Here’s a typical, if sad, example from a real software testing organization: The employer hires four low-wage testers with no formal training, no passion, and no real interest in learning testing. They are there to do the testing, and suddenly, you are one of the only adults left around to lead them. It is almost certain that things you respected in some colleagues—their ability to ask questions and challenge methods—were not appreciated, and that was the reason you see these four new, pliant testers sitting next to you. What do you do?

I’m going to look at this situation through the lens of the Gervais Principle, but you don’t need to read that novel-length document to understand this. For now, let’s look at four methods of change and how to decide which one is right for you.

The Methods of Change

Now, let me add my own set of ideas into the mix. When confronted with a culture problem, there are four things you can do:

• Leave. Run away. Find a new job. Quit. Become a consultant.
• Attempt to change the system from within via persuasion, appeals to logic, and, frankly, a few less honorable methods. The goal here is to get promoted enough to be in charge, creating the change yourself.
• Break the system using things like whistle-blowing or going to the owner(s)/stockholders.
• Shut down mentally and become a cog in the machine.

So you notice a system that has one or more cultural flaws in it—flaws deeper than could be fixed with just a minor process improvement or one simple discussion with your boss. Often your boss is not high enough up to do anything, either. Based on the Gervais model, your boss is likely a highly motivated individual who at one point worked really hard and is now in middle management. They may view the things it takes to get into upper management as distasteful, or not know how to get there, or simply be too junior. They may be passionate about their work, but they are likely to view their contribution as a specialty (testing, quality management, etc.) and not see the bigger picture of the business.

I have found that when you talk to your boss about the systemic problem, the response will fall into one of three categories. Your manager might know but not care (or care but has given up); she might not know and not care; or she might know and is attempting to create change very slowly. The third category is the least likely although the most often stated by a boss. The reason a manager will almost never discover an issue of culture from your reporting and then care about the problem is because an issue of culture takes a long time to come about. If they didn't notice how the culture works, either they don't care about the job or they can't understand your point of view because they accept the culture. So talking to your manager is unlikely to create change.

Understanding the Organization

Start by keeping your eyes open. All cultures have items that cannot easily change because of their identities. Can you name those? Look at the individuals running the organization. Attempt to pick up on their thought processes. Equally important, look at those out of power—these people might be ignored, bullied, or removed from their job posts. These sorts of power dynamics will tell you about the culture of the company. Without these sorts of fundamental understandings, you can’t even perceive what pieces are cultural and what are current business operations. They are fundamentally different things, as your business operations can change relatively quickly while culture is much more difficult to change.

The next most useful tool is asking strategic questions. You have to know that you don’t get unlimited questions in any given meeting. You get maybe a maximum of three nonclarifying questions in most meetings. There are exceptions, and certainly some meetings are more open to questions, but most cultures I’ve experienced don’t expect too many questions. In QA we are expected to ask more questions than most, so we get a little bit of leeway, but there is still a limit to the number of questions people will put up with before you lose credibility. I won’t recommend particular questions, but I do suggest you understand what information the question is attempting to get and who your audience is. If possible, organize your question in such a way as to maximize the information. I once asked my former director of development about how he expected the manual and automation testers to work together. He gave a nonanswer that was off topic. I changed my wording to see if he misunderstood. He had not. It was clear he either didn’t understand QA or had not considered the question and didn’t want to lose face. (Later, it was revealed that he didn’t understand QA.) This explained in part why development didn’t work well with QA, as there was no encouragement from the top. It was a cultural flaw. (Note: People are afraid to lose face in meetings; hallway conversations can often be more revealing.)

Examples of Change

You can try to replace your manager, or go around her. Perhaps you notice how lax security is, but you're in testing. I'm not just talking about a bug in security, but the general philosophy surrounding the company. Telling a test manager about lax security seems unlikely to help. It is a cultural thing. You can file bugs. You do. They are ignored and deemed low priority. "I mean, we don't store credit card data in our database," you hear from the product owner. You mention it to the VP of engineering and she asks, “How would fixing that impact our dates?” You say no more than a week or two, and she says, “No, but maybe after the holidays” in a way that tells you, “Maybe after next year’s holidays ….”

Now, maybe if you were in ops, you could fix this problem using firewalls and the like. One option is to attempt to move into ops or become the product owner to get people to change. This is costly and takes a lot of time, but it is an option. If you are the person in charge, maybe you can effect change. Maybe you just keep asking ops over and over again, and they change things. Maybe you use your friendship with a developer who is friends with the VP to push the change.

Now, these sorts of tactics are just that—tactics. They don't fix the long-term culture that the team learned, but if you can get the top to care, maybe the culture will slowly shift. You have your friend and the VP on your side; you might have a fighting chance. Persuading powerful people or becoming the person in power can change a culture, but it’s a long-term fight.

Using that same security concern, there is a drastically different method for changing culture. You try changing minds and nothing sticks. You talk with all your friends and people in power until you are blue in the face. Nothing changes. No one cares. You get angry. You call the company lawyer. The company lawyer knows liability law, and he begins to get scared. He calls the C-levels together and tells them what has been going on. However, you are not a hero. You're hated. You went outside the chain of command. In an alternative world you could do worse yet: Instead of calling the company lawyer, you could call your own and sue the company as a customer. You will definitely effect change, but then you'll be shunned. Your name might be mud in the town you live in. You might lose your job or even be sued for causing problems at your company. You will not be a hero, even in your coworkers’ eyes. Whistleblowers are often considered traitors who betrayed the group, no matter how good their intentions.

The last and final act you can perform to effect change is to simply quit. Sometimes the company will ask why, and you can try one last time to fix the culture. It might get better or it might not, but it won't matter to you. This is best if you already have another job. It can be a sad, terrible choice, but it might also be the best choice. You have your honor and your standards, and you can sleep at night.

The last option—a nonoption, in my view—is to do nothing and hope something changes. Maybe it will. But that is, at best, random chance. Random chance is not a method of change, but I did want to include it for completeness’s sake.