Being a good PM at Google

Things I've learned over 12 years building products you love

Being a good PM is hard

Nobody told me how to be a good Product Manager at Google and I made a lot of mistakes. Mistakes help you learn, but are also no fun to make. Here are eight things I've learned the hard way. Read this and learn the easy way.

Being a good PM is hard because:

• Every PM job is unique and requires different skills

• Success is extremely hard to achieve

• Most of your success depends people and variables you can't directly control

GenAI makes this even harder: the centre cannot hold. Everything is changing all at once: the product landscape, the jobs to be done, the capabilities, and the job of a PM itself.

Lastly, being a PM is hard work. You are not an ivory tower professor mandating brilliance from on high. At every stage, at every step, you need to be the most engaged, hardest working member of the team to ensure success.

As Indira Gandhi said, "There are two kinds of people, those who do the work, and those who take the credit. Belong to the first category". Do the work.

1. Shipping is all that matters

Ship something good. Everything you do is a means to this end.

If you want to:

• make users happy

• keep your team employed

• get promoted

…you need to ship and you need to ship now. As David Lieb used to say to me, "a PM is the arbiter of focus". Focus the team on shipping.

Sometimes this will mean you make mistakes and not ship the right thing. This is okay. Be incremental and unbundle launches to reduce the impact of a mistake. You can iterate your way to a good product much faster than you can think your way to a good product. Default to shipping. You learn more and it's more fun.

We pushed Gemini CLI really hard to get it out as fast as possible. The whole effort was less than 3 months from conception to launch, faster than the cadence of some teams' roadmap updates. We finished the last P0s hours before we shipped. We cut a lot of things that we all would have liked to get in, but we all felt very strongly that competitive pressure demanded we go fast, and we believed in the unique advantages Google offered. It was the right call: we saw enormous adoption and quickly surpassed all competitors in GitHub stars.

To paraphrase Von Moltke, "no product roadmap survives first contact with users." We made some mistakes in the process but those weren't the problems we thought we would have. We learned something important in shipping, and we fought different fires than we anticipated. Shipping let us see reality in a way we couldn't before.

Now I'm not a monster, don't ship garbage. Don't ship something you don't believe in. Never ship something that you are embarrassed to put your name on. But ship fast, ship early, ship often.

2. Help everyone yearn for the vast and endless sea

If you want to build a ship, don’t drum up the men to gather wood, divide the work, and give orders. Instead, teach them to yearn for the vast and endless sea.

-Antoine de Saint-Exupéry

If you learn one thing, learn that shipping is all that matters. If you learn two things, learn that inspiring people is the best way to ship. First, look to inspire. Don't start with the tactical.

People who are inspired by what they can achieve will work faster, work harder, and work with more focus than people who get a PRD thrown over the wall to implement, say, FedRamp compliance lists. Nothing against FedRamp! But you're going to get a lot more mileage out of showing the glorious future of, for example, how the Department of Education needs FedRamp compliance to be unblocked and use your product to empower millions of under-privileged students across the US. Sometimes we need to do grunge work to achieve something beautiful! Humans can do insanely difficult work over really long periods of time if they have the right motivation. Humans are amazing.

When we were designing Google Photos, David Lieb would get up in meetings and he'd walk up to the screen in the front of the room and he'd interact with the monitor like it was a phone. He had developed slide animations to mimic the exact flow he wanted to showcase and he'd use a clicker to advance the slides with his other hand. He'd scroll through the gallery to let you feel the magic; he'd tap on search and show you face clusters and the rest, he'd just navigate the app to show us all what it could be. Those were the most inspiring product updates I've ever seen. You could see the future right then, and get the feeling of what we were going to ship.

You are in a position to help everyone you meet find success. There's enough success for all of us. Help everyone to yearn for the vast and endless sea.

3. Use your product every day

I'm sorry. I know of no other way to succeed.

If you're not using your product every day you're going to miss critical issues and areas for improvement. Someone else is competing with you and they're using their product every day and they'll win. You cannot win if you're not using your product every day.

When I joined Colab, my TLM told me [team x] never uses their product, if they did, they'd fix the problems. He told me to go try to do [thing x]. I spent a few hours trying to do the thing and I had to contact support twice and ultimately couldn't do it. It was the most enlightening thing I've done as a PM.

Since then on Colab, I've had a bunch of teams ask me about what our launch bar is. They talk about process and write docs with lists, and sure sure go do that all you want, but on Colab we launch stuff we like. Our launch bar is that we use it and we like it. If we build something we use and we like, we launch it.

There are some people who think that they can replace the lived experience of using a product every day with user research. Replace might be too strong of a word: let's say they'd prioritize user research over lived experience. I disagree with this position; nothing against research, research has helped me in many ways. But it's not everything. Get your hands dirty and do things yourself. You develop way more intuition that way.

You are building for yourself. You are the user. If you're not the user, how will you know what to build? Set aside time to use your product. Don't build products you don't use, let people who use them do that.

Using your product every day will lead to 1000% more success than teams who don't.

4. Be a good friend

One of the most important lessons about long term success that nobody ever taught me is you must make friends along the way. If you aren't making friends, you're doing it wrong.

I hate the word networking and I will never tell you to network as that sounds like making forced fake friendships. You must be friendly and authentic and make friends who you will help throughout your career if you want to maximize your impact. This is a critical aspect of your success. If you do not make friends you will limit your success.

When my previous company Bump was acquihired by Google (as I like to say, the easiest way to get into Google!), only a subset of employees were part of the acquisition. I found out years later that I wasn't on the original list and one of the senior staff had successfully lobbied on my behalf even though she herself wasn't hired. Nobody told me. She has never once implied I am in her debt even though I am. She is a great friend.

You don't need to be an extrovert to make friends. You do need to be helpful and friendly and reliable. You will naturally accumulate friendships over time as people will know you do good work and they can depend on you. Charismatic extroverts definitely have an advantage in friendship making, but you can still make lots of friends by practicing basic kindness.

You should spend some time figuring out what other people want and how you can help them achieve it. How can you change someone's life for the better? Find what people want and help them get it even if it means a sacrifice on your part. Friendships and trust are the backbone of fast execution. Be the kind of person who has both.

5. Give more than you receive

Part of being a good friend is always giving more than you receive. When someone asks you for help, help them. When you see someone needing help, ask how you can help them. I end almost every 1:1 with the same question: "what can I do to make your life better?" You'll be happy you asked.

When someone comes to you, choose to be collaborative. Be additive in discussions. Remember the "yes, and" of improv. Find the good in what people have to say at work, then find what you can add to that. Most ideas are dumb, but all good ideas started as dumb ideas. One good live conversation is worth a thousand PRDs.

One thing I hated about game theory when I took it in college was the prisoner's dilemma. I hate that the game theory optimal choice is defecting, it seemed completely wrong to how I would approach it in life. I then learned that the game theory optimal choice is collaborating if the game was repeated. Technically the strategy is tit-for-tat but that's a detail. Your optimal strategy in life is collaborating.

Sometimes people will steal your idea and not credit you and you will feel undervalued. I had this happen once and I talked to an engineering fellow I really admire. I told him what happened and he said, "I love it when people steal my ideas, it means I've helped the world". That completely changed my outlook, and I think about that all the time.

I want to work with someone who is generous like this, and you'll find people want to work with you.

6. Use the right lever

When you need to get something done, you need to find the exact mechanism that will get that thing done. Is it a bug? Is it a doc? Who specifically needs to say yes, and what will get them to say yes? Too often I see projects floundering because there's no clarity on what exactly needs to be done.

When I was on Google Photos I had something I needed from a partner team. I went to their PM and asked them for help. The PM kind of debated me for a bit and was completely non-committal either way. I went back and had the same meeting again.

My boss asked me what was up, and I, kind of exasperated, was just like, this lousy PM isn't helping me. My wise boss asked to come along to the next meeting, which we had, and went about the same way, but my boss seemed completely unperturbed. We walked out of there and I was like, "see what I was talking about?" and my boss shook his head with some obvious disappointment and was like, "didn't you see? The PM has no power to make that decision, you need to talk to the design lead." Though he probably said it more colorfully than that.

Everyone has opinions in life. Some of those opinions are good. Shipped is the best kind of good. When you need approvals, find the right people. Those people don't always have the highest level or the right title. But don't let casual opinions from the wrong people block you from shipping.

If someone comes to you and asks for your opinion, render your opinion to the degree you have the right ownership scope. There are enough forces slowing people down, you don't need to add to them. As Michelle Pokrass said, "the universe does not want you to ship, but you must do it anyway". There's enough entropy already without you.

Find the right lever and pull there. Pull in the wrong place and nothing will happen.

7. Only go where you add value

When I was helping redesign the Assistant, I had a phenomenal UX lead and eng lead, to the point where I stepped aside for a lot of execution and let them vibe. At one point I told them I didn't think they really needed me, and my UX lead was so kind she said, "we need you to help us to know what problems to focus on." I don't think that's a globally true statement so don't get all strategic on me, but the point I'm making is find where you add value and do that.

Be like water, fill in all of the cracks. Find every gap and get in there and close it. You're not needed in every instance, connect the right people and cut out the middleman (you!). As Stephen Covey said: "If two people have the same opinion, one is unnecessary." Be necessary.

Sometimes you need to help your excellent coworkers know what to focus on. Most of the time you need to do the work nobody else wants to do. For years I was the only person who read user feedback in Colab. Every single bug, hundreds per week. Logan Kilpatrick literally just yolo tweeted out his email for API support. I wrote all the first docs for Gemini CLI. In my first month at Google I carted around a bunch of pre-release Android phones and took pictures of random places on campus to give ground truth data to the research team for calibrating across phones.

There are many types of work you can do. You can talk to users, use your competitors, forge alliances, or anticipate and derisk a myriad of challenges. You need to do whatever is needed to plug any gap that exists, doing a million things that don't scale. You have to get dirty. I've never known a PM good at strategy who wasn't also great at execution.

Do the dirty work, make everything happen.

As a corollary to this, you shouldn't go to places where you don't add value. If there's a meeting that you don't add value to, don't go, you're wasting your time. If you're on a team that doesn't appreciate you or the value you're adding, you should leave. The most precious thing you possess is time. Be careful how you use it.

Add value however you can. Go where you add value.

8. Success has many parents, but failure is an orphan

If a product launch succeeds, everyone gets the credit. If it fails? It's your fault. I'm sorry but I don't make the rules. I know it's not fair. I have failed many times and every one hurts more than the last.

When you are building, you need to be hyper vigilant that you're doing the right thing all. the. time.

What is success? What is the right product to succeed? Who are the users? What will they do with it? What is the full set of activities you need to engage in to achieve success? Who will do those things? You don't need to solve every problem, but you do need to elevate people who have the solutions and make sure the right idea wins and the right thing is happening.

At a previous company I had this TL who was such a stickler about pushing me to explain every piece of logic for any decision. I was pretty annoyed about it the first few times? But I quickly realized what a gift that was! The best kind of engineers are engineers who care enough to push you to have the best plans and who have opinions. The worst partners are partners who don't care. I miss working with that engineer a lot.

My biggest failure at Google was shipping a Drive <> Photos integration. It was super heavyweight, it took months of work (quarters? I've blocked this out) and it was a giant monolithic launch with a huge amount of complexity. It was so bad it was unlaunched a few years later. I launched it even though I didn't love it, I didn't want to use it, and we only built it because [unnamed exec] asked for it. We all knew then what I know now, that a much simpler import flow solves 99% of problems. I should have done that. I've been sad about my part in that integration ever since.

Success is shared, but at the end of the day, if your product fails, you alone fail.

(please don't deteriorate your mental health over my language here, I've failed many times, you can recover)

You are a conductor of a symphony orchestra

I once got into a confusing conversation with a prospective team who said a PM isn't the CEO. I was scared off by this statement because it sounded like they wanted PMs to just follow orders from above (which I am maybe not the best at doing) instead of being independent creative problem solvers. But I was wrong for this, and if you're reading this, you were right, hiring manager. A PM isn't a CEO because a CEO can fire people and be dictatorial. You cannot! You are a conductor of a symphony orchestra.

I played the bass in a few orchestras in high school and spent many hours in the pit playing for the high school musical. The conductor cannot play all the instruments themselves, and may not even be able to play any of them. There are many personalities and different motivations. There are lots of jobs to be perfectly orchestrated. If you do it well with confidence, music will happen. If you are weak or bad at your job everything will fall apart.

Your job is to help orchestrate the work of many people. You do not have the privilege of having a big ego; you need humility to coordinate everyone else. As I've discussed, you need to:

1. Focus the team on shipping

2. Inspire the team for what is possible

3. Stay grounded in the reality of today

4. Develop friendships

5. Be generous with your time and energy

6. Be precise in how you will achieve an outcome

7. Fill in all of the gaps

8. Own the ultimate decision

This is hard work but it's needed work. It's hard to describe your job to your friends because it's a complex job!

Embrace the chaos before you. Embrace the diversity of the orchestra around you. Do the hard work and find order and create art. I believe in you. Yearn for the vast and endless sea.

You're doing good work.

-cperry

Special thanks to Shrestha and Rory for feedback on an earlier rough draft and thanks to everyone else who has helped me learn the lessons here. Opinions here are my own, do not reflect those of Google, and this does not constitute an official stance or guidance.