How to Create an Internal Innersource Community

Jono BaconIn recent years, we have seen more and more interest in a variance of open source known as innersource. Put simply, innersource is taking the principles of open source and bringing them inside the walls of an organization. As such, you build collaboration and community that may look and taste like open source, but in which all code and community is private within the walls of the organization.

As a community strategy and leadership consultant, I work with many companies to help build their innersource communities. As such, I thought it could be fun to share some of the most important principles that map to most of my clients and beyond. This could be a helpful primer if you are considering exploring innersource inside your organization.

Culture then code

For most companies, their innersource journey starts out life as a pilot project. Typically the project will focus on a few specific teams who may be more receptive to the experiment. As such, innersource is almost always a new workflow and methodology being brought into an existing culture. Making this cultural adjustment is key to being successful and the biggest challenge.

There is an understandable misconception that the key to doing innersource well is to focus on building the open source software development lifecycle in your company. That is, open code repositories and communication channels, encouraging forking, code review, and continuous integration/deployment. This is definitely an essential part of the innersource approach, but think of these pieces as lego bricks. The real trick is building an environment where people have the permission and incentive to build incredible things with those bricks.

As such, doing innersource well is more about building a culture, environment, and set of incentives that encourage and reward the behavior we often associate with open source. Building that culture from scratch is much easier, but adapting to this culture, particularly for traditional organizations is where most of the work lies.

Cultural change can't be dictated, it is the amalgamation of individual actions that build social conventions that the wider group are influenced by. To do this well you need to look at the problem from the vantage-point of your staff. How can you make their work easier, more efficient, more rewarding, and more collaborative?

When you understand the existing cultural pain points and your staff associate the new innersource culture as a way to relieve those issues, the adaptation goes much more smoothly.

Methodology

Given that innersource is largely a cultural adjustment that incorporates some proven open source workflow methodologies, it begs an interesting question: How do you manage the rollout of an innersource program?

I have seen good and bad ways in which organizations do this. Some take a top-down approach and announce to their staff that things are going to be different and teams and staff need to fall in line given a specific innersource schedule. Other companies take a more bottom-up approach where a dedicated innersource team informally tries to get people on board with the innersource program.

I wouldn't recommend either approach explicitly, but instead a combination. For any cultural change you are going to need a top-down approach in emphasizing and encouraging a new way of working from your executive team and company leaders. Rather than dictating rules, these leaders should instead be supporting an environment where your staff can help shape the workings and implementation of your innersource program in a more bottom-up way.

Let's be honest, everyone hates these kind of cultural adjustments. We have all lived through company executives bringing in new ways of working—agile, kanban, pair-programming, or whatever else. Often these new ways of working are pushed onto staff and it sets the dynamic off on the wrong foot: Instead of encouraging your staff to shape the culture you are instead mandating it to them.

The approach I recommend here is to put in place a regular cadence that iterates your innersource strategy.

For example, every six months a new cycle begins. Prior to the end of the previous cycle the leadership team will survey staff to get their perspectives on how the innersource program is working, review their feedback, and set out core goals for the new cycle. Staff will then have structured ways in which they can play a role in helping to shape how these goals can be accomplished. For example, this could be individual teams/groups that focus on communication, peer review, QA, community growth and engagement, and more. Throughout the entire process a core innersource team will facilitate these conversations, mentor and guide the leadership team, support individual staff in making worthwhile contributions, and keep the train moving forward.

This is how it works in open source. If you hack on a project you don't just get to submit code, but you can also play a role in the operational dynamics of the project too. It is important you bring this sense of influence into your innersource program and your staff—the most rewarding companies are the ones where the staff feel they can influence the culture in a positive way.

Asynchronous and remote working

One of the most interesting challenges with innersource is that it depends on asynchronous collaboration extensively. That is, you can collaborate digitally without requiring participants to be in the same timezone or location.

As an example, some companies require all staff members to work from the same office and much of the business that gets done takes place in in-person meetings in conference rooms, on conference calls, and via email. This can make it difficult for the company to grow and hire remote workers or for staff to be productive when they are on the road at conferences or at home.

A core component of what makes open source work well is that participants can work asynchronously. All communication, development (code submission, issue management, and code review), QA, and release management can often be performed entirely online and from any timezone. The asynchronous with semi-regular in-person sprints/meetings combo is a hugely productive and efficient methodology.

This can be a difficult transition for companies. For example, some of my clients will have traditionally had in-person meetings to plan projects, perform code review over a board-room table, and don't have the infrastructure and communication channels to operate asynchronously. To do innersource well, it is important to put in place a plan to work as asynchronously as possible, while blending the benefits of in-person communication at the office, while also supporting digital collaboration.

The side benefit of doing this is that you build an environment that can support remote working. As anyone who works in technology will likely agree, hiring good people is hard, and a blocker can often be a relocation to your city. Thus, your investment in innersource will also make it easier to not just hire people remotely (anyone can do that), but importantly to have an environment where remote workers can be successful.

Peer review and workflow

For companies making the adjustment to innersource, one of the most interesting "sticking points" is the peer review process.

For those of you familiar with open source, there are two important principles that take place in everyday collaboration. Firstly, all contributions are reviewed by other developers (both new features and bug fixes), and secondly, this peer review takes place out in the open for other people to see. This open peer review element can be a tough pill to swallow for people new to open source. If your engineers have either not conducted code review or it took place privately, it can be socially awkward and unsettling to move over to a more open review process.

This adjustment is something that needs to be carefully managed. A large part of this is building a culture in which critique is something that should be favored not avoided, that we celebrate failure, and we cherish our peers helping us to be better at what we do. Again, this is about framing these adjustments to the benefit of your staff so that while it may be awkward at first, they will feel the ultimate benefits soon after.

As you can probably see, a core chunk of building communities (whether public communities or innersource communities in companies) is understanding the psychology of your community members and converting those psychological patterns into workflow that helps you encourage the behavior you want to see.

As an example, there are two behavioral economics principles that play a key role with peer review and workflow. The first is the Ikea Effect. This is where if you and I were to put together the exact same Ikea table (or build something else), we will each think our respective table is somehow better or more valuable. Thus, we put more value into the things we make, often overstated value. Secondly, there is the principle of autonomy, which is essentially that choice is critically important to people. If we don't feel we have control of our destiny, that we can make choices, we feel boxed in and restricted.

Just these two principles have an important impact on workflow and peer review. In terms of the Ikea effect we should expect that most people's pull requests and fixes will likely be seen as very valuable to them. As such, we need to use the peer review process to objectively define the value of the contribution in an independent and unemotional way (e.g. reviewing specific pieces of the diff, requiring at least two reviewers, encouraging specific implementation feedback, etc). With the autonomy principle, we should ensure staff can refine, customize, and hack on their toolchain and workflow as much as possible, and regularly give them an opportunity to provide feedback and input on making it better.

Reputation, incentives, and engagement

Another key element of building an innersource culture in a company is to carefully carve out how you will track great work and incentivize and encourage that kind of behavior.

This piece has three core components.

Firstly, we need to have a way of getting a quantitative representation of the quality of that person's work. This can be as involved as building a complex system for tracking individual actions and weighting their value, or as simple as observationally watching how people work. What is important here is that people should be judged on their merit, and not factors such as how much they schmooze with the bosses, or how many donuts they bring to the office.

Secondly, based on this representation of their work we need to provide different incentives and rewards to encourage the behavior we want to see.

There are two core types of rewards. Extrinsic rewards are material in nature, such as T-shirts, hoodies, gift cards, money, and more. Intrinsic rewards are of the more touchy-feely kind such as respect, recognition, and admiration. Both are important and it is important to get the right balance of these rewards. Based on the behavior you want to encourage, I recommend putting together incentives that inspire action and rewards that validate those actions.

Finally, it can be helpful to sub-divide your staff into different groups based on their work and engage them in different ways. For example, people who are new will benefit from mentoring, support, and wider guidance. On the other hand, the most active and accomplished staff members can be used as a tremendous source of insight, guidance, and them enjoying playing a role in helping to shape the company culture further.

So, there are a few starting points for broader brushstrokes that typically need to be made when painting an innersource picture inside your organization. As usual, let me know your thoughts in the comments and feel free to reach out to me with questions!
 

How to create an internal innersource community was authored by Jono Bacon and published in Opensource.com. It is being republished by Open Health News under the terms of the Creative Commons Attribution-ShareAlike 4.0 International License (CC BY-SA 4.0). The original copy of the article can be found here.