Twenty Years a Scrum Master

It was twenty years ago this month I finished my first assignment as a Scrum Master at Barclays Bank Head Office in London.

Of course, Scrum hadn’t been invented then, and my role wasn’t actually called “Scrum Master”. But anyone looking back at what I was doing, day-to-day, would probably recognise much of it as Scrum Mastering. Facilitation, identification and removal of impediments, a buffer between the team and distracting influences, guidance in a common approach, and so on.

Actually, it was more than just a Scrum Master role, because as well as delivering some key NeXTStep projects for Barclays, we were also inventing our own approach to software development, involving self-organisation, inspect-and-adapt, short time-boxed interactions, etc., which we came to call “Jerid” (and which over the years evolved into “Javelin“).

The label “Agile”, along with the Agile Manifesto, did not emerge until the Snowbird gathering in 2001, some seven years after we started our journey. So at the outset, we chose to describe what we did in the terms then prevailing, including, variously, RAD (rapid application development), RID (rapid iterative development) and JAD (joint application development).

Labels aside, our approach was a fusion of my own experiences of software development up that point, together with:

• Tom Gilb’s Evo method, in particular, quantification cf Principles Of Software Engineering Management.
• De Marco and Lister’s Peopleware.
• Ed Yourdon cf Decline and Fall of the American Programmer.
• A gaggle of works on software risk and risk management.
• Ideas from the quality movement (TQM, PDCA, et al).
• A sprinkling of stuff from software metrics.

Since those early days, I’ve played the role of Scrum Master or Agile Coach on some dozen occasions, interspersed with a number of other, generally more organisation-wide roles. Notably, the experiences with Jerid/Javelin at Barclays and other clients in the mid-nineties led to me joining Sun Microsystems’ UK Java Centre, and thence to starting the first 100% Agile software house in Europe (Familiar Limited, 1997).

Looking Back

Looking back today, at the things I’ve learned and how the Agile approach has emerged and grown in credibility and awareness over the past twenty years, some highlights stand out:

• We started from much the same position as the later Snowbirders: As practitioners disillusioned and frustrated with the “Big IT” approach to software development projects, often called “Waterfall” but more honestly described as “chaos, ineptitude, arse-covering and panic”. Practitioners wanting to do a good job, with the belief that we knew better how to go about developing software.
• Pretty soon, after a few projects and clients, it became obvious that being in charge of our own work and approach, at the team level, was only shifting the burden. I realised that most key impediments to teams’ progress and productivity lay outside the control of the team, in the wider organisation.
• Aside from Agile – as it came to be known – being a “symptomatic solution”, it has also failed to address the key issues of deliberately developing the “right things” and of explicitly managing the risks inherent in software development.
• Management can be very supportive of teams working on critical projects, even to the extent of allowing much leeway in adopting new approaches, and circumventing prevailing rules and mores.
• Management can also make bat-shit crazy decisions guaranteed to make delivering anything near to impossible.
• Businesses generally have a host of pressing matters, and only rarely is software development and its effectiveness anywhere visible on the business radar.
• The Agile (Snowbird) approach was never “designed”, and in particular never designed for effective adoption by real, fallible human beings in less than ideal conditions.
• Unlike other “agile” approaches, Javelin started from the position that “software development is risky, and success requires we manage those risks”.

“Risk management is project management for grown-ups.”

~ Tom DeMarco and Timothy Lister

From Processes To People

Over these twenty years, I’ve been seeking an answer to the question “Why is software development so badly handled, so ineffective, so broken, almost everywhere we look?” I’ve tried more or less formal project management, risk management, heavyweight process (CMMI), lightweight process (Agile, Javelin), management, leadership – all kinds of approaches and good-on-paper solutions. It’s taken me this long, and a circuitous journey down many cul-de-sacs, to arrive at my present understanding.

And what is my present understanding? That that only thing that really matters is the people involved in the endeavour. They don’t have to be particulary smart or experienced, but they do have to be curious and interested. And above all, motivated. All else pales into insignificance.

After twenty years, I still enjoy helping people develop software, and make a pretty good Scrum Master. Not that I’d suggest you ever hire a by-the-book Scrum Master – the role as defined has just too many built-in dysfunctions to make it useful.

– Bob