Subscribe
Share
Share
Embed
课代表立正的官方Podcast
深度访谈，有用干货，亲身验证的「真本事」
Superlinear Academy创始人，Maven Top AI Instructor
前Statsig布道师（OpenAI收购），腾讯副总监，Meta，Amazon；康奈尔经济学博士
社区：Superlinear.Academy
课程：ai-builders.com
个人：lizheng.ai
The 101 is your best opportunity to manage up to your managers. They're like, oh well, my manager manages me. No, a lot of the time, the relationship is actually you managing upward. Using your opinion without data.
Would that stop a PM or a leader to voice their opinions? No, they're gonna make a decision. So what's stopping you? Apple model is when you're pulling in your manager. The push model is when I need to be pushing them for updates. But if I still have to do that for a long period of time and I don't see that transition to a pull model, that for me is a concerning sign.
Hi, Mushtaba. Welcome to the channel. How you doing? Nice to be here. For audience who do not know, Mushtaba is a senior manager of data science at Meta. You have been at Meta for seven years, but you didn't start as a data scientist. You were kind of SDE at college, training, but you then joined Meta. Is Meta your first job?
Yes, Meta was my first job out of school. I originally studied engineering, so not like a formal math or stats background. I had like one stats course in university, but yeah, I had done a couple internships in software engineering, and then I joined as a data scientist as a new grad, and slowly worked my way to senior manager over the past seven years.
Nice. Maybe it doesn't take a dependence then of math or stats, or maybe you do that with engineering. I think you do it with engineering to some level, but I think the bigger aspect of the role that probably engineering helps a lot with is like it's about problem solving, which is a lot of what you kind of like cover throughout your engineering time. I think.
Data science and specifically data science at Meta is really about identifying root causes of issues, identifying problems, being able to like put together clear proposals and influence the rest of your team on putting them in the right direction. So a lot of that, I think, is agnostic of like whether you study stats, whether you study science or math or engineering. It's really about like your ability to break down problems at its core.
Love it because today, like you can, I can learn math or stats concepts very quickly with AI. But the ability to solve problems is what defines a good data scientist. Well, one hundred percent. And I want to ask you kind of three questions. Just pick your favorite to start with, and we'll go with the flow. The three questions are how to do one-on-one with your manager.
And the second question is from a manager's perspective, what is a good IC? Like when you pick people for promotion, what kind of person are you more likely to pick? And the third one is, how can I get help? When do I get help? What kind of help can I get from managers as an IC? I think a lot of people they are not manager, but they interact with manager.
Just how to best use a manager, I guess is the theme of my question. And before that, to put your question in context, how many people have you managed before? So, I initially, when I had started managing, I was managing a smaller team of around four people. I think at the most for the past little while, I've been like as an interim lead on ecosystem. So, I was supporting a team of approximately like fourteen to fifteen people.
I'm going to be transitioning soon and switching teams. So my new team is starting out at ten people with also like a layered manager underneath me. So I think a bit of experience and exposure to like.
different levels or layers of management, as well as like a larger and smaller team, where it also requires a different role of me as a manager. Cool. I think from the questions, I honestly I like the order that you just said them in because I feel like they naturally flow into each other. So I'm gonna kind of stick with that to start. Your first question: How do you kind of use your one-on-one with your manager?
Personally, I think there's a one thing people learn eventually, and they should learn over a period of time, is that your one-on-one with your manager is something you should be driving.
Early on, especially when you join a company or you're onboarding, you are hoping that you're going to go to your one-on-one and your manager is going to tell you everything you need to know. But over enough time, once you've onboarded, the one-on-one is really something you should be driving and you should have ownership on. You should be setting an agenda. You should be coming with clear like questions or topics or things you want to cover. But for me, and this maybe comes up in one of your later questions, like a strong IC is one who has a clear ownership of the one-on-one agenda and what they're trying to get out of it.
But I would break down like how to use your one-on-one into a couple clear categories and topics that even managers should take this advice because I use this for my one-on-ones with my manager.
First is like I cover FYIs, which are more just like tactical updates. These are things that your manager should know are going on, like things that are on your roadmap or on your workstream that you just want to flag are in progress, so your manager is aware of them. I like to honestly put them in the notes, and I have like I usually a one-on-one doc with my manager, so we can keep track of everything over the months and over the weeks as time goes on.
But I don't usually discuss them. They're like, I just need you to know this is happening, but I don't need to talk about it because what I really want to use that time for the most is actual discussions, things I need your input on, things I need you to action on, or information I need from you.
I don't want to spend too much time just giving like updates. Like, is it a standup? I think once you're, if your one-on-one is more like a standup, and you're just listing out things you're doing to your manager, then it's not really an effective use of either of your time.
It should be more about discussion. You don't want it to feel like it's a stand-up where you are kind of just listing things to your manager that you're doing. You want it to feel a lot more like a discussion, and you're getting input and going back and forth.
Nice. I remember hearing this advice from many successful people who who have successful careers. Ah, one mistake, common mistake people make in one-on-ones is making it a project update. But I think I remember the advice of not making it a project update. But if I don't do that, then people feel like, oh, should I just talk about my personal life?
Yeah, so I think that that gets kind of into the other things I I had in terms of categories to cover. So the first is like just FYI is that you wanna honestly list and have as async updates so your manager's aware of. The real use of your time is gonna be in things you need to escalate. So the one on one is your best opportunity to manage up to your manager.
As a concept, some people might not think about it too much. They're like, "Oh, well, my manager manages me." No, a lot of the time, the relationship is actually you managing upwards, you informing your manager things you need them to act on or things they need to know. So, the first set of updates after the FY FYI for me is like risks to projects, things that you need to escalate to your manager. So, like you were saying, sometimes people treat the one-on-one too much like just a project update.
But when I give this feedback, sometimes people over pivot and then they won't tell me anything, and they'll tell me when it's too late and the project's about to like fall at risk or be or fall behind. Your skill set is really about letting me know when the project is at a point that I need to get involved. Is there difficulty in an XFN collaboration? Is there some risk with some of the analysis going on? Is there blockers with data or with like XFN like?
Getting some context into like risks to the project or risks to the timeline is the type of thing we should be talking the most about. Not every single thing on the update, so that's a key area to kind of like highlight in your one-on-one.
I think we might talk about this in the third question as well, but there might be a stigma in asking for help or surfacing risk. Like some people feel like if I flag a risk to my project, it means I'm not competent. I cannot handle things.
So the mistakes some people make, in my experience, and something I have tried to coach a lot of my team on, is you can let me know what projects are at risk or what what analysis or what work is going on that might be at risk, but also clarify you don't need me to do anything about it. There's value in you just letting me know that information and letting your manager know that information, because one thing PICs might or like reports might not be aware of is
Your manager is likely in a lot of meetings and reviews and rooms where you aren't, and they might be asked about the status of this project. And it's important you and your manager are on the same page about things. So if there's risk to work, it's not bad for you to tell your manager. There's no stigma around like, oh, my manager is going to think poorly of me. No, it's like you can tell me, hey, I have this under control, but I'm having difficulty with this collaboration, or I'm actually having difficulty with data access, but I think I can unblock myself.
But as long as I know that level of risk is there, I can make sure when I discuss with the rest of my XFN or other leadership, I communicate a certain level of like hesitancy or like risk with the scenario. So I don't put you in a position where I say, "Oh, actually, no, I talked to using the project's perfectly fine. It'll be ready by Friday." Because then that puts you in a position where you're more likely to fail and puts a lot more pressure on you.
Because it'll be like, well, your manager said everything was fine. For me, this is a lot more about making sure you and your manager are on the same page about things, rather than me letting my manager know I need their help. Because a lot of the times, even when I have ICs and they tell me what projects are at risk, they'll let me know they don't need me to do anything about it, or.
When they come to me, for me, the difference between like, "Hey, you are not doing a good job," and I need to get involved, is if you tell me something's at risk and you haven't tried to unblock yourself, like you were just coming and hoping I'm the first person who's going to solve it for you, versus you come to me, you let me know what you've already tried to unblock, and then I get involved to help because now it's at beyond your control, and that's fine. Like you should be relying on your manager to help unblock you because that's part of my job responsibility.
Do your homework, and when you cannot do it by yourself, then ask help from your manager. But that analogy reminds me of something. I think what you are saying is we should be objective, we should be honest, and of course, like to get things done well, you need to actually understand not only the good part, but also the bad part and the risky parts. But that comes from kind of a trusting relationship.
And I think a lot of people are in this mode of I don't trust my manager, because maybe even from school, they feel like, oh, my teacher is someone who are gonna punish me for not doing my homework, or grade me, or evaluate me.
So they like bring this mindset to work, and they feel like managers here to evaluate my work, to fire me if I do things wrong. I understand where that stigma comes from, and to be honest, I've had that same fear myself in the past. But I think the difference is your manager, as much as you might think they're there to evaluate you, they're also there to advocate for you. Because when it comes to performance reviews, when it comes to these conversations that go on in rooms, your manager will be the one who's trying to represent you the best.
So you might feel like I can't trust my manager. I don't know if I can trust them to tell me. But you also have to ask yourself: Have you give them an opportunity to see whether or not you can trust them? I had this experience early on when I had switched teams and I joined ecosystems. I was a really new senior IC, and in my head, I was convinced: Oh, I need to be autonomous.
Like because I'm senior, there's an expectation that I can solve things on my own. I can solve problems on my own. So I was working on a project, and it got to a point where I was having a lot of difficulty with like some XFN partners.
And I didn't escalate it. I didn't tell my manager. I assumed telling them meant, like you were saying, oh, it's like the stigma that that means I'm not autonomous. Like I'm not being independent enough that I need their help to like resolve these X, Y, and difficulties. And as a result, I had to like push back the deadline for the project. I had to talk to my manager, be like, hey, we're not going to hit.
This is not going to be ready for the leadership review. We're having difficulty with the model, and I'm having difficulty with my partners. The learning experience for me, because after that, my manager came to me and told me, he's like, "Hey, you just have to let me know. Like, let me know. You can tell me when there's a problem. You can ask me for help." And he's like, "I'll tell you common mistake most people make, and I tell a lot of my reports this, which is they think they're asking for too much help."
more often they're not asking for enough help, and he gave me advice. He's like, I will tell you this: you can ask me for as many questions or as much help as you need.
And when I think you're at the point where you're asking for too much, like you've crossed the line beyond what my expectations is for your level, I'll let you know.
And the reality is, I never got close to the line. And most people don't get close to that line because we convince ourselves on what is.
This idea of a senior, very strong, independent person, but that's rarely what the reality is. The more senior I've gotten, the more managers I've met, I've met, I realize everyone asks for help. Everyone asks their managers or other people for help. So we think we have to be very independent, strong people, but that usually just means we're not asking for as much help as we should be, or as much help as everyone else is.
Yeah, and my personal experience is as a manager, it actually feels very good to be able to help, because we are so far away from doing the actual work. I tell my team this all the time. I was like, every so often, I wonder, am I adding value in your life? Like, what am I doing? How am I being helpful? And then they'll come to me with a problem, and then I'll be like, oh yeah, that's I can still help. I can still add value. It is reassuring as a manager. At some point, the idea is we like.
Build our team up to a point where we like automate ourselves out of a job. Where like you guys are all very strong, independent. You don't need me. It is very reassuring at times when they ask questions and you're like, okay, I can still help out.
Yeah, yeah, and I think the other thing is, managers, their performance is not how many people I fire, is how many people I help get promoted. Hundred percent, hundred percent. Performance management is a part of the job and is something you have to do.
But I would say it's the worst part of the job, and it's also not something that we highlight the most as like something for your promotion or your strength as a manager. What really matters a lot is your ability to build your bench, build your team, kind of get them to be more senior, get them promoted, build their strengths, such that you can scale yourself more. So at any point in time, it is in your manager's best interest to build you up in your career as well, like beyond.
just the fact that they probably care about you and care about your career growth, it's also in their best interest to build you up in your career. Nice. I think this naturally transitions to the second question: like, what is a good, strong IC that you are more likely to help get promoted? Good question. I think a lot of this also kind of aligns in my head of like how I think.
What are the skills that I think get a person to like get promoted, get to senior levels? Like, what are the type of stuff we care about? It varies at different levels, but for me, one of the biggest things is like someone who has a good balance between the idea that I'm talking about of like autonomy, but also managing up. Like, I know when one of my ICs I can trust when I know they'll come to me when I need to get involved.
If I can trust to leave them on their own and know that the project's not at risk, because before it gets to that point, they'll let me know, they'll pull me in, they can manage me up correctly. That's a level of like, okay, this is a strong guy. See that I can trust and I can kind of like leave to own a scope and own an area.
Because I trust the level of communication they have with me, that I'll be aware of what's going on, I'll be aware of problems, and then I'll be aware when I need to get involved. I think.
ICs that I don't trust as much is the ones where I have to pull a lot more. So I talk about this push and pull model a lot with some of my ICs. So a pull model is when you're pulling in your manager, you let them know, hey, you need to get involved here, hey, you need to do something. And I trust that they'll pull me in when I need to get involved. If not, I kind of leave them on their own. We'll have our weekly check-ins, we'll have whatever conversations, but then I can trust to leave them on their own.
The push model is when I need to be pushing them for updates. I need to be checking in because I don't know what's going on. That for me is more something I do for my more junior ICs, and that's fine because they're still learning and growing, or someone new who's onboarding.
But if I still have to do that for a long period of time, and I don't see that transition to a pull model, that for me is a concerning sign. Then I was like, "Hey, I need you to own this more. That's owning the agenda, owning communication, and owning managing up to me, such that I'm can trust you to leave you on your own in places."
Let me try to dive one layer deeper. Sure. So, what triggers you to push to feel like you have to be in this push model? Yeah. What are the signals? So, for me, a lot of the times it's like.
We are working towards a project. We are working towards a review, and I don't feel like I'm getting regular updates. I don't feel like I'm aware of the current status of projects. I don't feel like I'm getting updates on milestones or situations that I should know about in terms of the progress of the work.
A strong IC is someone who can kind of like break that down and communicate that regularly to stakeholders. So not just to their manager, but to the rest of the team. So people don't feel like they're in kind of the dark on what's going on. That for me is a very strong IC when they're able to manage all the stakeholder expectations, all the collaborator expectations well.
And in some scenarios, your manager also falls in that definition of a stakeholder collaborator. Sure. So I start getting worried, and I start getting into a very push model when I was like, "Hey, we're working, we're getting close to deadlines, we're getting close to this place, and I don't feel like I know what's going on." That's when I start pushing more.
And it's because I feel like they haven't pulled me in yet. I don't think you have to only pull in your manager when you need their help. You should be pulling them in just so.
You let them know things are fine, and they don't need to worry. Like if if I get if I get pulled in earlier and I get I see earlier versions of drafts and I see things are fine, then I can continue to take a step back and I don't need to get involved. Like I don't need to do anything. I just need to be aware of the state of things. I'm trying to kind of do an abstraction of this. First, most people make the mistake of undercommunicating.
Yeah, very, very few people make the mistake of over communicating. We should ever on the side of try to do more communications. That's number one. And the second is, like, what are we actually doing? We are trying to make sure the project is successful. It's about the project. It's not about individual performance. And when people under communicate, often they are worried about, oh, how other people would evaluate me if I say this, if I do that. But if you look,
Take the perspective of let's make sure the project or the things get done well. Then you kind of have a natural like understanding of.
呃，when should I do the communication and what kind of what level of of information should I disclose? 100. I think that's a really good like way of framing it where.
If you can separate yourself and your personal like concern of like, hey, I don't want to look bad. I don't want to look like I can't be independent. I can't be autonomous from the situation, and you focus more from like the project ownership standpoint of like, hey.
Am I making sure the project is going to succeed, or is the project at risk? Then you might be a bit more willing to like actually move it forward and have a better communication style. There's one other thing I wanted to kind of highlight in terms of like strong IC and I think strong performance. The stronger ICs that I have had, I think have a very good self awareness on what their strengths versus weaknesses are, and I think this is.
probably also like correlated with just seniority, where the more senior you get, the less hands-on I have to be when it comes to like career development or career trajectory. That's not to say I don't have career conversations with my senior ICs, but I think my senior ICs have a lot more ownership or advocacy on what their career trajectory they want.
Like they're very well aware about what their strengths are. They know what their weaknesses are, and they communicate that to me. So my role and my focus is really about like how do I put them in the best position.
to capitalize on their strengths. I think with the weaker and not necessarily more junior ICs, but with the weaker ICs, they don't have a good self awareness of where their strengths or weaknesses lie.
They are still like kind of trying to push through things that they don't realize are not strengths of theirs. Is it obvious to you in most cases, like the weakness, like they don't know their own weakness? But is it obvious to you and other people?
Yes, I think for the most part, if you give with enough time to actually properly calibrate and evaluate someone, I think it is fairly obvious to me. I think the gap and the issue becomes when it's not obvious to them, because then there's a level of like strength that I need and competency that you kind of need to get to in every skill by certain levels. But I think beyond that, like when you really think about growth in
more senior levels, the focus is really on like how do I get the most out of my strengths? How do I put myself and focus on positions where I can capitalize the most on my strengths? Because after a point, it's not really about improving on weaknesses. It's more about just like ten times they kind of get investing a lot more in your strengths.
Yeah, like obviously, like if everyone can do your job as well as you do, or everyone becomes replaceable, so you need to amplify your strengths. But I want to deep dive, like why do you think it's the case, like?
I I do like it's also my experience that someone's weakness is obvious to other people and their manager, but not obvious to them. And it's very hard for manager to actually point that out without you know hurting feelings.
The less you are obviously able to like identify, obviously the harder it is as a manager's job for me to be the one to kind of break that to you. But I think a lot of it actually does have to do with self-reflection. One thing I try to do, kind of going back to like the one-on-one question, is I separate out weekly one-on-ones with career conversations. I think it's like really hard to have a proper career conversation where we're so focused on tactical detail.
And in a lot of my career conversations, I usually open up and talk first about before anything about performance evaluation. I ask you to reflect on how everything's been going.
What do you think is going well? What do you think is not going well? What are your career trajectory goals, and how do you think you are progressing towards them? I think when people aren't asking themselves those questions or aren't thinking about those things critically, they're not fully aware on their strengths or weaknesses. And even after all of that, if they still aren't, it's likely they're just miscalibrated, like they have a false sense of where the bar is for certain skills and strength. And then it's my job as a manager to make that very clear.
And sometimes that's totally fair. Sometimes that is on me as a manager. That's my failure point if expectations for you aren't very clear on what the bar is, and that's where I start putting in. I have to put in more effort to make sure that's clear to you.
But in my experience, the stronger ICs are the ones who are very self-aware on where they are below the bar or what gaps do they have to like the next level or where their strengths lie. Such that career conversations are way more driven by them. Similar to just one-on-ones are more driven by them. They have a strong understanding of like.
这些是 strengths I have. These are weaknesses I have. I would love to improve on these things, but I think I would really strive in roles that let me make the most of these skill sets. And then we focus on like projects and opportunities that kind of like match their skill set and their interests. I want to double click on this own your own career and career tag part, because one mistake I made in my career at Meta was I feel like I talked about promotion half a year too late.
Once I started talking about promotion with my manager, then we have a plan, and we have constant checking of where is the gap and how to find me more projects to fill the gap or to have more visibility. That is a very important starting point.
Talking about promotion with my manager, but why did I do it half a year too late? I was afraid. I was afraid how my manager would think of me. Like, oh, maybe I do not deserve a promotion, and then my manager is gonna be why is why is he asking for promotion when he is below expectations?
Like now, if those were true, my manager do not think that. But yeah, people have this imagine imaginations, right? I think this comes back to a point you're saying earlier about like it's about building trust with your manager. As a manager, one thing I have to evaluate and figure out early on is like what are people's motivations? Like what motivates different ICs on my team? There are people who are very clearly motivated by motivated by promotion.
Like they have, every time they get a promotion, they're like, okay, how do I get to the next level? Like, what is the focus? What are the gaps? And then there's some people who have, who honestly rarely talk about promotions. They just talk about like their life, work-life balance, skills, strengths, stuff like that.
As a manager, it's still my responsibility to make sure I set up both of them on good paths to promotion. Like, just because one asks me more about it, doesn't mean as a manager I should not be putting an equal effort for the person who doesn't ask about it. But the reality is, like at the end of the day, you are your biggest advocate as an IC. So
I do the same thing too. In ideal scenario, my manager is putting in the equal effort to think about my promotion, making sure I'm aware when I'm close or what my opportunities are. I still want to have that conversation, and I still push for that conversation. So I'm aware how far I am, and I usually start these conversations from a place of like vulnerability and very much clarity. Like I'll let my manager know, like, hey.
I don't think I'm that close to a promotion. For me, I think it's more than a year away. But it's helpful for me to understand where my gaps are and what you think my gaps are. So even if it's over a year away, I can focus on working on those things. So when we have our career conversations, you can let me know if those are being getting better, if I'm improving on them. I'm not asking you to promote me in a year. I'm asking you to just let me know what are the things I need to work on and let me know if you see that improvement or growth.
And coming from that level of transparency and building that trust, I think helps a lot. And the reality is, sometimes, to be honest, like you might think, "Hey, I think I'm more than a year away, or I'm very far away from a promotion." And that's not the case. Like I've had that happen. I told one of my managers at a point in time, "I think I'm a bit away from a promotion, so I'm not really sure. I'm thinking about team switching or other options and stuff like that." And they were like, "I actually don't think you're that far away."
And that's just because I didn't really know where the bar was at that point in time. So, I think being more transparent and being the owner and advocate for your own career.
helps ensure that there's like no lack of clarity, like there's no miscommunication, or you're not missing out on any clarity of expectations. Because in an ideal world, a good manager should be communicating all this with you, regardless whether you ask. But you don't want to leave that up to your manager. It's perfectly fine for you to push and ask for that clarity.
Yeah, I think clarity and nothing should be unexpected, like the right expectations. Yeah, there's like two failure scenarios in this. I think when you get to your performance reviews, if you are surprised.
by your feedback or by your rating, it's either like obviously there's only two ways it fails, right? Either you didn't have a very good understanding of where where you are and where your expectations are.
或，I think the bigger issue though is your manager has not done a good job of like communicating those to you. Because even if you were unaware and you thought things were going great, your manager should have already been giving you this feedback throughout the process. Your
Performance review should not be the first time you're getting any critical feedback that there's weaknesses or gaps. How do you kind of break the bad news to the reports in a constructive way? I think the most important thing is to just be straightforward and honest. It's hard, obviously, having like tough conversations, whether that's performance management, whether that's giving someone a lower rating, or whether that's even like
Giving someone a good rating but lower than they expected, or they didn't get the promotion that they were hoping for, it's tough to be the person who breaks that news. But I think it's also a responsibility in the role. I try to understand that we're all adults. I don't like beat around the bush. I get very straight to the point of like, "Hey, this is what happened." I am very clear on like what the gaps were that were discussed, where I think the gaps lied.
And then before I get really far into just like tactical discussions, I asked how they're feeling. Like this is disappointing. This is sad. I realize that you were hoping for something else.
And we could talk about that. Like, what are your feelings? What are your disappointments? What do you think I could have been doing better throughout the half, such that like this disappointment wasn't there? As a manager, sometimes you worry. You're like, hey, did I build up their expectations too high? Did I let them think that their promotion was gonna happen 100%, and I gave them a false sense of like confidence that it was 100% going to happen, right? So it's also important at that time as a manager to reflect on like, hey, what could I have done better in this process for you? Do you mind if I play a difficult scenario?
Sure, play a difficult scenario. You ask my promotion, my emotion is, it's not fair. I I work so hard, right? I did did overtime, I committed the most lines of code, and I did all those analysis, and I ended up getting a bad rating. It's not fair. You think it? Totally right. You did put in a lot of time, and you did write probably the most lines of code.
I think this is something we acknowledge and we did talk about in your performance rating. But I think the thing that we need to really focus in on and think about is not the lines of code or the hours of work, but it's the impact of the work that's coming out.
I think that's why, especially going into next half, something we should work on together is making sure your roadmap is set up in a way that you don't have to spend so much extra hours and so much more time writing so many lines of code, and you can actually hone in your efforts on the most impactful projects.
I think the gap in your promotion case wasn't the volume of work, was actually the impact of work. And I think there's right now you're taking on a lot of these like more execution, operational things, smaller things. That is taking time away from the more impactful work you could be doing.
So, it's a partnership, and I'm hoping going forward we can kind of work together to set you up on a roadmap that really focuses in on the type of projects that will set you up better for a promotion.
Yeah, it's it's like it's obviously hard. It's a hard conversation to have, and sometimes at the end of the day, regardless as a manager, whatever you could say, people will be disappointed. And it depends on the company. It varies from the company you're at, at place to place. But I think something is also important to understand is like it's not really just up to your manager whether or not you get promoted. Like your manager's role is to like advocate and push for you in performance conversations. But depending on how your company does it, it's not always like
their individual yes or no say on whether or not it happens. I think a good manager will communicate that to you. will let you know what the feedback was, what the gaps were.
But it it takes time. It's not like you people get over it right away. Yeah, I think I don't know how many people have the realistic expectations, but I feel like.
People are becoming more mature and having the realistic expectations. Like things are not entirely in our control. Industry, company, stock market, economy, like everything contributes to a certain path. One last question about this. Uh, what is a good easy part? Maybe asking for the audience. Some audience may wonder, especially a data scientist. You talk about communication. You talk talk about project management. You talk about uh, you know, autonomy.
And no surprises, manage up. But you didn't talk about technical complexity. And a lot of people, when they think about what is a high-level data scientist,
Imaginary, if you think about someone who are expert in certain domain, can do complex things, technical, uh, complex. Yeah, can elaborate. It's a common question. It's a very fair point.
And the way I think about technical expertise, and the way I've seen it come up a lot, is like there is a baseline level of competency when it comes to technical skills or technical complexity ability that is required for success in a lot of places.
And the reality is, I think a lot of people who get hired at these companies and work with us in places that we work at have that level of technical competency. What ends up being majority times the blocker for growth and the blocker for more impact our work is not.
The technical expertise, it is some of that softer skill. It is the product intuition. It is the communication ability. It is the collaboration skills. I have seen way more cases where that is what blocks someone from having more impact. That is what blocks someone from being able to like work at a larger scale or on a larger scope. That ends up blocking their promotion case. Then.
the technical complexity barrier, and even if it is the technical complexity barrier, it's not their lack of ability. It's the lack of like scope within the area they're working for that complexity to exist. And it usually is a sign that you need to be working on a different team or on a different project that has the right amount of complexity.
Yeah, I have worked with some really smart data scientists at Meta. I'm sure you have as well. And I've worked with some really strong technical people. But even amongst the most technical people I have met,
The most successful ones are the ones who are also able to like connect and bridge that with like the right communication skills, the right prioritization skills, and the right collaboration skills to actually translate that into impact. I have met really strong technical people who get really deep into areas, but if they can't convince other people why that's important or why that matters or collaborate with other people,
It doesn't really change a lot of things. It just kind of stays with them. Yeah, the the sentence we influence other people. We don't write production code. Like company hire us to produce impact. And if we solve a hard problem that has no impact, we actually wasted a lot of resource. Not yeah.
We say this. I bet I'm sure we say other places say it too. Like our currency is influence, and it's your ability to convince people to make the right decisions.
How to increase that influence? It's a good question. There's a lot of stuff that kind of builds into it. Honestly, over time, I think it's similar to like building your brand, right? Like as like influencer, as someone who's trying to build their brand on LinkedIn and YouTube and other places, it's the same thing you do internally as a company. You have to build trust with other people. They have to be confident in your technical abilities, your competency, but then also
Have confidence in your intuition. Like they'll want your opinion on additional things. They'll reach out to you for your input on other things. Like when you have a history of successful influence and a history of like being able to like identify opportunities or make the right recommendations, then there's a feedback loop where people will come to you more.
Usually, when I'm new and I start off, I I usually try to find low hanging fruit, quick wins, kind of build relationships and connections because I feel like that's like the foundation of then being able to influence anyone.
That's such a important point that I don't think is discussed enough. The right intuition, because I actually ask in the early days, like marketplaces org, the centers have a lot of influence. Like people listen to data, and I ask the early director Ashish and Nav, how did the marketplace have such a strong data culture?
And the answer was, early days, Ashish just insisted on his opinion, and he was right a lot of the time. So at beginning, people may not listen to him, and over time, he builds the reputation, influence. One thing, and this is maybe unique to like larger tech companies that you realize over time is, even if you don't provide
Good data or right data, people are going to make decisions. Like people are going to make decisions and move forward. And like the company is gonna, there's going to be motion.
So it is important and imperative that you kind of just help direct them in making the right decision and moving in the right direction. Because if you don't, they're going to move regardless. No. I used to get this feedback a lot. I have never been really good at like sharing my opinion. Every time like I enter a new room or enter a group, like I usually tell myself, everyone here is more senior than me. Everyone here has more context than me. Everyone here has more life experience than me. So like, why would my input and opinion matter?
And my managers and Skip have always pushed me a lot to like voice your opinion, or like you clearly have an opinion, but you should share it more. I'm getting comfortable with the point. Everyone is just like acting and sharing their opinion based off of the information they have.
As data scientists, we tend to actually have maybe a lot more information than the average XFN or collaborator we work with. And if you think about it from this point of view, data is like the voice of the user to some level. So, to some degree, it is your responsibility on behalf of the users of your product to like bring voice to their concerns and bring voice to their data on like what do they like doing, what do they not like doing, what do they want more of.
So if you don't do it, and if you don't provide that voice to the users, then we're gonna likely make decisions anyways that might just be bad for the product and the user experience.
Yeah. This reminds me one of my most valuable lessons at Facebook was a senior data scientist circle led by Roberto. He's now VP at Instagram. Really smart. He asked this exactly scenario: How many of people who went to the circle are comfortable reasoning opinion without causal evidence? And I think half of the room raised our hands. And then he asked, How many of you are comfortable reasoning your opinion without data?
And I think only he raised his hand, and he said, "Would that stop a PM or a leader to voice their opinions?" No, they're going to make a decision. So what was stopping you? Yeah, you get to a point, and honestly, as like when you in a manager, a lot of the times, if you're like a lead in an area, that is what you're trying to do. Like you're just
If it's a new area and there is no data, you are basing it off of whatever intuition you have, whatever historical data and everything you understand, and you're just making a recommendation.
I think the important thing is to make sure you have like safeguards in place, so we don't just one hundred percent invest based off of like faulty information. But every other function is probably going to make recommendations and decisions without data, anyways. So
There's nothing stopping you from doing it, but also trusting that that is still not zero data. You have a history of seeing a lot of data points that you've kind of internalized that is influencing your intuition. Yeah, yeah. The decision is qualitative. It's never like A is bigger than B, so let's go with A. Yeah, yeah. Even with causal data, it's a strong evidence, but it's part of the evidence, not the entire evidence.
If we could wait every single time for causal data to make any decisions, we did, we would not be moving fast at all. I think judgment is very important, and intuition is very important, and that's probably like the core of product analytics at most these tech companies.
Nice. I will summarize this as: strong ICs should have a good mental model, right? They should have the right intuition, and they should be able to be right a lot, even without data.
So that's the data scientist internalized within the person. Great. Let's go to the third question. So.
How to ask help from a manager? When to ask help and what can you ask help on? For me, and I think we talked about this when we talked about like what's a strong IC. You should ask your manager for help once you have tried to unblock yourself first, right? Once you've made an effort to unblock yourself and you are still blocked, then you kind of go to your manager with.
What is going on? And your manager can help in a lot of ways. Whether this is like an issue with collaboration model, like there's people you're working with and there's a difficulty, and then you need.
An escalation at your manager's level, and then their manager's level. Like you kind of need to get your manager involved when there is any level of like seniority involved. It's important to kind of like work with your manager because at some point it is.
Beyond your scope, like I would not expect some of my ICs to go argue with a director or go argue with a VP, and they shouldn't have to. Like that should fall to my level, and that should fall to me to like protect them from that level of like pushback, because there's clearly power dynamics at play that make a conversation and pushback hard to have.
But it is important for me to know what what have they already tried, what have they already communicated, what has already been attempted. I think there is like a lot of areas where your manager can help with just additional context that they have.
And this comes through a lot in communication and kind of like going back to the idea of a good one-on-one. Like, part of what you do, what your manager helps with, is just context. Like, your manager's in more rooms and meetings than you to get context on leadership or stakeholders on the top of mind problems for the company that you might not have context in. So, your manager can help connect your work and your analysis to the larger strategy.
And you should lean in on that context your manager has. You not having that context is not a failure point of yours. There's rooms your managers in that you aren't, so there's no expectation that you should have that. But you should be able to like connect with your manager to get that information. So a lot of what your manager can help with beyond like unblocking you with problems is also connecting you to the right information and context to like make your work more impactful. So for me.
That's something you should. Your manager should be trying to tell you on your own, but that's also something you should be pushing your manager for. If you are working on a project and you're not fully sure why we're doing this, or you have questions behind.
What the actual question is, the motivation is. That's something you should be pushing your manager on. That's something your manager should be helping you on and providing you with. Nice, yeah. Like you mentioned, the influence is the capital, right? And get influence to get people listen to you. You have to talk about what they're interested in. You are talking about a different dimension. Even if you are saying all the right things, doesn't matter to them. Yeah, like.
I've I tell my team this sometimes because I think on ecosystems a lot of our work is just about like influence at the overall app level.
Impact is a combination of you doing really good analysis and people really caring about that problem. At times, it's hard to have impact. You could do really amazing analysis, but if the
Stakeholders and audience aren't thinking about that problem right now. It's a very hard hill to get over to convince people that the problem matters, and then your recommendation and solution matters. So at times, it's really about like connecting the right skill set with the right problems people care about.
And that's where like your manager helps a lot. Your manager should be finding you the right scope, the right opportunities, the right problems to connect you with. And on your own, you might not have access to all that information. So a lot of
What I do for my team, especially for my most senior ICs, is just really helping, like, brainstorm and kind of build their roadmap with top priorities, top of mind concerns. And then when they're actually like writing and communicating their work, making sure they're able to connect their findings with other things going on around the company and app that are of
呃， strategic priority, just so it has more opportunity to land impact. How often do you review their kind of writing, communication, and provide feedback? It varies on the IC and on the seniority. I think for more junior ICs, I would probably have a couple iterations. Like I'll review a more rough draft where I provide, like I usually try to.
For my team, especially now as it's larger, I like review the very initial outline and plan just so I make sure when they start they're going in the right direction. And then I'll review at the end when it's like pulled together and like clean it out, help iterate them on framing. So.
Depending on who the audience is, like if this is going out to just more their people at their level and their ex-fn, then I might review it once. If it's going to fairly senior audiences, then it's not just me. It's usually like even my layers of skips who are all reviewing it, and there's like a couple layers of iteration where we're like working together on the framing and cleaning up the communication and trying to connect it to a narrative that matters to leadership. But I would say.
Um, two to three times. Nice. Yeah. So as your manager, escalation when there is power dynamic involved, try to unblock yourself. But when you couldn't, and also ask your manager for information, contacts, and connecting you with the right people, information, contacts.
And all of these things are not just true of your direct manager. This is like the scales when you think about your management chain, right? There are certain things that even your manager can't help unblock. You have to go to your skip, like.
I like talk about the meta example, right? I work on Facebook app. If I have issues ever with like one of the other orgs or one of the other apps, like at some point, even my manager is not the one who can help unblock. I have to go to my skip. I have to go to a very senior VP level, and you kind of like rely on them.
So the same situation kind of scales. Like, understand when is something escalated to a certain difficulty or collaboration difficulty that you need to go even further up, and then when is context that you need at a much senior level. Like, I need context at times on what's going on in leadership reviews or leadership meetings, and maybe even my managers. Not in those. So then I reach out to my VP of like, hey, I just need some context on this ask we got. I just want to know what was
Where was it discussed? What were you looking at when this came up? What additional things were you guys talking about? I just need some additional context. Nice, those are good questions. Can you give a few more those kind of questions? How to extract more information from your manager? What questions to ask? I usually start off by just asking what was the context in which this question came up, because
As many data scientists probably know, they usually get asked a very random or specific point, like "Hey, we just want data on this feature," or "Hey, we're looking for some data on these users." And I'll be like, "Okay."
Can I get some more context? What was going on? What were you looking at or discussing when this topic came up? Okay, I think the most important question to be asking is, what decision are we going to make with this data? What is like the purpose of this data? What is the decision you're trying to make? Because
What's often going to be the case is there's likely better data you could provide to help with that decision beyond what they're asking.
But because they don't know the data, they're gonna ask for something. But if you know the question, you can actually provide them better data or more data. So I always tend to try to figure out what is the context in which these discussions were going on. What were you looking at? Were you looking at some existing data? Were you looking at some existing trends? Can I see that? And then two, what is the decision you're trying to make?
The reason I think sometimes it also helps to ask what is the decision you're trying to make is if the answer is there's no decision, you can also push back like is this a priority? Is getting this data going to help in anything? Is this going to change our mind? Is this going to change a decision? If not, is this a higher priority than everything else I'm working on?
And that's often like for me a sanity check on prioritization and pushing back, because I think every data scientist has experienced getting asked a lot of.
data asks and everyone wants to always have more data, and an easy way to push back at times is asking what is this data going to help us do? Does it matter if the data's not going to change any decision? Then do we really need it? Yeah, I think that's the key question between kind of a SQL monkey versus a data scientist who can influence decisions.
And it also helps to understand, like, what level of precision are we trying to go for, right? Do you need, like, what is the level of?
Precision you need in whatever decision you're trying to make. Are you looking for like a ballpark estimate? Do you need to know like is this like a one percent or ten percent or like a seventy percent opportunity or sizing? Or are you getting trying to get very specific where you need like very clear confidence intervals? Like there's a high level of risk of being wrong, because that lets me know how much time I should be investing.
And I think, especially the more senior you get as an IC or even as a manager, that's like what you get better at being able to gauge of how much time should I be spending on this ask and asking the right questions to figure out how how much investment you should.
谢谢。Thank you. I think that answers my first question very well. I think we got to both the message and action items, and they can just take away and start using tomorrow. So, any questions you think I should ask you? Do you think they care about the IC to manager transition at all? Like what it was like? What was easy? What was hard?
Is that too personal? I think people care about that, but it has been asked and answered too many times. Like every time, like people ask that question, and it's hard to say something new for this. Totally fair. No, I think because the focus has all been here about like how do you use your manager, what is the how do you
Kind of like connect and talk to your manager. I think we answered. I think it's more than three questions because, but we went really deep into all the topics and the follow-ups. I think we kind of covered a lot of it. Nice. Thank you. How do you feel about the conversation?
It's for sure awkward the first time you do it. Like I feel I I probably talk too much at times and I ramble a bit. But it's nice. It's like I.
It's interesting because I'm talking to you directly, but I can feel like I'm talking in front of people, even though they're not here. Like you get the sense that you're trying to talk to a larger, broader audience.
Yeah, that's I try to bring that perspective to the guest I have conversation with. I don't know, like the anchor that I try to use is if you send this video to our direct reports, is it going to help them be more successful? If that is yes, then sure, we got enough value out of this conversation already. Yeah, all the rest is I don't know algorithm, many different things that out of our control.
Makes sense. I said the same stuff I tell all my direct reports, anyways. So I would happily send this to all my direct reports. Yeah, and I don't think you talk too much or too little. I think it's the perfect amount, and it definitely do not feel awkward to me. Yeah, natural. Next, we should be on more podcasts. I, well, I'll be on another one in two weeks with you. It'll be fun. Nice. Cool. All right.