If you have inept or inefficient managers, then you can end up having a very inept or inefficient organization. The middle management actually becomes the most critical piece of a hierarchical organization. If you get that right, then you get excellence recursively and if you get that wrong, then you get mediocrity recursively. When you look at any startup, when you think of innovation coming from like the ground up, those startups don't start off with a hierarchy, right? They don't say like wait a sec, we're going to be set up like this and have a manager and then an individual contrary, like they try to set up to be as flat as possible because that's how you know most people feel.
So, it will create an environment which is more innovative where the ideas can be treated as more equal. Hello and welcome to The Engineering Leadership Podcast brought to you by ELC, the Engineering Leadership Community. I'm Jerry Lee, founder of ELC, and I'm Patrick Gallagher, and we're your host. Our show shares the most critical perspectives habits and examples of great software engineering leaders to help evolve leadership in the tech industry.
Let's get ready to rumble. Welcome to our community's first ever debate. We have Farhan Thawar and Jerry Crickelli joining us to hash out which org structure should rule them all. Should you go flat or should you build a hierarchy? One disclaimer, I am so excited for you to listen to this episode. I went all out for this conversation, mixing in just about every sports and politics analogy and metaphor that I could put in.
So, you've been warned and one request as this is our first ever community debate while you're listening, tell us what you think about the format and give us your opinion on flat versus hierarchical which org structure would you choose? So, kick off a conversation and tag us in a comment on LinkedIn or Twitter.
Now, let me introduce you to our candidates in today's debate. First up, in the flat corner, hailing from Toronto, Canada, we have Farhan Thawar, VP of Engineering at Shopify. So, Farhan joined Shopify via the acquisition of Helpful.com where he was co-founder and CTO. He was the former CTO mobile at Pivotal and VP of Engineering at Pivotal Labs. He's an avid writer, speaker, and advisor at Y Combinator and one of Toronto's 25 most powerful people.
And, in the hierarchical corner, hailing from San Francisco, California, we have Jerry Crikelli, senior director of engineering at Facebook. Jerry is the former VP of Engineering at House. He was an engineering director at Google and helped develop early versions of the display ads serving infrastructure. He helped launch YouTube ads as well as video ads on mobile apps. Enjoy our first debate with Farhan Thawar and Jerry Crikelli.
Gentlemen, thank you both for joining us today. For our audience, for everybody tuning in, both parties have agreed ahead of time on the rules and for matters of the debate. They actually have not heard any of these rules. So, this is the first time they're being exposed. But, for the sake of this conversation in the metaphor, they have agreed to these rules ahead of time.
So, here's the rules. We expect a clean debate, no funny business, especially no touching of the hair or face. Each person will get two minutes for an opening statement. And then we'll get into a few topics chosen by the candidates. So, we'll talk about culture, innovation, and velocity. And then, each candidate will get two minutes for a closing statement. We may have an opportunity for a question from our town hall.
So, if you are in the audience, have a question, submit them ahead of time. They'll be screened by our committee to make sure that they are good to go. So, with that, Jerry, since you won the coin toss backstage, you'll be going first. So, gentlemen, to begin our match, what I need you to do is to fist bump the screen and then we'll begin.
Okay, fantastic. Jump into our opening statements. Jerry, since you've won the toss, you'll be going first. So, Jerry, kick us off, make us an opening statement about hierarchy as an organized design structure.
So, first of all, let me just say that in any organizational design, there's a lot of different options and it's not necessarily that one organizational structure is better than another. It's more that it's critical to know which organizational structure is better and when to use which style in this particular case, I'll try to argue that I think a hierarchical organizational style, although in some ways less popular because it has more of a
serogatory connotation because it's hierarchical, actually provides a lot of flexibility in terms of scale because in general, you can actually incorporate a flat organizational style into a hierarchy as well. because many large organizations have siloed, have siloed teams or have different parts of the organization that are run completely independently.
And if you think about a hierarchical organization as a tree structure, then within each branch of the tree, you can actually practice a flat organizational style as well. You can have a pretty large fan out in each branch of the tree and give a lot of independence to whoever the manager is for any given branch. Another advantage I would say of hierarchical style is that it easily allows for geographic distribution and a variety of talent composition.
If you have a flat style and you have a large variation in talent, it's pretty difficult to move quickly because you're not going to have every person comparable to all other people. Whereas in a hierarchical organization, if you want to have an injured team, you can easily accommodate for that and collect the group of ninjas, have them execute in specific areas with full autonomy and then have the rest of the, have the rest of the organization unchanged.
More importantly, it easily allows for geographical and nowadays remote distribution because it's very easy to pick specific projects, place them in particular geographies and have them operate independently.
And finally, flat style becomes harder and harder to manage as the end, the number of people in that organization scale, it becomes more difficult to communicate and it becomes more difficult to coordinate you end up having like end to end communication as opposed to one to end communication or org to org communication, which is why in most larger organizations or one organization scale.
If you look at the structure for governments or if you look at the structure for corporations or even the story for that matter in most instances, kind of human society has evolved into a hierarchical structure just because it tends to naturally size tends to naturally gravitate in that direction. Jerry, thank you so much for your opening statements far on your opening opening statements leave the floor for you.
Great thanks Jerry I can't believe you said governments is a good example of hierarchy I think just walking right into a trap there. We all know governments are not actually designed to get many things done in effective manner so I think maybe you're you're giving me some points there. So the reason I really admire the flat structure is actually because what you're seeing is a you want you don't want to have a big distance between the CEO and individual contributor.
So I've definitely been in many situations whereby you've got management chain that has a broken telephone effect and in many companies that ends up diluting the information that goes back to the CEO. I remember this at a large company I worked at about 10 years ago where we would have red light green light yellow light on our slides showing how things were going.
So I think there's information lost there when you look at really effective leaders you'll find that they tend to cut through any sort of hierarchy that's in place anyway. So if you look at like a Warren Buffett who has a 250 director ports actually because he sets individual compensation for all those CEOs and there is no hierarchy between myself when I was running a startup called extreme labs I had 120 director ports.
So I was running a light manager and the reason for that is people tend to fall back on the management techniques and they just use it as like a checklist. So for example you know maybe people have read somewhere you should have one manager for 10 director ports five managers to a director. So they just blindly implement these things without understanding what they're trying to implement.
So I think that's the downside to a hierarchies that it just tends to be set at some in some place in the company as a standard and then everybody rolls it out, whereas it should be different for different teams and can be much much flatter. I remember you came from Google. I know Google had a few experiments around having 30 or 50 director ports to a manager as an experiment to see could we get more done.
So it's actually a great line one of my friends from McKinsey says which is managers create work for each other. So when you have managers and they create work for other managers that's not real work customers are not seeing that impact.
And actually all it is is just sucking up the air in the room and we know that there's no way to actually figure out what part of that work is effective because you know smart folks will just do work right for 40 hours a week and you'll never know if that's actually manageable. So what I would say is there's actually ways to create a flat structure shorten the distance of communications of the CEO and have people really enjoy their their work life.
I know that extreme labs many of my former employees come back and say that they wish they could find that type of environment again because they were able to get so much done without any managers in between. Thank you so much for a hunt. Jerry, thank you again for your statement. So you both have laid out a good high level case for why for hierarchy or why why for flat. So for the next portion of our debate we're going to be diving into a couple specific topics of a couple key categories.
Our first category is culture. And so for this for a hunt I'd like to turn to you first to give you an opportunity answer and then Jerry will have an opportunity to rebuttal as well. So for a hunt can you talk a little bit more about culture as it relates to a flat organization. Sure, I would say the biggest takeaway for me in a flat organization is that what you try to figure out is where the individual contributor is getting feedback from and how they're learning in the organization.
So for companies people feel like the craft leader can only be the manager and so they're learning directly with the manager and the manager is probably spending one to two hours a week with that person and helping and maybe you know if you read high output management right that one to two hours of input can be very valuable and can maybe drive 20 or 40 hours of work.
So I always the case and I saw this again working at large companies where my one on the ones over a three year period where I wouldn't say lead to any useful output even though I was doing them every week versus in a flat organization you tend to be since head of a structure where you're learning from your peers.
And with your peers you're actually if it can be high fidelity and intense you can learn from them on 20 40 hours a week because you're collaborating so much and the flat structure really allows is especially when you got many, many more people than just five or 10 in your peer group imagine 40 50 and Mike is 120 Warren Buffett has 250 right all of that they're learning from each other and that actually I find is much more effective from a cultural standpoint because it's part of it's not part of the one I want that happens once a week.
It just happens as part of your every day working hours for on your your response.
Sure so first of all let me kind of address the government issue it's not that all governments are bad it's just governments with bad people are generally bad and I think that's also emblematic of if you think about a hierarchy at the end of the day it's not so much the problem that the problem is not that there's a hierarchy the problem is that if you have inept or bad managers and the hierarchy then they create worse and worse.
Organizations or they apply a Peter principle where in general if you keep promoting and promoting people within a hierarchy and you know that you promote them to a level of incompetence then you end up having only incompetent people and management levels and that's a terrible way to run an organization.
I think Google to some extent solved this problem explicitly by only promoting people to the next level one day forward even functioning well at that level and that essentially guarantees that you don't end up in this Peter principle where you have inept managers on the contrary. You end up taking people that are very capable at the level below and are already operating at the level above and then you put very capable people at a level higher to mentor the people below them.
I think one thing that from a cultural perspective that a hierarchy allows is it enables multiple cultures to evolve within a single organization because if you think about a hierarchy and you have various branches each branches actually allowed to have its own culture and to develop its own culture typically in a flat organization.
You certainly have an evolution of culture but that culture tends to be monotonic because if you have too many cultures at the same time it becomes much more difficult to get things done and so you end up having an environment where you end up having to hire a lot more people that are very similar and have similar viewpoints because otherwise it would be difficult for them to execute.
Naturally in a hierarchy each branch of the hierarchy can hire different types of people and you can have multiple cultures that kind of coincide at the same time. As I mentioned at the same time it is a lot easier to scale a hierarchical organization because even if one branch is dysfunctional it doesn't mean that all branches are dysfunctional.
You can see this in many organizations especially even large successful companies have departments that are very successful and very efficient and in some cases you even have departments in a hierarchical organization that have a very large fan out factor to far hunts point I used to manage 30 people directly before we just it became untenable from an individual perspective and you couldn't assign as much attention.
So I began to become more of a believer that at some point in order to scale effectively you need to have some sort of a hierarchy because Google did start out with a model that used to have a very very large fan out per director and it became detrimental to the individual contributors underneath that director because although they had complete independence they had very little attention and you could give them very little feedback as a consequence.
So if you think in a hierarchy you have a lot more of an opportunity to have a diversity of different cultures and you have a lot more of an opportunity to have diversity of management styles for manager the most important thing is just getting the management right because I think Parna's aptly pointed out that if you have an after an efficient manager is then you can end up having a very an after an efficient organization and so the middle management actually becomes the most critical piece of a hierarchical organization.
So if you get that right then you get excellence recursively and if you get that wrong then you get mediocrity recursively. Thank you Jerry for like the review 30 seconds response rebuttal if you'd like to respond. Sure yeah so you know back to the government point I think what what I was saying there is that it's not a model that most companies try to emulate right like hey this is the hierarchy that we think is an efficient organization.
So if you're a feedback point what I think makes more sense is to develop a feedback culture whereby the teams and the customers and the peers are giving feedback so you're not dependent on like trying to get 30 or 15 minutes with your manager which allows you to have that much much larger fan out and I'll say the last thing is.
You do the same thing at Shopify which is like we want to make sure that people who get promoted perform well in the level before they do get promoted but it's still because of a management layer possible in any company to hide more easily when there are lots of managers around because there is all sorts of work that can be generated from that level.
Yeah I mean I think at the end of the day in any organization the excellence of that organization is going to be a function of the people in the organization and the management top down I would say management culture of the organization if the management believes in effective middle management and they hire and try to optimize for that then you're going to get.
Pretty well functioning organization to far hunts point if you drop the ball and you start promoting people based on popularity or based on whoever can get along with the most number of people as opposed to who can effectively lead organizations and obviously has is collaborative.
And you end up getting much better results even as you have larger and larger organizations naturally the more number of people involved in anything the more difficult it becomes right in some ways not to get kind of too political but the more people that vote for something the harder is to make a decision on something because there's a lot of opinions in the room and so in that regard.
When you have hierarchies or when you start to break things down it's like problem solving if you if you break things down into pieces or you divide on conquer it's easier to solve the piece of a problem and to have more and more people try to solve the same problem at the same time I think even within flat organizations you naturally end up selecting a subset of those of the people in the organization to make some key decisions otherwise you just end up having endless debates and it's very difficult to draw consensus as the number of increases in any given debate.
Thank you, Jerry. I think that segues into our next topic really really naturally so topic number two we wanted to touch on was innovation and and so you just laid out a little bit of a case of more opinions equal more time which to extend that a little bit means maybe less innovation and so I think to open this up I'd say far on if you like to open up about flat organizations as early innovation the floor is yours.
Yeah it actually segues nicely because Jerry says something that I think I'm super worried about which is that you want to break ties or have opinions of like the hierarchical leaders matter more than individuals on the team right because they're used as tiebreakers I think that's dangerous because what you really do want is like an idea meritocracy and you want that whatever the management layer looks like whether flat or more hierarchical to be more like air traffic control to help the teams not running to each other versus saying well I'm the director and I'm going to say I want an idea to do that to help you get to the world.
the director and I'm the most senior person here so let's go with my opinion. And so the same idea makes sense for innovation because you do want to have the best ideas flow to the top. And from what I see in some of your arguments Jerry I think you can still get them without having
to worry about a very strict hierarchy. So for example you could have multiple teams working on things that doesn't mean that they have to have individual managers to manage them even though it might be a team of like I said I was running a team of 120 people with with with which were probably 60 different teams. So it didn't it didn't matter that they didn't have their
individual managers. I think that the idea of mistake making and allowing the teams to make their own mistakes and not being like rubber stamped or accountable by a manager is actually a very important concept because right now in big companies what happens is you work on some project
and then some executive wants to put their stamp on it or have to have sign off on it and very late in the project they might overrule some you know decision that was made very early on in a project versus having a flatter hierarchy and actually saying to the manager hey your job is not actually be accountable. It's actually run this machine let the responsibility be pushed down to the individual individual individual teams and if they go out get feedback and or fail that's
okay because that's part of the experimental culture we're trying to build. Very your response. Sure I don't think that there's anything inherently wrong in a hierarchical organization that prevents managers from not necessarily being decision makers but more like decision facilitators. I think that there is a presumption in hierarchical organizations and in fact I would say that's almost a hierarchical organization done wrong when the person at the top and recursively applied
every person in a management position has to have a stamp of approval. In a good well functioning hierarchical organization every node of a tree ends up having an opinion and in some cases you know anybody who's at the vertex could be a facilitator and in many cases you allow everybody and if
you have a team you can have even a large fan out and the manager basically plays the role of gatekeeper where they're allowing everybody to have an opinion and so you could have a meritocracy of opinions synthesize all of those opinions and then either roll them up or present the top three to your manager and subsequently so on and so forth so it doesn't necessarily have to be that
the manager is always the decider. In fact some of the best managers are the best listeners and so they essentially facilitate the conversation across their teams and they make sure that every person on the team has an opinion that's being heard and given an opportunity to speak and then take the best ideas and either have them be voted on by a group or collect as many as possible and then
present them to their manager and so on and so forth until the best idea surface. I think the the failure of execution in such organizations is specifically when you have managers that function in a hierarchical style and then it's the job of an organization whether it's flat or hierarchical to weed out managers that have those tendencies because I think that's where the failure occurs. It's not so much that the problem is the hierarchy. The problem is the execution within the hierarchy
itself and how you end up executing. Lastly I think that ideas or how ideas are generated as more philosophical if the organization is open and allows for a natural democracy of ideas than any individual regardless of how far down in the chain they are if they're given an opportunity and
their managers presents it such that they're given an opportunity to escalate their ideas and those ideas will naturally gravitate to the top and I've seen that happen even in fairly large organizations or small organizations that grew organically when you had good people presented with good ideas
then those ideas ended up escalating. The problem occurs is when the manager either presents those ideas as as their own or only prefers their own ideas to be escalated for the sake of getting themselves promoted or whatnot and then you end up having a cultural failure but that's more of a person failure that those are just bad managers as opposed to the structure itself being inherently in a position that doesn't allow people to express their opinions or have those opinions escalated up.
Thank you Jerry. I'll give you about 30 seconds for a response and then I want to make sure we have time for our third topic so for your response. Sure, cool. Yeah, I think like Jerry I'm agreeing
with you on some of those points except that that you really have to thread the needle. A lot of your comments are around like well you have to have good managers and well you have to promote correctly and well you have to hire correctly and I think you know not every company is like a Shopify or a house that can get those things right so that you've got a management layer in there
that is effective and can help all of the you know company execute. In most cases all of those things actually anywhere it's super hard to do so while it's nice to say hey yeah you just have to do this x, y and z I find it in practice extremely difficult and some forgs can get it right
and some camp. The other point I'll make is that when you look at any startup when you think of innovation coming from like the ground up those startups don't start off with a hierarchy right they don't say like wait a sec we're going to be set up like this and we'll have a manager and then an individual country we're like they try to set up to be as flat as possible because that's how they you know most people feel it will create an environment which is more innovative where the ideas
can be treated as more equal. Thank you for hand. Jerry I wanted to call to one thing you had mentioned beforehand about the process of rolling up ideas and the meritocracy of ideas our third topic is about velocity. I think one one criticism that may come up of that process may be around velocity so I was wondering if you could make opening remarks on our topics about velocity as it
relates to hierarchical organizations. Sure I'll sort of make one comment in terms of startups and the reason that startups of higher velocity is that they typically have almost by definition most startups start small that's why they're called startups and so it's easier to have faster velocity
when there's fewer people in opinions in the room in some cases there's even prevailing thoughts around if you want to make a project go faster reduce the number of people on the project as opposed to increase the number of people on the project so I do think that there's that there's fundamental benefits to having a smaller group and the smaller the number of people that you have involved the more I actually do believe that you can have a flat style versus a hierarchical style.
At the same time I think that if you have a major initiative that you want to pursue in some instances having a hierarchical organization allows although maybe it's bad to put it in the terms of command and control it's easier to have directives in a hierarchical organization because you can
funnel them down more efficiently and then if you have every layer of the organization sign up on those directives or everybody believes that this is the right path for the company and this is the right direction and as Jeff Bezos puts it if you know even if people that disagree they disagree
and commit you end up having fairly efficient execution within a larger organization and so you can steer a fairly large ship well if all the different layers within that organization align on a particular goal however from an innovation perspective it becomes trickier because by definition
the larger the organization the more ideas get generated in many cases it's easier to lose specific ideas somewhere in the process but from a velocity perspective if you're trying to execute on a large project then having a hierarchy in some cases is a lot easier to get that execution done
then if you have 200 people that you have to convince to go in a particular direction and every single person has their own opinion because by definition there is no leader involved imagine how many sidebar conversations you have to have because by definition an meritocracy
the merit of anyone given individuals idea has to be better than all the rest and so you have a lot of convincing to do and a lot of people to persuade versus in a hierarchical organization where if you pick a direction and you can convince only the managers in that organization that this is the right path and they sign off on it then you can kind of get that execution going a
lot faster. Are you Jerry? For a hundred response. Yeah I think it's again a bit dangerous to think that the authority only comes from being someone's manager like I agree with what you said about having certain sets of folks who you can help convince who can then help convince like other
people in the organization I'm just worried that the only way to do that is because you've got like the title of manager or director or it has to come like top down like one of the best ways I've seen innovation combination of innovation and velocity flourish is that you've actually got
people who are best situated to make those kinds of decisions and they don't necessarily have to be a manager right so imagine even just squads of a team who have like whether design, engineering and product and different leaders emerge from that not HR leaders not like direct reports but
actually leaders with opinions or the best situated make decisions around other design and UX or technical decisions or product decisions and I find that those teams tend to thrive because there is a push and pull between those peers versus saying well hey from the top we're I'm doing
this because I'm your manager is why you should do it and so I think that you can get that same sort of effect and remove the extra potential depending on the organization baggage that comes around with managers because you're those are people who are necessarily or maybe only in part
contributing to the actual output of the product versus individual contributors who are all working to build the thing right when you've got lots of layers and those people are not directly involved in building a thing it's just strict overhead thank you for on so we want to transition to
closing statements to give you both one minute of time to to end this I want to break the fourth wall a little bit and just say like this has been so much fun having this conversation with with both of you and I know that both of you have been able played along and had a lot of fun and taken
you know the extreme of your position I think it's just really powerful and looking at the chat and what's going on there I think people have seen this conversation is really refreshing and different and really enriching because you two have really put a lot of critical thought into
into these positions so this has been a ton of fun just breaking the fourth wall but so with that to bring it back in all right I'd like to give the two candidates a one minute for closing statements so for for this one we'll have Jerry you can kick it off with your closing statement
sure as I mentioned at the beginning I think as fun as it is to debate flat versus hierarchical I think the most critical thing is to know you know it's less about which style and to know when to use which style and obviously I think if a team is small then you're able to get away with having
flat organizational style as the team's scale or as the company's scale I think a hierarchical style makes a lot more sense and provides a lot more flexibility more importantly if you have diversity of talent and not everybody is as good as everybody else or to Farhan's point if not you know
everybody is a house engineer or a Shopify engineer or a Google Facebook engineer and you end up in instances where the managers are more capable than the individuals then I think hierarchical organizations actually provide for much more efficiency in that particular regard but even if you
do have a ubiquity of style and everybody has fairly comparable skills I still think that hierarchical organizations a scale in an easier fashion as long as you have effective managers that are open transparent and allow the facilitation and also the transference of ideas up the chain.
Lastly it's very important to have managers be multipliers rather than adding craft into the ecosystem or just basically being dead weight a good manager is a good multiplier and so if they have four people on their team then the net output will be more than four or five
whatever it is a bad manager is essentially just at overhead and in that particular regard yes a hierarchical organization wouldn't be as efficient if the managers within that organization aren't as efficient and so ultimately just comes down to hiring good people whether it's a flat organization or a hierarchical one. Thank you Jerry Farhan you're you're closing remarks.
Cool thanks Jerry I think that an endadre I think that there is a there's a point there that I agree with Jerry which is that you have to choose the right structure based on what you're trying to build at the time what I would say though is I would tend more to start with flat for as long as possible and likely push it much farther than most people would because I do feel that their people automatically move to some other desired state much too early so the idea that hey
after I have 10 engineers I probably need a manager or after I have 50 I probably need director I think is misguided of course people matter not everyone wants to or feels like they want to be capable to or have like a hundred director reports or so I remember talking to the VP engineering
at Twitter they had 80 he had 80 direct reports before it split I think most people should try to push it farther than they are comfortable with because of the velocity increase because of the ability to have everybody working on the product versus people trying to manage the people who
are working on the product and so my default view is go flat try to figure out of course have great people like everybody wants great people easier to say than than to do but to have a flatter structure than a higher archer one whenever that trade-off comes about and then of course as you
have to break it off even for me at 120 I had to break it break the model and start putting in more management but I think it's different for not just situation but different per person or depending on what company you're in and I think we should push it much farther to flatter
versus if you do a survey it's probably 99% higher archival and 1% flat in the world I think it's to be much more the pendulum is to swing much more towards flat for the next little while we figure out what that structure could mean Jerry for on thank you both so much for your sportsmanship
and for having fun with us while diving into such a critical topic I mean there's so many interesting lessons at the extremes of these strategies and I think you two did an incredible job about really pulling out some of the most powerful and interesting considerations and essentials
of of pursuing these different strategies and just appreciated your final remarks about choosing the one that works best for your situation and your strategy I think that's like a very powerful way to end and close the debate now it's up for the world of engineering leaders
to decide so thanks so much Patrick for hosting thank you thank you we'd like to give a special thanks to Mesmer the exclusive accessibility partner of the engineering leadership podcast Mesmer's AI bots automate mobile app accessibility testing to
ensure your app is always accessible to everybody to jumpstart your accessibility and inclusion initiative visit mesmer hq.com forward slash ELC you can also follow the link in our show notes that's mesmer hq.com forward slash ELC if you enjoyed the episode make sure that you click
subscribe if you're listening on apple podcasts or follow if you're listening on Spotify and if you love the show we also have a ton of other ways to stay involved with the engineering leadership community to stay up to date and learn more about all of our upcoming events our peer groups and other programs that are going on head to sfelc.com that's sfelc.com or you can also follow the link in our show notes see you next time on the engineering leadership podcast