$30 off During Our Annual Pro Sale. View Details »

Promise theory: from configuration management to team leadership

Promise theory: from configuration management to team leadership

How I applied Promise Theory to managing a loose team of people to most of whom I don’t have a line of command. PT was born in 2004 as Mark Burgess was looking for a model to describe CFEngine 3 but it soon proved able to model any complex system, and I had used CFEngine a lot when I became the Head of IT in Telenor Digital. With a lot of both operational and managerial tasks and an initial team size of one and then two, the challenge was big and borrowing help from other functions in the company was a necessity. A traditional management style was not going to work, so I decided to shape my own management style by using PT. This is the report of the first year.

Marco Marongiu

February 05, 2018

More Decks by Marco Marongiu

Other Decks in Technology


    ...enjoy the ride!
    From: Configuration Management
    To: Team Leadership
    Class Ticket type Adult Child
    Start date Route Conductor
    February 5th, 2018 Config Management Camp Marco Marongiu

    View Slide

  2. Agenda

    Me and Telenor Digital

    A primer in Promise Theory

    Putting it all together...

    What’s here, what’s next, what we don’t know yet

    View Slide

  3. Me and Telenor Digital

    View Slide

  4. Who I am

    Sysadmin for several organisations; in Opera Software (Oslo,
    Norway) 2010-2016, did lot of configuration management
    there (Puppet first, then CFEngine);

    Leader of the 24/7 squad of the IT Services Management team
    in Tiscali (Italian ISP), June 2005 June 2006 (5 reports);

    Head of the Systems group in Sardegna IT (agency of the
    Regional Government of Sardinia),
    November 2008 March 2009 (8 reports).

    View Slide

  5. Meet Telenor Digital

    Part of the Telenor group, the biggest telephony operator in
    Norway with subsidiaries in Europe and Asia

    Immersed in a “traditional” environment, yet modern
    – Agile methodologies
    – Heavily cloud-based
    – Freedom of experimentation, freedom to fail
    – Focus on innovation
    – DevOps culture

    View Slide

  6. Telenor Digital’s IT before November 2016

    Routine work distributed across the company (e.g. on-/off-
    boarding of employees, order of equipment...)

    Bigger IT tasks (e.g. office network, company-wide systems)
    looked after by skilled engineers, but on a case-by-case basis

    IT was no-one’s job and no-one’s priority

    Technical debt accumulated year after year and the
    organisation was now feeling the pressure of it

    It was time for TD to get a dedicated IT team

    View Slide

  7. Telenor Digital in November 2016

    I joined the company on November 1st, 2016

    Day-to-day tasks, the pile of legacy on my lap, different
    expectations coming from all directions

    Lots of highly-important things, difficult to understand what
    should be on top and what should wait

    One-man team until March 2017, a dozen helpers but no line of
    command to anyone.

    View Slide

  8. View Slide

  9. Rethink leadership

    Traditional, “line-of-command-based” leadership cannot work

    To succeed, I needed to exploit my helpers’ autonomy
    – Notice how my helpers’ autonomy is exactly what makes the
    traditional leadership style fail…

    What leadership style is suitable to coordinate a group of
    people who don’t report to you for the most part?

    View Slide

  10. The three traits of my “leadership style”

    You start by trusting each person in the team and being on
    their side

    Don’t be a jerk

    Don’t impose unless absolutely necessary

    View Slide

  11. A primer in
    Promise Theory

    View Slide

  12. Key concepts

    An agent is anything that can promise an outcome or a
    capability (explicitly or implicitly)

    View Slide

  13. Things to notice about agents

    There will be agents making promises and agents using promises

    They make promises voluntarily, and they try keep promises voluntarily
    (agents are autonomous)
    – Autonomy and voluntary cooperation are so key to promise theory that you must
    always keep them in mind when using it;

    There may be boundaries, limits or conditions to where the promise holds;

    If an agent is unwilling to cooperate you can try to incentivize its will to
    cooperate somehow, as long as promising what you want is in its capabilities;

    ...or you can try to impose on an agent, YMMV.

    View Slide

  14. Assessing promises

    Who decides if a promise was kept or not? The agent making the
    promise or the agent using the promise?

    There is an entire chapter in “Thinking in Promises” about assessing promises and it’s not
    only about assessing the outcome of an agent; however, in this particular talk we have to
    focus on assessing the outcome only.
    – Hint: you either have children or you were a child to someone: when parents ask
    a child to tidy his bedroom, who does the assessment of the outcome?

    A successful outcome requires that the outcome is well specified in
    advance and is clear to both the agents making the promise and the
    agent using that promise.

    View Slide

  15. Superagents

    View Slide

  16. Superagents

    View Slide

  17. Putting it all together...

    View Slide

  18. View Slide

  19. Thinking in promises...

    Helper == agent offering/promising me a service
    – Help with video conference systems: orders, accounts, contracts…
    – Help with hardware orders and supplies
    – Help with managing the office network in a remote office
    – Help with managing Jira projects and other requests regarding Atlassian tools

    Agents may not be able to keep their promises
    – e.g. they may receive conflicting orders from their managers, and those take a higher priority

    Me == agent using helpers’ promises, super-agent offering IT services to the
    – Promise IT services to employees and manager, based on the promises used
    – Represent my network of promises to others,or disclose as much details as needed, depending
    on the agent requesting information

    View Slide

  20. Advantages

    I was not imposing on people;

    I knew clearly what I could expect and by when, and what I
    could not expect
    – I was myself able to set realistic expectations for the IT team
    towards others
    – I was able to decide what tasks were better to take on and what were
    simply impossible to achieve at the current state.

    View Slide

  21. Expanding the horizon of PBL

    Promise theory can be used to model processes
    – Design processes so that involved “agents” can promise to follow

    Promise-based leadership can be used with direct reports
    – Remember that impositions don’t guarantee the outcome, so there is
    no real advantage in using impositions instead of promises

    Time to see some concrete examples

    View Slide

  22. What’s here
    What’s next
    What we don’t know yet

    View Slide

  23. Weekly planning

    How it works: what tasks can we promise for this week?

    What’s new with this process: easier to set realistic goals and
    realise what’s a stretch instead;

    Does it work? Still influenced by our traditional way of
    scheduling, but it’s OK.

    Open problem: does it scale to larger teams?

    View Slide

  24. Establishing processes

    How it works: understand what people can promise, then design
    processes around that;

    What’s new with this process: minimum reliance on imposition,
    effective processes, almost no need for “police”;

    Does it work? It does.

    Open problem: one year is not enough to tell if it’s sustainable.

    View Slide

  25. Establishing OKR for a team

    How it works: incomplete objectives from previous quarter and overall
    goals for the year are evaluated; we pick from the most important things
    what we can reasonably promise for the quarter;

    What’s new with this process: easier to set realistic goals and realise
    what’s a stretch instead;

    Does it work? Didn’t work well initially because of high priority given to
    unplanned work (helpdesk); working better now with more focus shifted to
    long-term improvements (OKR 2017H2 == 66%)

    Open question: not tested at scale.
    If you don’t know OKR: https://www.youtube.com/watch?v=mJB83EZtAjc

    View Slide

  26. promise
    PB-OKR for a company: seems familiar?
    can’t keep - reformulate
    can’t keep - reformulate
    can’t accept - reformulate
    can’t accept - reformulate
    can’t keep - reformulate
    can’t keep - reformulate
    can’t keep - reformulate
    can’t keep - reformulate

    View Slide

  27. IPv6 MTU path discovery

    View Slide

  28. Summing up

    A promise-based leadership style contributes to better relationships with
    the team, helps setting more realistic goals by making it clear which
    objectives are doable, which are a stretch and which are impossible; it also
    contributes to set the right expectations on the people “using” your team;

    It requires a leadership attitude that doesn’t rely exclusively on
    imposition, empowers team members and helps them develop so that they
    can promise more;

    It requires a company culture where it is OK to say “no” and “I can’t do
    that” and is open to negotiation.

    View Slide

  29. Have a nice trip!
    Marco Marongiu
    Email: [email protected]
    Twitter: @brontolinux
    Web: http://syslog.me/
    LinkedIn: https://www.linkedin.com/in/marcomarongiu

    View Slide

  30. Inscription

    This work is dedicated to:
    – Ann Karin Tonstad
    – Anna Kennedy
    – Anne Omholt
    – Christoffer Viken
    – Eva Reisersen
    – Eyvind Bernhardsen
    – Gro Kvamme Johnsen
    – Holger Ihrig
    – Jarrod Harbrucker
    – John Corrigan
    – Kristian Barek
    – Miha Pirc
    – Ole Kristian Brattli
    – Tina Stenberg
    – Tina Tan
    – Volker Hilsheimer
    – ...and all those who helped me and
    the IT team since I joined Telenor

    View Slide