The war for talent. How to become a successful organization with DevOps.
Author: Peter Graef
Are you an IT manager or specialist? Or do you work on the business side and deal with digitalization and innovation? Then it is very likely that you have already heard of DevOps and understand this concept. And at least as likely you suffer from the lack of DevOps experts in your company when implementing DevOps.
In the US/UK language area people like to talk about the «War for Talent». At DevOps conferences, managers complain about their internal IT and the lack of skills. And the collective complaint is mixed with envious looks at the «Unicorns» like Google, Amazon, Netflix etc. - how did they manage to have these innovative, well-trained and highly motivated IT engineers? Where did they get them?
Adrian Cockcroft (VP of Cloud Architecture at Amazon and former CTO of Netflix) answers this question:
So the top talents are already in your company! But they are dormant talents that need to be awakened and developed into DevOps experts. DevOps is therefore also an HR topic! In Switzerland, there are already a few companies that include the HR aspects as an important workstream in their DevOps Journey or Agile Transformation. Proven Practices and process models are hardly available and if they are available they are mostly incomplete. But the ToDo's are clear:
Let's start with what just isn't working today: Our trainings.
Our trainings produce professionals who are strong in «learning/reproducing» mode. Instead of solving problems together, we focus on individual knowledge. (Project) management is understood as a planning and control function. ICT profiles (=silo roles) such as business analysts, developers, architects, testers etc. are defined in different levels (junior, senior, professional). Incidentally, success-critical roles from the IT operational organizations often do not occur... Furthermore, there is no comprehensive lifecycle analysis of (software) products. Instead, we think in terms of «Plan-Build-Run».
If you don't want or can't run away, there is the possibility to become an internal DevOps Leader in your company. Maybe you can approach the HR department in your company and promote new ways of talent development. Maybe you are even from HR and want to contribute to DevOps leading to better organizations.
Develop in-house training courses that focus on systemic thinking, lean management and continuous improvement. Optimizing the whole vs. optimizing subsystems. Provide learning organizations with feedback loops across the value chain. Use Brown-Bag Sessions, Tech Talks, Show and Tells, Ignite Talks, Demo Days, etc. and create curiosity for innovative learning practices that put «collaborative solution development» before «learning/reproducing». Pick up on trends like design thinking, co-creation, value stream mapping, etc. to stimulate cross-team collaboration.
Send your employees to many external trainings and conferences. And then demand internal sharing of what you have learned. Your employees will be doubly motivated, because a training course in Munich, Berlin, London or San Francisco is fun and passing on what they have learned to their colleagues is even more fun. Spend a lot of money on training - it's worth it! Do you know the joke of the sceptical manager who tells his colleague: «Imagine we send X to this expensive training and then he leaves». His colleague replies: «Imagine we don't send him and he stays.»
Create internal mentoring and coaching roles. These key people can have a tremendous scaling effect by inspiring and motivating other employees to work across teams. Give your best employees less professional workload and more responsibility for community building activities.
On the organizational aspects that keep top talents in the company: Do you know what makes your employees happy? Don't read on. Pause and write your assumptions separately on a piece of paper...
Hopefully, your note will contain phrases such as fairness and transparency, your own scope for decision-making, flat hierarchies, appreciation and trust, top-class colleagues, interesting and varied work, cooperation across teams, fair remuneration, no chronic overtime.
Many of these features have to do with the corporate culture. It is influenced 'from above'. More about this in a following blog post. In the DevOps context we say that leaders take responsibility for the creation of a 'blameless culture'. If you have such a culture in your company, you can see if wrong decisions with negative effects have been made. What happens then? In a DevOps organization, we say: «There are only two types of decisions: First, good decisions and second, opportunities to learn...». This is also true for managers, who talk about their own mistakes themselves and proactively, thus achieving learning effects.
Some of the organizational good luck charms mentioned above have to do with the structural and process organization in your company. New German: The Operating Model. But you do not have to perform a reorg (again) immediately. You can leave line functions as they are and still start building stable teams. Employees can come from different organizational units and still take joint responsibility for a software application or product, for example. Start with co-location, because short distances make effective communication easier. Always keep Conways Law in mind and break down hierarchical levels and organizational boundaries. At least in the pilot tests and DevOps initiatives. The next Reorg will certainly come and with it the opportunity to adapt the operating model with regard to a value stream organization.
And last but not least: If you have developed more top talents with these measures and lose less top talents...then you make it easier for the recruiters in your company to attract new top talents from outside. Do it like Adrian Cockcroft!