I consult, write, and speak on running better technology businesses (tech firms and IT captives) and the things that make it possible: good governance behaviors (activist investing in IT), what matters most (results, not effort), how we organize (restructure from the technologically abstract to the business concrete), how we execute and manage (replacing industrial with professional), how we plan (debunking the myth of control), and how we pay the bills (capital-intensive financing and budgeting in an agile world). I am increasingly interested in robustness over optimization.

I work for ThoughtWorks, the global leader in software delivery and consulting.

Monday, February 29, 2016

How an Operational Gap Becomes a Generation Gap Becomes a Valuation Gap

A decade or so ago, when an IT organization (captive or company) hit rock bottom - bloated payroll, lost confidence and ruptured trust resulting from rampant defects, rocky deployments, functional mis-fits, and long delivery cycles - it would give anything a try, even that crazy Agile stuff. It didn't matter if it was incumbent management with their backs against the wall, or new management brought in to clean house, desperate times called for desperate measures. To people looking to shake things up, Agile made intuitive sense and a lot of it was appealing, even if its proponents had a bit too much evangelical zeal and some of it sounded a little strange. So they'd bite the bullet and do it, selectively anyway: a build server was easy to provision; releasing software for testing every couple of weeks was done easily enough, and getting everybody in the same physical workspace (or at least in close proximity to one another) could be done with a little arm-twisting; of course, developers were instructed to only pair program "opportunistically", and automating QA was a fight nobody wanted to fight, so they'd sit with the team and test incrementally but go on testing the way they always had. Still, even with compromises, there was a good story to tell, usually to do with much higher quality and eliminating rework, and that made Agile A Good Thing for all parties concerned.

Fast forward to today, and we see that Agile is much more ambitious. A few short years ago we were content to have an automated build execute every few minutes; today we want every check-in to trigger a build that is promoted through progressively more complex tests in virtual environments managed, instantiated and torn down by scripts. We used to be content releasing for user testing every other week, and to production every couple of months; we now aspire to release to production many times a day. We used to want Master Story Lists to guide incremental development; today we want to iteratively experiment through code and have the feedback inform requirements definition and prioritization. We used to be satisfied with better delivery of internally-facing software; today we want to evolve software products that are used by people across our ecosystem, from interested parties to customers to investors to employees. Today, Agile wants to push against everything that creates an artificial or temporal constraint, be it organization, management, accounting policy, or even capital structure.

Although Agile has evolved, the entire tech world hasn't moved with it. In fact, some of it hasn't moved at all: it's still common to see non-Agile organizations that do big up-front design; work in functional and skill silos; have manual builds, deployments and testing; and make "big bang" releases. And, it's still common for them to face a "rock bottom" moment where they conclude maybe it's finally time to look into this Agile stuff.

As hard as it was a decade ago to inject Agile into a non-Agile organization, it's much harder today for a non-Agile organization to complete a transformation. This seems counterintuitive: since the non-Agile to Agile path is so well trod, it should be much easier than it was in those pioneering days of yore. Although there's never been more tools, frameworks, languages, books, blogs, and countless other resources available to the individual practitioner aspiring to work differently, organizational change tends not to be self-directed. The challenge isn't taking an organization through the same well-established game plan, it's finding the people - the transformational leaders - who are willing to shepherd it through its journey.

Unfortunately, re-living the same internal negotiations to reach the same compromises, solving technical and organizational problems long ago solved, only to end up with an operating model that is considerably far short of the state-of-practice today is not a destination opportunity for an experienced change leader. Even assuming, as my colleague Alan Fiddes pointed out, that the change agents brought in still have the vocabulary to carry on arguments last fought so long ago, any change agent worth their salt isn't going to reset their career clock back a decade, no matter the financial inducement.

This might simply mean that the approach to change itself is what has to change: require far less shepherding from without by expecting more self-directed change from within, brought about by setting the right goals, creating the right incentives (motivate people) and measuring the right things (what gets measured is what gets managed). Why shouldn't it be self-directed? It isn't unreasonable to expect people in a line of work as dynamic as software development to keep their skills sharp and practices current. For people leading an organization that's a little dated in how it develops software, then, the question to hold people to isn't "why aren't you doing Agile" but "we're going to deploy any time and all the time effective fiscal Q3, so how are you going to operate to be able to support that?" It's amazing what people will do when sufficiently motivated, change agents be damned.

Whether there's a more effective means of change or not, being far adrift of the state of practice points to a more severe threat to the business as a whole: a generation gap.

* * *

Three decades ago, the state of practice didn't vary that much across companies. Yes, there were people coding C over Rdb deployed in VMS on minicomputers and people coding COBOL over IMS deployed in OS/390 on mainframes, but the practice landscape wasn't especially diverse: waterfall prevailed and a lot of code was still data-crunching logic run in batch. At the time, captive IT, consulting firms, governments, new tech firms (think Microsoft in the mid-80s), and established tech stalwarts (H-P, IBM) could reasonably expect to compete for the same labor. College grads in Computer Science or Management Information Systems learned practices that reinforced the modus operandi common to all consumers of business computing.

The landscape has changed. Practices are far less homogeneous, as they've had to evolve to accommodate a diverse community of interactive users pushing for features and functionality with little tolerance for failure. The familiar combatants still slug it out for labor, but must now also compete against tech product firms untethered to any legacy practices, norms, policies or technologies. Today's grads are trained in new practices and expect their employer to practice them, too.

Companies lagging behind in their state of practice will struggle to compete for newly minted labor: why would somebody with highly marketable tech skills go to work at a place stuck in the past, when they can work in a more current - even cutting edge - environment?

This isn't just a hiring problem. A practice gap is fuel for a generation gap if it deflects young, skilled people from becoming employees. By failing to hire the next generation employee, a company develops an intrinsic inability to understand its next generation customer.

A company isn't going to reach a new generation of buyer - consumer or business - if it is tone deaf to them. A company ignorant of the next generation's motivations, desires, values and expectations has little chance of recognizing what it isn't doing to win their attention, let alone their business. Since social systems are self-reinforcing, a company is unlikely to break the deadlock of ignorance and inaction.

Failing to bridge a generation gap not only cuts a business off from growth opportunities, it sets the stage for long-term irrelevance. Investors recognize this, even when management does not. Growth changes from being a "risk" to being an "uncertainty", and when that happens a company's future1 is no longer priced at a premium, but a discount. In this way, an operational gap becomes a generation gap becomes a valuation gap.

Outdated practices are an indicator that management has it's head buried in the sand: it has a problem it can't see, that it doesn't know how to solve, that is starved for information it can't get because it has elected to disassociate itself from its source. The motivation to change how you practice shouldn't be to become more competitive today, but to still be relevant tomorrow.



1 By way of example, Yahoo net of its Alibaba holding and cash has frequently been valued at or near $0 by investors in recent years.