answer is NO… or some kind of “no”.
● YES is a counterintuitive answer so it needs to be proven.
If yes then….
If you want to go far, together is faster.
(still would be an african proverb since I am african).
A solution
AND relation
A solution… explained
1.- In addition to collaboration and community health
metrics I recommend to consider product delivery
process performance ones.
2.- Correlate both groups of metrics.
3.- Focus on decisions and actions that creates a positive
correlation between them.
A solution… explained
4.- Create a reporting strategy to developers and
managers based on such positive correlation.
5.- Use such strategy to turn around your InnerSource
story: it is about creating positive business impact at scale
through open collaboration.
Let’s focus on the correlation between those two group of
metrics: collaboration/community health metrics and
delivery performance ones...
Collaboration and community health metrics
Check the CHAOSS project: metrics.
Software delivery process performance metrics
Software delivery process performance metrics
● Both metrics characterize the performance of a system that
process a flow of elements.
● Widely used: networking, lean manufacturing...
● Can be applied at system level (globally) and locally.
● Simple units. Easy to measure.
● They can be positively correlated when applying Cont. Delivery
principles and practices.
Example: correlation between both group of metrics
Goal: to show what do I mean by positive correlation between both
groups of metrics.
Metrics:
● Collaboration/community health metric: new contributors (just released!).
● Delivery performance metrics: throughput and stability.
Assumptions:
● Ideal system. This is about illustrating the idea, not about accuracy.
● Effort and budget have been applied to increase the number of new
contributors by 20%. The goal was achieved. But...
● ….Change Failure Rate grows so Stability decreases.
Software delivery process performance metrics
Example: correlation between both group of metrics
Example: correlation between both group of metrics
Software delivery process performance metrics
Summary
1. Collaboration within a commercial organization matters more if it
has a positive (measurable) business impact.
2. To take and evaluate decisions combine
collaboration/community health metrics with throughput and
stability.
3. Prioritise those decisions/actions that produce a positive
correlation between these two groups of metrics.
4. Report, specially to managers, based on such positive correlation.
5. Adapt your Inner Source story: business impact through
collaboration.
Additional business related metrics that can be used in
combination with throughput and stability.
(requested through slack)
Business metrics: costs
Business metrics: software delivery process (simple)
Thanks to ISC Fall 2020 for this opportunity.
Thank you very much for attending to this session.
Agustín Benito Bethencourt
You can go fast by going
together: software delivery
process performance metrics

You can go fast by going together: software delivery process performance metrics

  • 1.
    Agustín Benito Bethencourt Youcan go fast by going together: software delivery process performance metrics
  • 2.
    Who am I? ●Experienced manager in open/remote environments. ● Consultancy/management on creation of complex Linux-based systems/platforms/distros/products. ● Contributor in several FLOSS communities (KDE, openSUSE…). ● Worked at known FLOSS orgs. like Linaro, SUSE, ASOLIF, Codethink… ● Currently at Daimler Group Services Madrid for MBition. ● Metrics lover. ● More about Agustín at his site.
  • 3.
    Problem statement If youwant to go fast, go alone. If you want to go far, go together.- African proverb
  • 4.
  • 5.
    The challenge Can wego fast by going together?
  • 6.
    The challenge. Canwe go fast by going together? Could an increase in number and frequency of interactions among developers (collaboration) improve efficiency? Could an increase of developers improve effectiveness? Is is worth the additional cost? ● For many senior managers the answer is NO… or some kind of “no”. ● YES is a counterintuitive answer so it needs to be proven.
  • 7.
    If yes then…. Ifyou want to go far, together is faster. (still would be an african proverb since I am african).
  • 8.
  • 9.
    A solution… explained 1.-In addition to collaboration and community health metrics I recommend to consider product delivery process performance ones. 2.- Correlate both groups of metrics. 3.- Focus on decisions and actions that creates a positive correlation between them.
  • 10.
    A solution… explained 4.-Create a reporting strategy to developers and managers based on such positive correlation. 5.- Use such strategy to turn around your InnerSource story: it is about creating positive business impact at scale through open collaboration.
  • 11.
    Let’s focus onthe correlation between those two group of metrics: collaboration/community health metrics and delivery performance ones...
  • 12.
    Collaboration and communityhealth metrics Check the CHAOSS project: metrics.
  • 13.
    Software delivery processperformance metrics
  • 14.
    Software delivery processperformance metrics ● Both metrics characterize the performance of a system that process a flow of elements. ● Widely used: networking, lean manufacturing... ● Can be applied at system level (globally) and locally. ● Simple units. Easy to measure. ● They can be positively correlated when applying Cont. Delivery principles and practices.
  • 15.
    Example: correlation betweenboth group of metrics Goal: to show what do I mean by positive correlation between both groups of metrics. Metrics: ● Collaboration/community health metric: new contributors (just released!). ● Delivery performance metrics: throughput and stability. Assumptions: ● Ideal system. This is about illustrating the idea, not about accuracy. ● Effort and budget have been applied to increase the number of new contributors by 20%. The goal was achieved. But... ● ….Change Failure Rate grows so Stability decreases.
  • 16.
    Software delivery processperformance metrics
  • 17.
    Example: correlation betweenboth group of metrics
  • 18.
    Example: correlation betweenboth group of metrics
  • 19.
    Software delivery processperformance metrics
  • 20.
    Summary 1. Collaboration withina commercial organization matters more if it has a positive (measurable) business impact. 2. To take and evaluate decisions combine collaboration/community health metrics with throughput and stability. 3. Prioritise those decisions/actions that produce a positive correlation between these two groups of metrics. 4. Report, specially to managers, based on such positive correlation. 5. Adapt your Inner Source story: business impact through collaboration.
  • 21.
    Additional business relatedmetrics that can be used in combination with throughput and stability. (requested through slack)
  • 22.
  • 23.
    Business metrics: softwaredelivery process (simple)
  • 24.
    Thanks to ISCFall 2020 for this opportunity. Thank you very much for attending to this session.
  • 25.
    Agustín Benito Bethencourt Youcan go fast by going together: software delivery process performance metrics