The Power to Lead
How do you systematize knowledge work?
Peter Drucker noted that so much of what we call management consists in making it difficult for people to work. No wonder companies grow cultures with special talents for resisting change.
The great conceit of agile has always been reliant on heroes to overcome corporate cultures which impede effective knowledge work.
How do you systematize knowledge work?
I can hear many Agilistas out there protesting: “No, no, Curtis! Agile teaches that relying on herioism is an anti-pattern. Scrum establishes new ways of working, new systems, that overcome limitations.”
Except that it doesn’t. Not really.
What Scrum does, if implemented effectively, is isolate a development team from the traditional cost & schedule pressures of the firm, allowing the team to become, de facto, “self-managed.”
Such a team will produce work as quickly as they can, assuming they are highly self-motivated (this is not always a given, of course). Initially, they will be less productive until they get used to the new ways of working, and then often will ‘gel’ and become more effective. If left to their own devices, however, most teams will experience “entropy”. Productivity tends to devolve to something no better off than before.
So-called “scaled agile” frameworks really just attempt to apply this pattern of isolation recursively throughout the organization. While this makes for a great training course and sales presentation, scaled agile frameworks typically fail to address 3 key components: existing hierarchies, budget ownership, and management incentives.
Existing Hierarchies & Communications Structures
Budget Ownership [Suzy works for me 20% of the time, for you 40% of the time, for Ellen 20% of the time; and is still expected to do all the ‘good corporate citizen’ things- mandatory and voluntary training and self-development, compliance, etc.]
Management Incentives: How are managers measured and rewarded? Is it on the basis of speed of delivery? How about cost-cutting (euphemistically referred to as “effective resource management”)? The increasingly popular employee engagement score (usually a variation of NPS)? Or perhaps some combination of all the above?
Knowledge Robots vs. Human Creatvity
Why can’t Johnny iterate?
Compared to his teammate Kobe Bryant, who had a lifetime free-throw average of almost 90%, the 7-foot tall Shaquille O’Neal was an appealing target for opposing teams to foul if they were down by just a few points with only seconds on the clock. NBA players even had a name for this tactic: “Hack-a-Shaq.” That’s because O’Neal, a Hall of Fame Basketball player, was a notoriously poor free-throw shooter, with a lifetime average of just about 50%. This, despite the fact that the ball was always exactly 22 ounces, the rim was always 10 feet high, and the free throw line was always XX feet from the basket. So, why was Shaq so much worse (relatively) than Kobe?
Was it the opposing crowd noise? Was it the closeness of the score? Was it whether Shaq was coming off a loss or a win, or how he had played earlier in the game? Was it the meal he had at lunch, or the weather outside? All of these are potential causes of the statistical noise in Shaw’s free-throw history; that is, his erratic results at the line.
If basketball were like auto manufacturing, there is no doubt that Ford or Toyota could design a “free-throw robot” that would sink the baskets every time. But who would watch? Shaq’s nerves were impossible to predict each night, but they were also part of the reason why people tuned in to watch- to see if he could overcome that adversity.
Knowledge workers suffer from similar erraticism in their predictions of completion. But unlike a sports match, managers are not amused when targets are missed. Couldn’t we use robots to build software reliably? In some cases, yes… sort of. Part of the reason for the movement to “the cloud” and other Software as a Service platforms is precisely because they are more discrete and robotic in their ability to enable certain features and functions. Unfortunately for the managers, humans’ ability, especially those in the Marketing Department, to conjure up new feature lists far exceeds to ability of platform developers to anticipate and pre-deliver these features, much less customize them to the extent often required in global enterprises. Otherwise, we would simply install “self-managing knowledge robots,” and encourage human developers to seek new opportunities outside the company.
What about Bob?
Creating self-managed teams also begs the question of “What do we do with the team’s manager?” Most agilistas will tell you the answer is simple: turn them into an agile coach, and redeploy them elsewhere in the organization to convert the unwashed masses to the agile religion. Or, if not, then encourage them to seek other opportunities outside the company.
Culture vs. System
True culture change does mean to change what we do every day. But that does not necessarily mean that effective systems for knowledge work will be created. What’s the difference, and how do we create effective systems?
Let's agree to define productivity in terms of throughput. We can debate the meaning of productivity in terms of additional measurements of the business value of delivered work, but as Eliyahu Goldratt pointed out in his critique of the Balanced Scorecard, there is a virtue in simplicity. Throughput doesn’t answer all our questions about business value, but it is a sufficient metric for the context of evaluating the relationship of practices with productivity.