Navigating the Challenges of Offshore Hiring: An Interview with Christophe Louvion

I had the pleasure of speaking with Christophe Louvion, an expert in offshore hiring. His unique perspective on Conway's Law, which states that teams design systems the way they are structured, was fascinating. According to Christophe, hiring full teams instead of individuals is the key to offshore success. This is because hiring one person at a time can lead to entropy and gravity, preventing you from getting the teams you envision.

Instead, he suggests finding partners that can deliver teams that will mimic your culture. This means looking for the same caliber of talent, expertise, and full teams of three to seven people. Flying out to meet the first batch of teams and establishing the culture you want to build is also important. By hiring whole teams, you can realize your vision faster and build a strong offshore team.

Interestingly, the recent layoffs at Twitter have shed light on the dangers of overstaffing and the resulting lack of productivity. While some argue that Elon Musk's decision to fire 80% of the company was crazy, others believe that he was right. The real lesson to be learned is the importance of working on what matters the most. Most companies have a lot of work that does not matter, so they need to figure out what matters the most. If they have too many people to do this, then they need to make deep cuts that hopefully won't kill the company.

Twitter's layoffs have been successful because they cut deep to what matters. While there may be some latency issues, Twitter is still up 24/7, and many users feel that their feeds are slightly better. It is hard to argue with the outcome, and it is likely that Twitter will continue to thrive.

Transcript

[00:00:05.450] - David

Hello, everybody, and welcome. Today I get the opportunity to talk to a longtime friend, fellow member of the Lacto Forum, accomplished CTO and COO Christophe Levion. For you. Don't know about don't know. Christoph A. You should. Good. But let me tell you a little about him. As I mentioned, he's a CTO. He has lots of experience in lots of different companies in product development, engineering operations. He's had leadership roles both CTO and COO at companies like Episource. He's a founder of innovaxis chris office. Done a lot of stuff, a lot of wisdom. I'm always happy to hear him at Lacto Forum meetings because inevitably, he has unique insights and things to say. He's had a lot of recognition, including the Noir Award from the French Academy of Innovation. He has Bachelor of Science and a Master of Computer Science degrees and just a generally smart guy. So we get to talk with him today and hear more of his wisdom.

[00:01:26.410] - Christophe

Good talking to you again.

[00:01:28.200] - David

Good talking to you, my friend. Good talking to you. So the genesis of this call or this talk was something that you said in a recent meeting we're both in, and we were talking about hiring offshore teams, and you had a unique perspective about hiring people offshore, and it was about teams as opposed to people. So let's start off by talking about that. Let's talk about what's your perspective and how you think about offshore. Like it? Don't like it. How do you hire and what success that brings?

[00:02:08.980] - Christophe

All right, so a lot of people think about offshore as a cheap way to get access to some talent, and they kind of busted the work because that's what they do. They hire with no purpose. What my experience has been is very different. I look at the world as an opportunity to apply the old good Conway's Law globally. Conway design your team the way you want the systems to be designed. So if you do any modern development and you think in components and systems in the cloud, you could think about your teams the same way. And if you think about orchestrating diverse set of teams with clear responsibilities, the location doesn't matter as much as the structure of all those teams. And if you're in a good-sized company that has a good growth, hiring one person at a time is just extremely painful. It's very slow and tends to drive you away, actually, from this concept of Conway, where you're going to have a new component, a new system, and a new team. So for me, it's about mirroring the intent of the things you want to build and act really fast to get in that position, because entropy and gravity, if you hire one person at a time, will prevent you from getting the teams that you envision.

[00:03:43.600] - Christophe

So you're going to make a pretty diagram of how your system you would like it to be, and you're going to start to hire, and your team will look like nothing in your system. Your system will never end up the way you want it to be. So if you believe in Conway, then you have to hire full teams very fast, or you have to hire your teams very fast. And for me, the fastest way is to bring a whole team, five, six, seven people, not bigger in one shot. Give them a Pivot system and repeat.

[00:04:08.910] - David

So how do you hire a whole team as opposed to hiring individuals? How do you how do you think about that?

[00:04:19.570] - Christophe

So now you have to think about geographies. Ideally, the first thing that will come often, people say, what about time zones? It's important. It's not very important. A time zone is an artifact of likely having a fully designed system that requires very strong day to day interactions between people. So if you have the ability to really orchestrate the system you want properly, a time zone in most cases are not going to matter much. If you can have good teams around the world, the same talent, you can hire United States, in South America, in Southeast Asia, in Europe, wherever that is. You need to have the same caliber of talent, the same expertise, and full team. It can be autonomous. Product design engineers make them fairly small, like three to seven. Five is a good number. So if you have that, then you need to find a number of partners. And you have plenty of partners that are not so good, and you have plenty of partners that are very good, that can deliver teams that will mimic your culture. So I think what I'm looking at when I go abroad is like, do you have partners that try to sell you widgets like people one widget at a time, or they have the ability actually to form those teams and to deliver those teams such that they will be integrated into your culture.

[00:05:48.440] - Christophe

So you look for this. There's plenty of people that are capable of doing this.

[00:05:53.430] - David

You're hiring companies that are hiring the teams for you in the various different geographies. You're not flying out there and do it yourself.

[00:06:02.610] - Christophe

Yes and no. I find partners. But if I get into a new geography, I will fly there and meet the first batch. You need a very strong anchor. I've teamed in Vietnam. So I've been in Hanoi and Hochimin. And you find the first nexus. You build the first team, you meet the people, you describe the culture you want to build, and you really hired that first team. And a potential trick if you grow fast, a few months later, if you have a team of five or six, you hire very senior with the intent of raking that team, where each of them can become a leader, the next teams. So you can go from one to five teams, and then you hire more regular stacks of teams. So flying there is definitely valuable. If you intend in growing a geography beyond a handful of people, I would.

[00:06:59.910] - David

Argue you have an abstraction problem or abstraction issue you have to solve, because you're then, yes, you're hiring the team, but you're hiring the people that are hiring the team. And so you need to be able to evaluate them. You talked about some of the attributes that you look for or the ones that you avoid. Can you step us through with thinking about how do you find the right partner and how do you interview them? What do you look for?

[00:07:28.110] - Christophe

Yeah. You don't interview the partner. You describe the type of team leads that you're looking for. You interview them, and once you find the right leads, then you go back to the business and say, okay, now I want to build teams around that. Here's the conditions for building those teams. You have to drive that. You don't let the partner drive how you work. And every company has their own philosophy of how to build good teams. If you have a good philosophy for us, just replicate the exact same thing and find a partner that will let you do that.

[00:07:59.470] - David

But are there attributes for that partner that you look for before you engage with them to know that their recruiting is going to be effective, their filtering is effective. I understand that you're setting, like, here's the kind of people we want to hire, but how do you evaluate those partners to see if they're going to be good partners for you?

[00:08:23.970] - Christophe

The way I think about it is very simple. You have a quick conversation, and you test, like paper, conversation and powerpoints. They don't cut it. So I've probably tried 50 in the last 20 years, and there's five left that are really good. It's the same way as how do you hire. You want to hire a director in engineering in Chicago. What do you do? You interview them and you make a decision, and you hope to be right, and you do all kind of testing. And sometimes you're right, sometimes you're wrong. You have to put some effort at the beginning and understand it may not work. But I don't think there's more to do than just hiring a good maybe hire. They couldn't have a director of manager of engineering and do that work properly. And if you hire in the United States, you would count on that director to build a team. Well, if you hire in another place, you can count on them if you've done due diligence. And you'll be right sometimes, sometime you'll be wrong. Got it.

[00:09:30.970] - David

Okay. And when you're putting together the relationship with that partner, are there any things you think about differently than you would do for a recruiter here in the States? Are there any other T's and C's that you think about?

[00:09:46.040] - Christophe

No. It's really like, I'm trying as much as possible to set them as a PO ENTT as much as possible where it's mostly about payroll and recruiting. I'm interested if you have maybe some.

[00:10:00.650] - David

Local training capabilities, but I give the same no matter where you work, no matter if you work directly on W Two 1099 Peo Contractor I expect my company to treat everybody the same and to give you the same coaching, training, everything the same. So once we have access to the first talent and you just replicate the same way, I don't think there's much difference. Now, some companies don't like that. They want to control their people, right? And it may work for some projects if you just send requirements on one side and expect software to come the other. But if you want a team, a real team, you have to control all of that yourself.

[00:10:47.790] - Christophe

Let me see if I'm mapping to what you're saying is make the outsourced partners, thin a layer of veneer as possible once the hiring happens, and then they're just people that work for you. They just happen to be in Vietnam or Thailand or Brazil with the caveat.

[00:11:06.540] - David

That you have local regulations and laws and all of this, you need the partner. Because even if you were to hire yourself, if you were to build your own company there, you would have someone that has all that expertise to deal with everything legal, HR matter, and it varies quite a lot, right? Like there's requirements around the world that are very different. So you need a partner for that that is ahead of. And the world is changing left and right. If you have people around, you could see not only salaries are moving, but just everything's changing around the world. So you need partners to help you navigate through that. If you were to do that directly, you'd have the same issue.

[00:11:43.650] - Christophe

Are there any countries that you prefer or don't prefer? I know India's got employees need to give 60 days notice to their okay. Yeah. Before they leave their company, which means they get a job somewhere else, and then they are really on the market. So sometimes it's hard to hire. In India, are there certain countries that you find in some way better or worse? As far as just any attributes?

[00:12:12.720] - David

Yeah, I haven't done all of them, but the ones that today work pretty well for me are all the countries in Latin America. You're usually in the same time zone. And again, it's not only about costs, it's about finding people that for which you have a chance to build and retain a lot longer. I think that's the competition in the world is to retain talent, not to acquire talent. So Latin America, all of it, southeast Asia, Vietnam in particular, but the Philippine is also pretty useful right now. India is great, but it's great talent, but it's very difficult. The 90 days is a killer in the process.

[00:12:53.110] - Christophe

I like South America too, just because of the time zones. And there's a lot of great talent in Brazil and Argentina particularly. That's why I found yeah, but that means also because of time zones are you picking? Let's go from hiring to what kind of projects you're giving them, what kind of things they're responsible are you choosing? Go back to Conway's Law. Here's the thing, you're responsible team in country X, are you picking what areas responsibility based on time zones, on culture, on language?

[00:13:33.190] - David

Strategy for me is pretty simple. Try to go further possible for United States, assuming your roadmap is fairly stable such that you don't need too frequent of synchronous conversations. That's the driver. So again, I would never separate by functions. QA in one place, engineers in the other. That's a disaster. So full teams, so anything that it's stable. So it's not front end, back end necessarily and stable. I don't mean like from a system perspective. I think from a roadmap, like the flip side, you try to do an MVP, your customers are in the US. And you don't have a long roadmap. You're changing your opinions a lot. You do a lot of customer meetings and design Sprint, and we do as close as possible as the customer. So if the customers are in Europe, build a team in Europe. If your customers are in US, build in US. If you have customers around the world, maybe build a few teams around the world. So I would look more about what needs to stay versus everything else should move around.

[00:14:49.240] - Christophe

Okay, got it. And product is in these teams.

[00:14:52.760] - David

Oh, yeah, for sure.

[00:14:54.100] - Christophe

Yeah. By the way, for what it's worth, that's what I feel. Okay, let's talk about the other side. Let's go from hiring to what we're seeing a lot trend now is layoffs. Let's just talk about elon for a moment and then we'll go general layoffs that are happening. Yes, Twitter. So he walked in and said, we're way overstaffed. I'm going to fire 80% of the company. And then he did. There's an argument which is, oh my God, he's completely crazy. There's an argument. Is he's right? There's an argument he's wrong. But we haven't seen the effect significant downtime of Twitter. We've seen some because there's latency. What's your take on what he did? What lessons do you draw?

[00:15:52.460] - David

All right. And it could be all of the above at the same time. But listen, I don't work there, so I can't comment on the specific. Though I have maybe some ideas. It's likely a piece of it's cost, like he got Twitter and wants to cut costs, but I don't think it's the driver. I've seen over the years that the more people you add, the less outcome you get per person. I'm sure you've seen that. Most people have seen it. And what's happening in most cases is you have different dynamics at play. The two that I can think of are a very powerful negative dynamic is the first one is the functional division. So people start more and more to be divided by their expertise and then sub sub expertise. And so making any simple little thing becomes a complicated problem of coordination and that just slows down significantly. At the same time, you also have a layering that is happening, removing people that used to be very close to their end users, their customers, to the value they get further and further away from the outcome. So the chance for having work, to be busy, work that doesn't have value, just explodes.

[00:17:08.350] - David

So companies, usually they try to fix this with project managers, which makes things worse and adding even more people. Because if it's too slow, you need to add more people and it implies even more the problem. Like it's the natural entropy, I guess, like gravity. It's hard to ignore the weight of all those people and the way companies, I think, can fight that. There's two ways you have to work on what matters the most. And likely most companies have a lot of work that does not matter. It matters to somebody, but it doesn't matter to the organization. So you could try to figure out what matters the most, but you have too many people to figure it out because everyone will think their thing is important. So if you really don't know what matters, well, you cut deep in a way that hopefully doesn't kill you. And what matters will be very visible and it will get done really fast. I'm a twitter user. I see it up twenty four seven. I even feel my feed is slightly better right now. Maybe it's just placebo effect, but it's hard to disagree on the outcome. It doesn't shock me.

[00:18:21.630] - Christophe

No. You don't think that there's latency that we're going to see something break or actually, maybe your argument is you don't know yet. Maybe that was what your argument up in the beginning. But actually, let me say what you said back in a couple of sentences and you tell me if I got it right or wrong is over time, things get overbuilt as far as charts and there's not an obvious way to clean that up besides making arbitrary cuts. See where the bleeding is, stop the bleeding, reorganize, and the things that people will tend to scream more about the things that matter.

[00:19:00.770] - David

Yeah, because it's really hard to figure out at scale unless you really have a very strong deliberate mechanism for it. Are we doing something that matters a lot? It's hard to know. The litmus test could be if that person or that group is gone, are we going to fail our 2023 revenue or whatever metrics? Do you know that? And when you don't, that means you likely have way too many people because it's not correlated. If you don't have a good sense of the work, most of it is supporting the success of the business. Well, you likely have too many people you don't know and you're going to keep on hiring more and more. And likely he will cascade to even be further away from value. It's hard to know unless you have a billboard system to manage that.

[00:19:50.910] - Christophe

Okay, so team topologies fan or foe?

[00:19:55.470] - David

Fan.

[00:19:56.670] - Christophe

Fan.

[00:19:57.610] - David

Yeah.

[00:20:00.270] - Christophe

By the way, for me too. Me too. It was a blinding glimpse of what should have been obvious.

[00:20:05.830] - David

Yeah. And it's synchrology like everything else. I don't think it's a framework as much as opening the conversations for thinking about what are teams for, you have set up teams that likely ideally, at least two thirds of your teams are directly and it's back to what we're talking about. If you apply team topology, all your product engineering team should be very connected to some customer value. And you should know that that team is providing this value. This team is running that value. It should be very visible. And if those teams are fairly small, and if every team owns a little piece of the puzzle but itself has a very direct connection to value, you will know if you have teams that you don't need or not, and then you have supporting teams and all of that. I think it's a great framework. I don't think it's a framework, but it's a great opener of conversations to evaluate a non functional, organized company by those autonomous team that should most of them provide value. And then you have supporting and enabling and all the other flavors. Yeah, I think it's a great frame. It's a great solution to solve the overstaffing problem.

[00:21:28.190] - Christophe

Do you recommend people I'm going to say something a little more fascist than you said. People at CTOs at regular times slash their teams to see where things should be done with fewer or do you have another mechanism you suggest?

[00:21:48.860] - David

Well, ideally, you have the framework in place to not get there, that you always identify that your organization is supporting directly the business KPIs. Now, if you haven't done that, you likely would be getting value of of cutting to see how much over staff you've been. But you shouldn't be there. And you have to understand very clearly, you have to work with finance, sales and marketing. If you don't do that every day, you'll actually have like I think that's the job actually, I believe today mostly now in this economy and everything else, where valuation on companies based on revenue multiple, that is not going to cash flow. Talking about value. Well, cash flow, I mean, if you read the PnL, you should be able to 80% of your team, this team is plugged here into the PnL. If you know that, you're likely not going to be overstaffed. But if you can't align your team again, all the product engineering team, 80% directly to align in the PnL directly, not through five indirections, you likely have too many people. I think that's the job of a CTO. Yeah, but platform teams are very important. But I put a caveat to this.

[00:23:13.550] - David

I think about a platform team as a team that is demonstrating, accelerating other teams, not mandating that you must use my technology for the sake of it. Their KPI should be explicit because you exist, other teams go faster and you and platform team need to demonstrate that because if you can't demonstrate, you're likely then slowing other teams down because now they need to integrate into you, which means you're not autonomous anymore.

[00:23:41.910] - Christophe

That's a great bit of wisdom about measuring platform teams.

[00:23:45.920] - David

I think it's the only thing a platform team does. You should measure, you should demonstrate, you make other teams faster. And if you can do that, you're not a platform team. You're a bottom neck to everybody else.

[00:24:00.270] - Christophe

Yeah, I completely agree. I completely agree. What do you suggest for CTOs that have to do layoffs? That they either choose to or they have to because the macroeconomic effects about communicating with their team and explaining what's happening, why it's happening. My experience is that when you do a layoff, you lay off N people, you always lose N plus M. There's people that go and flee, that are scared and they may not be the ones you wanted to lose. So how do you decrease in that equation? M.

[00:24:40.730] - David

Yeah, I don't have a great method for the M, but for the approach, I think you have to be very honest. People are smart. I would go something like this. We're not making our numbers. Again, that should be very clear because if you are organized, you would know that. But we're not making our numbers. We're cutting deep. We're reorganizing teams where every team, 80% of the product engineering team will be aligned to numbers. We're going to cut the work. That doesn't matter. So the one thing the worst for me is, is when I hear we're going to do more with less. I hear we're going to burn down people to the ground and lose even more people. So instead of doing more with less, I would say we're going to do less with less people. We're going to work on things that matter more, we're going to bring more value. And I'm going to ensure that every single person in every single team has a very clear understanding of how their work is either directly interacting with business value or if in the platform team directly help others in ways. When you go home at the end of the week, be very proud of the value that you brought to the company.

[00:25:58.500] - David

That's the message, I would say, yeah, you'll lose a few more, deal with it. How your new team? That's needed, as we discussed. But if you're very clear with why you're doing it, and if you have a path to get people as much as possible in a way where the work they're going to do will be meaningful, you're not going to lose a lot of people.

[00:26:22.570] - Christophe

So it's about outcome versus output when you do that cut, and I explain that.

[00:26:27.580] - David

Yeah, absolutely.

[00:26:29.230] - Christophe

So one last question for you. What do you know now that twelve months ago, if I told you were to be true, would have surprised you?

[00:26:44.400] - David

Oh, chad GPT. It's not the topic, but yeah, Chad GPT four just came out, right? Yeah. I think to me it's as big as when I remember the internet in 97. And I know there's a lot of people saying it's amazing, a lot of people say it's just a little trick when you really play with it and really go, we've done some exercise on my company to go through all the things we do and where you can get value. It's a big list.

[00:27:19.800] - Christophe

I got to tell you. I love what Microsoft is doing to embed it in the other products. I think there's a big innovator's dilemma problem here. I think Google is going to try to do the same thing, but those are just the beginning integrations, and it's going to change markets significantly. Even if it's only 80%. Right? Yeah.

[00:27:39.670] - David

Because a lot of the workflows in organizations is about taking some data, doing some processing, and a lot of it is not only about the math of it, it's actually extract how many companies have Etl teams. And it's not only about, again, an Excel file and I put it into a snowflake or something. Most companies now look at data as the treasure, and likely 99% of all the information is unstructured today. So it's a huge stepping stone to dig gold in every single company. For me, that's the big surprise.

[00:28:28.320] - Christophe

Well, thank you, my friend. Once again, I've enjoyed our time together. If someone has questions for you, what's the best way for them to reach out to LinkedIn? Okay, well, then we will reach out to you at LinkedIn. You'll find Christoph there. And as I said in the beginning, he's a great mind in our field. And so if you have questions, you should definitely hit him up there. Well, thank you, Christoph.

[00:28:55.040] - David

David, thank you for the invite. Have a good day.

[00:28:57.320] - Christophe

Yeah, you too.