"There are going to be a lot more remote-only startups that are going to form with some of the best people all around the world." - Adam Odessky
Today, I'm excited and more than a little nervous to welcome Corey Quinn to the show. To be clear, all of my guests are crazy smart, but Corey is also professionally funny. He's the chief cloud economist for Duckbill Group, a consulting firm that helps companies fix their horrifying Amazon Web Services bill by making it both smaller and easier to understand. It is a complicated and arcane job that Corey makes much more accessible with his frankly hilarious newsletter Last Week in AWS and his podcast Screaming in the Cloud. Corey, thanks so much for coming on the show.
No, thank you for having me, Rachel. I am excited beyond belief to inflict my crappy excuse for a personality upon your audience. Give me a microphone. I'll be there with bells on.
Wonderful. We haven't had bells on the podcast before, but we are open to new experiences. Duckbill has built this amazing business, educating people about AWAs and helping them manage their costs. Can you talk about how the company came together and how it's grown?
A series of escalating poor decisions. We look back at my own career trajectory. I am pretty clearly one of the absolute best in the world, and that is not an exaggeration at getting fired from jobs. I have a number of challenges in the corporate sphere that basically distill down to my personality. So at some point it was clear that when you get fired because you had a crappy boss and that happens five or six times in a row, the only consistent feature in all of those scenarios is you and maybe other people aren't always the problem.
Some people start companies because they have this aspirational vision of a better world, of a problem they know how to fix. In my case, it was, well, lack of a better option. I couldn't really stomach the idea of doing another DevOps job or managing an ops team and the things that always got me in trouble, namely my opinions that every person I'd ever spoken to even casually had said, Oh, that's going to get you in trouble one of these days, either they're all right and this is going to be hilariously bad or instead I can try it and see and find out for sure. Four years in, nobody has hit me with the shut up stick yet.
For better or worse, it seems to be working with about 10 people now. Focusing on consulting was an easy approach because I'm a terrible developer. So why would I write a SAS product? Instead, I'm going to go and talk to people instead. The house bill sounds like a good intersection between a bunch of different disciplines that I have a bit of expertise within, and it's a small problem affecting approximately everybody in the fullness of time. I figured I will give this a shot and we'll pivot in 18 months if something else makes more sense, and it accidentally turned into a bit of a rocket ship and I have no plans to pivot.
It's fantastic. It's one of the things I've always loved about this end of the tech industry where we've built the island for misfit toys, where battle-scarred Unix systems administrators and failed English professors can come together and dink about with shiny lights. It's very satisfying to me.
It absolutely is. I started my career many moons ago as a grumpy Unix systems administrator because there's no other kind of systems administrator. We're all angry and upset and sad about these things all the time because computers are terrible. Software is terrible. Hardware is terrible. The only question is how it's going to break in different ways. Once you've been around the cycle, a few times of seeing today's brand new, shiny, exciting technology become tomorrow's legacy garbage that we're all going to have to support. It starts to rain in your enthusiasm for adopting this awesome code that we wrote last night. So too production with it.
Though, software is eating the world, so there's literally just going to be the gray goo and nothing beyond it. It's going to be Unix systems as far as the eye can see. What are some of the challenges you see your clients facing as they're trying to go from being a normal company to being digitally transformed?
Frankly, there's a big misperception that everyone else is doing it right and we're doing it terribly. No one in the world thinks that they have a lock on this stuff. There's always room for optimization, there's always room for improvement. But very often when you're looking at this from a perspective of, Oh, the cloud bill is way too high, we should do something about that. That is the end result of a game of corporate telephone where the person in finance has received a surprise bill from AWS that's 20 percent higher than last month’s.
Their question is if they're not looking at this the way that you or I would if surprise our rent is now 20 percent higher this month. Instead, it’s “this is interesting and it wasn't planned for, what does this mean for our longer term projections?” but those people never talked to each other. So five layers of corporate telephone later, the person who can actually impact the AWS bill winds up getting a question phrased as, “the bill’s too high!” But if the bill were 20 percent lower, magically, which never happens, they would get the same question asked “What does this mean for our long term planning?” It's not so much about saving money, it's about predictability, it's about understanding it.
There are exceptions in all directions, let's be very clear on this. But that's the most common answer for “what does this mean for the future?” Let's also be very clear, the biggest lie we all tell ourselves is that after these sprint finishes, then we're suddenly all going to start making smart decisions, pay off our technical debt and do everything the right way so we shouldn't make any long term commitments based upon the current state of the art. Those companies are still running mainframes, and there is, at the time of this recording, no AWS 400 for them to migrate to. So, yeah, maybe go ahead and just assume your architecture is going to sustain for the next 12 months.
It's just because we've been in a temporary crisis since 1995.
Exactly. All things are temporary in the fullness of time.
So how do you deal with that? You've got a client coming to you saying, our bill’s 20 percent higher this month. We tried yelling at the guy and the guy's like, I don't know what you want me to do. Where do you start with that situation?
Easy, best question in the world. And this is what I think a lot of the software companies miss is, OK, you have a problem with your AWS bill or you wouldn't be reaching out to talk to me. Why do you care? It sounds a little insulting on some level. Your bills too hide. You're asking me why it matters and well, yes, the internet lost its collective mind a few years ago when Lyft filed their S-1 and disclosed that they were spending in the hundreds of millions of dollars a year range, and I think it was 300 million over a three year time frame. People were saying, for $100 million a year, Lyft could go ahead and build their own data centers. And theoretically, maybe.
But first, people always forget that the bill is dwarfed by your payroll costs for people working on the environment. So it's an optimization and to look at the rest of the S-1. Sure, they're spending $100 million a year-ish on cloud, and they're losing $900 million a year on their business. The game of double or nothing, is pretty telling. Let's say I'm the worst cloud economist in the world, and I double your bill for no actual gain. What does that do for your business trajectory?
Counterpoint. I fly to Seattle and start taking hostages, and I drop your bill to zero. Does that materially fix your business challenges? No, it does not. So you want to optimize it. But the things you're optimizing for, in a scenario that most companies are, is less about optimizing an existing thing and getting it a little bit better and more about a capability story. What can you do faster, more effectively because there's an upper bound of one hundred percent, theoretically, if you're AWS bill that you can save, but you can bring in many times of that if you're looking at it from a revenue perspective by launching the new feature at the right time to the right market, focus on that personal psychology or on money and corporate psychology or on money or diametrically opposed in those respects.
How common is it the situation in which I once found myself where my employer was paying about $150 million a year to ingest Amazon Web Services and make them slightly slower and less reliable?
Generally speaking, if you want to make Amazon Services slower and less reliable, usually the answer is multi cloud, but that's a separate argument entirely. On some level AWS is spectacular at building the building block primitives that you can use VMs, object storage, load balancers, et cetera, et cetera. That's great, but that's not what companies for the most part are doing. If your business model is, we're going to offer object storage by slapping our own white label on top of S3, perhaps consider an alternative business model, but taking the bricks and building the house is what companies are doing. That's something AWS has long struggled with even articulating a story around. That's the thing that winds up being the differentiated aspect of what companies do.
Now are those houses less reliable than the bricks upon which it's built? Yes, definitionally they have to be. That's a common challenge. In isolation, we're just spending one hundred and fifty million dollars a year to make web services worse. Sure, that is on some level and in a cynical perspective, absolutely true. But on the other side of the coin, there is presumably at least that much in revenue that is coming in as an aspect of that, unless you're one of those newfangled VC companies. I come from the old world with the kind of business model our grandparents would have understood. Namely, you make more money providing services and it costs you to provide those services as opposed to. Well, all right, I'm trying to raise $20 million. What if I raised $2 billion? What am I going to do with that? Something monstrous? Then we wind up with, well, we're losing money hand over fist, but that's OK. We're going to make it up in volume. I don't understand those models. I don't pretend to understand those models and here we are.
I share your deep fear of the VC funded business model that is currently burrowing it’s way into San Francisco. There are similar challenges facing brick and mortar companies.
Absolutely, absolutely. No one has this right. Everyone has challenges, everyone has constraints. And I want to be very clear that I have loud, angry opinions in the general case. But my default approach when talking to individual companies about what they're doing is coming from a starting point of you're probably making the right decisions. Now, that's not always true, but with that baseline understanding or expectation that, OK, I see something that looks ridiculous rather than everyone I'm talking to is a fool, which is not a sympathetic position. It's oh, it's probably like this because of context. I'm missing. Can you help me understand why it is this way? And sometimes, well, it turns out We're Fools is the short version of the answer. But other times it's Oh yeah, there's this other constraint you're not aware of. Sometimes that constraint is no longer in effect and provides an opportunity. Other times it's Oh yeah, we can't really turn off that mainframe because it winds up doing $80 million a day in revenue. So we should probably have a plan.
I actually want to dig down into that because I think there's really something interesting here. The kind of shamelessness that you describe. Assuming good intent is one of my favorite things about DevOps, and we often encapsulated on the show as looking for win-win outcomes. And I wonder if that somehow diametrically opposed to VC, which is moving more and more towards this zero sum model like literally in the case of Peter Thiel's book zero to one, do you think there's something about Duckbill never having taken any VC money and having this independence that frees you up to look for those win-win outcomes?
Absolutely. Our entire business model is built upon fixed fee in return for a predefined outcome, and as a result, we have none of the conflicts of interest or perceived conflicts of interest that a lot of folks in this space have. We have no partnerships with any third parties in this context. We have no kickback or referral system or the rest, and someone winds up referring a client to us because we can help them. We thank them profusely or maybe buy them dinner. But that is as far as it goes, it's not a story about, recommending you implement this tool because I get three percent if you do it or I'm taking some fixed percentage of your savings or some fixed percentage or bill, because when you're going in a fixed fee basis, every piece of advice I give you is exactly what I would do in your situation.
Sometimes it very much becomes a you should spend more on this area because you have cut to the point where you are impacting your durability, and that's fine until suddenly it isn't. It's a business risk. Being able to have context behind what's going on is important. Now, if you're looking at this through a VC lens, how do I scale this to the entire world and then find a multibillion dollar opportunity here? I don't think you do. I think that as soon as you try and distill it down to a pure software play, you lose a whole bunch of important context and you start making boneheaded recommendations like, go ahead and reserve some capacity for that cluster you're turning off in two weeks on a one year commitment basis. Or alternately, those are idle instances, shut them down. How about we don't make too many foolhardy recommendations like that, and no one believes a thing you say anymore. That's why we start by asking questions, even for things that you would otherwise assume would be no brainers.
That touches on another, what I would describe as a sort of standard podcast stance that we take, which is that automation is for the 80 percent of your job, which is what humans should be concentrating on, is the 20 percent of stuff that's really hard to automate. That's nuanced and subtle and has a lot of variables involved in it, because that's where humans really outperform computers in analyzing very difficult and very complex terrain.
A question I've sometimes asked is about AWB and its intricacies. Why are you focused on AWS billing instead of architecture? And the answer that I think gets missed a lot is they are one and the same, and I don't believe otherwise. If you're trying to make an assertion in that direction, we're going to have a heck of an argument.
Money becomes a proxy for efficiency.
Exactly. You talk to most folks who are specialists in cost optimization in the official capacity, and the answer is an awful lot like, Oh, we're just going to assume whatever you build is right by some RIs and maybe enable this new service somewhere else or right size your instances, I instead look at what is that thing, why is it doing the thing that it's doing? What are the inputs? What are the outputs and how much engineering work goes into that?
You're taking in a petabyte of data a month in that application and passing it back and forth between availability zones 50 times over. Does that need to happen? If it does, can you store two copies of that data because it's going to be way less money than passing it back and forth forever? There become a whole bunch of different stories. There become a lot of misunderstandings around this. But yeah, it's easy to just consume some stuff API and say, buy some reserve instances, but it doesn't help people get into a better place than they are today.
The money being a proxy for efficiency also shines the lens back on the business, back to our grandfather's business models, where we sell stuff for more than it cost us to build them. So it becomes a. It's not so much about the money as the money being a way to tell you if you're getting the proportions right.
On some level. There is a point where you need to stop focusing on optimizing your bill because there's always going to be opportunities to do that. Left to their own devices, developers love games. We will, in the engineering sense, absolutely be thrilled to sit there and golf a hundred bucks a month off of our developer environment for six weeks. Let me spoil that for you. You cost a heck of a lot more.
Let me also say the uncomfortable truth that all of these software engineering worlds like to conveniently ignore. The reason that you have the wonderful compensation packages you do and the perks that you have and the respect in your company that you do is because you generate more value than you cost. The day that that changes, you're in trouble, and people like to forget that they like to comfortably keep that thought at the periphery, not something that they internalize. I think that's a mistake.
Having come up through journalism, I think ignoring those early warning signals that your industry is no longer profitable is definitely something to watch out for.
I tend to avoid commentary on specific industries, but as far as individual companies go, because it's easier to rationalize that, well, we're losing money every month and that shows no sign of abating and when I look at the financials, it seems depressing and sad, and half of our executive team is left in the past year. The correct response is generally not, I bet this is an awesome boat now that all those rats are gone. Might be time to start considering where else to go.
My buddy Heidi Waterhouse, who I believe you know.
Heidi is wonderful. She was the first guest on my Screaming in the Cloud podcast. Now I'm over two hundred recordings in. She came back recently for another showing and she was one of my favorite guests at the beginning and remains so now
She is the best
I love all of my guests, but I love her slightly more.
The way she formulates it is to understand how you make money for the company and how the company makes money overall and how those two relate. She keeps that as the watchword in all of her career decisions and all of her career advice. And it's golden.
Oh yeah. A lot of what we see too, from a bootstrap perspective, which is the condescending term that these companies use to refer to profitable companies that sell funds, is in the VC model, you wind up raising some insultingly large pile of money. Two hundred million or whatnot, it sits in a bank account more or less and dwindles and. At our current rate of burn. What is our runway? How long do we have before we to either raise more or go out of business?
It turns out that's tremendously unhelpful in a bootstrap context, because the answer is always our runway extends three to six months or so. And if you're a VC backed company, that is panic time in many respects here. That's the nature of doing business that is just fine. And there are other metrics you have to look at, and even referring to it as runway gives people the wrong impression here. On some level, how you treat money as a company fundamentally dictates how you think about your entire business.
It's back to the win-win outcomes. You can't afford to burn customers if you have to go back to them in six months and ask them for a renewal of the contract.
Exactly. And again, we tend to do an awful lot of one off projects. We don't have a model where you start off by paying us and then you continue paying us and you never stop paying us. That's not how we go about doing business. When I was running ops teams and I went out to specialist consultants, one of my key questions was great, at what point are we done with you? When do we declare victory and stop paying you? The answer was, well, five years from now, by the time you've hired a team, it's like, I'm going to go talk to the next candidate.
What do you think makes it so hard for more traditional companies to adapt to all of these changes?
By definition, an awful lot of those companies are fairly large, and companies in terms of process and in terms of culture are an expression of the trauma they have gone through at different times where you wind up with an executive, for example, who has embezzled money and gets fired, prosecuted, et cetera. The natural knee jerk response is, well, we need to wind up implementing a whole bunch of additional controls and a whole bunch of processes so that you can never embezzle money in the same way.
Every time something bad happens, you wind up rolling out an additional process or control and fighting that instinct is compelling. In some cases, you need process and you need control, and in others you just need to fire the jackass that made you think that you did. Misunderstanding that is a delicate balance. Now you look at companies that have been around for decades, they have found something that clearly works. They are still a going concern but a lot of the constraints that shape them and processes that have emerged as a result are remnants of an era that no longer directly maps to what we see today. That is said from a place that is in no way steeped in judgment, but rather from a position of understanding that times change people's people change cultures change.
You and I were joking about how corporate innovators are the least likely people to roll out innovation before we started this show, and it's common and easy to make fun of big corporations trying to innovate. But I find myself drawn back to it for exactly the reason that you articulate. These companies have survival strategies that have outlived their usefulness and there are great people, visionary people inside those companies who are trying to create new survival strategies. I like it because it's very meaningful work. You've got this corporation, which is the livelihood of everyone who depends on it and you've got people who are thinking about the future and trying to plan for it.
There's a cultural myth that we have, that innovation is something that only can happen in garages in San Francisco and is probably not the case. Remember, Bell Labs? Everyone wants to talk about the most recent transformational technology that has changed the face of the world being the iPhone. Apple wasn't a small company when they started development on that project.
Companies absolutely can innovate. Fantastic things come out of large environments, sometimes terrible things, too. I'm not here to say that everything is hunky dory. I'm not waving a flag saying yay, capitalism. But at the same token, it does empower certain forms of innovation that are interesting, that have the power to transform different organizations. And that's great. Startups have their place, but so do established corporate entities. I don't want, for example, things involving food safety or medical technology to wind up coming from four tech bros in a garage who were just given an insultingly large pile of money by VCs and their entire business model is Let's outrun regulation. That doesn't seem like a good thing for anyone.
As you look back over what you've done with Duckbill, what are you proudest of?
Oh, easily the people hands down. We have hired an amazing team and we've had some people leave and they were amazing as well. It comes down fundamentally to who are the people you want to spend your time with. What are you personally bad at? You spend a lot of your time on who you can bring in to do that thing way better than you ever could. It's a constant process of being humbled.
Looking at the way that we work, the way that we gel, the way that we think about these things and live out those beliefs as a corporate entity. Again, we are not a cult family. We are very much a company and treat it as a company. I'm very proud of the people that we have and the culture that we are maintaining.
If you had a do over, what would you do differently?
Oh, so very many things. I would have approached things in a bit more of an intentional way. On the media side of our business. We talk about ourselves as a consultancy, which we are, but we're also a media company in many respects. In the early days, it was one of those, I needed to do something that kind of looks like marketing, so I'll start writing a sarcastic newsletter and send it out to three people who are possibly blood relatives of mine. Now there's twenty six thousand people, and my family is not that big. People actually care what I have to say. It was sort of an accidental happenstance story.
I would have been much more intentional about it. I would have done certain things earlier on, and I also would have not second guessed a lot of things that I did in the early days of, Well, if I did this in my last job, I would have gotten in trouble, so it's probably the wrong thing to do. To be clear, we're talking about things like expressing my personality, not embezzling corporate funds. Let's be very clear on that point. I would have effectively embraced the dare to be different approach. I mean, fundamentally, our philosophy distills down to the cloud industry as a whole takes itself far too seriously, and we aim to change that. I think that realization would have been handy if it had hit sooner.
Something we talk a lot about in the fiction workshops I go to is lean into your weird, and I think it's true really for any kind of creative endeavor.
Everyone is bringing something to the table, whoever they are, wherever their experiences have brought them. And that's great. Embrace that. It's having people with different lived experiences. Bringing those experiences to the table is incredibly valuable. It comes down to listening to people whose experiences do not mirror your own and figuring out how things that you advise customers to do would work super well for some customers, but not others. It's important to wind up practicing shifting perspectives, and I think that that is something that companies miss.
That's another thing I really love about the corporate innovation gig is working with clients far outside Silicon Valley and taking the good things about Silicon Valley and sharing them, but also learning that Germany has this incredibly progressive regulatory system around renewable energy. And they're driving to being carbon neutral in a few years. And it just changes every perspective you have about where innovation happens and what it looks like. It's been really delightful to be humbled in that particular way.
There's a lot to be said for functional regulation that has not been captured by the industries that it is regulating. Regulatory capture is a real thing, and it causes a bunch of problems. But I'm not here to opine on politics. I'm here to opine on bills because at least those I know how to fix it. Turns out this is my big problem with VC culture. If we're being perfectly honest like, well, I bought a winning lottery ticket this one time and it panned out super well. Now I'm qualified to be an expert in telling other people how to win the lottery too, and also an expert on anything that I spend 30 seconds or more reading on Wikipedia. No, stay in your lane.
I've actually been in one of the startup companies where someone who scratched off a winning lottery ticket parachuted into a completely different domain and came in saying, I'm from Silicon Valley and I'm here to help, and sure enough, we took $6 million of investors' money and flew it into a mountain.
How would you distill your experience into lessons for our listeners?
Mostly, as a cautionary tale if I’m being honest. I think that there's a lot you can take from people's stories, but remember you're only ever getting their highlight reel at best. Very few people will go on stage and say, here's how we really screwed the pooch and this was a terrible decision and we're still recovering from it. It turns out when you're publicly traded, you don't get to tell those stories. So everyone has confidence where everyone has those highlight reels.
The way I actually disambiguate this as I look up the project they're talking about on LinkedIn and see how many people are claiming credit for it. If it's a low number, it was a disaster. It's one of those obvious and hindsight moments. Other than that, I would say depending on where you're starting from and where you're going, there's a lot that we got right, either through intention or happenstance.
I think that having a differentiated value proposition is key here. I could have said, Oh yeah, I'm an architect. Great! Spin, a dead cat. You're going to hit 15 people in any given room in San Francisco who can claim the same thing. But by mapping that directly to a very specific, very expensive, very painful business problem that suddenly triggered the Rolodex moments because no one looking at their environment said, You know what we need here? An AWS architect. No, of course not. They're saying, wow, that bill's in the stratosphere and the CFO is about to hit me with a belt. So we should do something about that. When you tell people I fix the AWS bill. The response has fallen to one of three categories either, that’s cool or that's interesting or holy shit, wait, right here, there's someone you need to talk to. It turns out that the right positioning statement can sell the entire enterprise for you.
Yeah, at a 30 percent response rate. Not too bad.
So far, so good. Well, those are the three outcomes. One person in three does not have an AWS bill problem yet. Ask me again in two years.
How do you think the pandemic might affect companies in the longer term?
I think it's hard to say. I think that it will cause more change than people are expecting, but also cause less change than people are expecting. If there's one universal truth, it's that people are terrible at learning from their experiences. People are going to be glad we're out and about no longer locked down and then revert to their old habits. We are seeing inflections right now in the labor market.
As far as remote being a major push. Suddenly you're seeing top dollar offers being extended to people who don't live in major cities, and it's too early to say whether that is a sustained trend or we just have a few examples we can point to and call it anecdotal. But I think that there is going to be a gradual shifting back towards, if not being remote first, at least being remote friendly. But we'll see. Every company is different.
We've started this company fully from day one, and the past year has been very different from the three before it. When you're remote everyone's working from home in a non-pandemic year, great. You wind up flying out to meet customers. You wind up having gatherings and talking to folks, and it becomes much more of a today. We're going to wind up being on site of the customer versus today we're going to be working from home when all these days start to look alike and we wind up finding ourselves sitting at various Zoom meetings all day, every day. That's something very different. And it is not the same.
When you're seeing the stressors that impact this of being trapped at home for lack of a better term, it is stressful in a way that normal times remote work is not. And some folks are absolutely going to conflate the two and think, Oh, remote is awful, because that year was awful. Yeah, this year was awful, but saying we're bad at doing remote because we had to basically instantaneously in a terrible situation where everyone is deeply depressed is not really a valid test of the best that remote has to offer. As a counterpoint, some companies have a strong in-person culture, and unlike most of Twitter, I don't think that's inherently a bad thing. What I suggest is that whatever side of that divide you're on as a company, you approach it with intention, not inertia.
I think we all learned a lot about what works for us temperamentally. I mean, for me, not getting on a plane for a year was great. I have to admit it. I don't like jet lag. I don't like travel. I don't necessarily like being out of my comfort zone. I adapted to remote working much better than I thought I would.
I think that there's a lot that gets overlooked as far as how remote works and the most disastrous experiences that I've had with remote in previous years have been the hybrid model. Where we're going to have a bunch of people in the office and then a couple of people who are full remote, no matter what intentions you go into this with, the remote people are second class citizens. Either everyone's remote or everyone's in person, by and large is the path to success. Trying to split that in both directions leads to disaster.
Corey, how on earth do you avoid burning out?
I had a wonderful conversation about this recently with Dr. Christina Maslach, who has turned industrial burnout studies into her entire career. She's a professor emerita at Stanford, she has been studying this since at least the 80s, and it is a fantastic series of conversations. I think that one of the most common things that gets overlooked across the board is that burnout means you're working too hard. It's not. Burnout has much more to do with your level of engagement. Do you like what you're doing? Because we've all been on projects we love and working ridiculous hours at various points. Not that that's healthy. Not that employers should expect that. Let's be very clear. We loved every minute of it.
There are also times where we're just going to work and phoning it in and burning out there, even though we're basically doing Excel spreadsheet paperwork most of the time. It comes down to, for me at least, I have a bunch of aspects to my personality slash psychology slash psychiatry where something I've learned about how I work effectively and where I remain engaged is I thrive when no two days look too much alike. Where every day is different than the one before it. When too many days start to look like the day before it. It is time for me to change something, and I ignore those warning signs at my own peril.
So you're a seeker of novelty?
Absolutely. So much of what we do as a company starts off as me exploring something, in search of novelty, discovering there's actually something repeatable and useful here. Then passing it off to people who don't have the constant exploratory urge to go down every weird back alley that they come across in the dysfunctional way that I do. Again, hiring people who are great in the ways that you aren’t is a terrific early stage lesson for any company. My business partner and I are almost completely opposed. He loves spreadsheets. I love having a personality. I love you, Mike. You're great.
He and I are very different. Our working styles are radically opposed. We have almost perfectly aligned values, though, which is what really matters when partnering with someone. As a result, we sure the other up where we're individually weak. We were both running independent consultancies before we wound up joining forces, we were solving a lot of the same problems together, and the problems that I was facing were the ones that he was struggling with and vice versa. We're better together than we are independently.
That raises two questions for me. One is, do you think that this novelty seeking is why this specific business model where you're doing one off engagements for a huge range of very different clients is taking off?
Let me be clear, we do have some recurring engagements and I want to be very clear on that. But that was why I started it. We wind up going in solving a problem and leaving because I wanted to seek after novelty. I'm no longer directly involved in most of our consulting work. I have people who are better at a lot of these things than I am, and they're much better at these things and doing ongoing support and ongoing engagements.
There are some that are terrific. But in many cases, the right answer, as it turns out, is go in, do an assessment and deliver an advisory opinion on a lot of these things and then leave. You're not going to add much more value by hanging out. And again, we're fixed fee. We're not billing by the hour. Once we're done, we're done. Let's move on to the next. There's a very large ecosystem of folks who have big bills or will have big bills in the near future.
I wanted to dig into what you said about yours and Mike's shared values. What are those values and how did you have those conversations where you established that you were a good fit?
It's hard to say what values are in a direct sense, but I can talk about ways that they manifest. For example, when we wound up joining forces, we wound up effectively agreeing even without a real conversation about this, how compensation was going to work. For example, we have a 401K with a six percent match at our company and invests instantly. It is not one of those long term things. We're not going to golden handcuff you so you stick around for a period of time. We want to increase people's mobility, not reduce it. It shouldn't be one of those. I hate my job, but I'm on the hook to get some more money if I stick around for another two months. No, if you're not happy, go somewhere else. We'll help you do it. That's one example.
There's also a philosophy we have where if we don't respect a company's business, we should not be doing business with them. It is why, on the media side, I do not have sponsors that I do not look at and believe truly that if you're in a particular situation with a particular problem, this is the right answer. If there are some that no matter how you look at it, there's no way to responsibly use this company to solve a problem. They shouldn't be sponsoring this stuff and we shouldn't be advertising it to folks. I don't care about the money because you can always make more money. You can't repair a damaged reputation. Not having to explain or justify those decisions to one another or doing the softly softly approach. Well, we might need to fire a customer because it isn't going well. Do I have to justify this or find a bunch of backfill? No, it's expected that these things will happen.
Speaking of the newsletter, what is the best way for our listeners to connect or to follow your work?
LastweekinAWS.com rounds up a whole bunch of Amazon news, discards the boring stuff and then makes fun of what's left because I have deep seated personality problems masquerading as my brand. We have the Screaming in the Cloud podcast and of course, aggressively posting on Twitter at Quinny Pig. If you like salty opinions, you will love my Twitter feed until it touches on something that you love.
What does the future look like for you personally?
I think on some level that I am here for the foreseeable future, and that is a weird feeling. Before this, I never stayed anywhere past the two year mark. Now we've been four in change and I love what I'm doing because again, all of the problems I had at previous jobs, I am empowered to fix here, and that's no small feat. But it's also a matter of how do I make this resonate more? How do I wind up seeing where the industry is going and aligning? So many companies see a problem that is inherently transient. They focus everything they're doing on the current day expression of that problem. Then when the industry slash ecosystem evolves beyond it, they're effectively left holding the bag with no real plan. How to evolve to meet the new day.
Everything changes, and a lot of things also don't change, and understanding the difference between those two is critical. There is a philosophy here of making hay while the sun shines. A hypothetical example, no one puts things in AWS anymore. They're in decline and Azure is ascendant. I'm a specialist with AWS, but past a certain point when that ecosystem starts drying up for a variety of reasons. Am I going to sit here and put my head in the sand and pretend it's not? Or am I going to start embracing what customers are doing and going where they are? It's really not a difficult question to wind up having to answer right now. You obviously, you solve problems customers have. When those problems start changing and the conversations change, it's time for some soul searching.
As you look at our industry over the next five years, everything pans out the way you think it should. Everybody takes your advice. Everybody does what you think is the right thing to do. What do things look like in five years?
I think people start contextualizing what a cloud provider is and what a cloud provider isn't radically differently. Because the early stages of moving into the cloud, you treat it like an extension of your data center, where it's a bunch of VMs that talk to each other and that's about it. If their problem is, is that your staff keep tripping over the power cables and you forget to replace hard drives, and yeah, that is a net positive, but it can be so much more than that.
You can start using differentiated services higher up the stack because if you're not, your competitors certainly are. Maybe you don't need to build your own load balancers on top of a bunch of VM. Maybe you could use their managed load balancing service in some cases, or maybe not. It depends on you, but sitting there building your own load balancer unless you're five is probably not the right answer for most companies. So instead of focusing on the undifferentiated work that everyone else does, focus on the things that make your business unique and consider passing the rest of it to folks who have made that their differentiation point.
It's the boss level of leaning into a weird. Figure out the thing that you bring to market that no one else can.
Exactly. This gets into Wardley mapping. Simon Wardley loves talking about this stuff. What's become commoditized, what's differentiated at a high level? What is the thing that you're best at? In my case, it's pretty easy to answer some of those questions. I have clearly at this point, I don't think it's much hyperbole to say that I am the undisputed best in the world at AWS bill understanding and optimization, but I'm not in any way, shape or form, even marginally competent at HR. So we wind up outsourcing to an HR company because that is not something that we need to do in house as one of our core competencies.
Making decisions like that across the board is critical. One area that I think a lot of folks get stuck on is they find that the technology that they work on inherently becomes a core component of their identity. I'm as guilty of this as anyone. I started my career running large scale email systems on Unix. It became pretty clear that that was not a growth industry. People started trusting Gmail a lot more, and it was clear that every company was not going to have its own mail server and house and need admins to run those mail servers. So it was time to either defiantly yell at the tide as it's coming in or find something else to specialize in.
I went from working on large scale email systems to configuration management to cloud at this point. I don't know what's next, but I have a sneaking suspicion that I'm not done yet. Cloud is the end and now the next 40 years are going to look just like this. I don't buy it. We're going to find out what those shifts are by paying attention. Spoiler it is not Kubernetes, but that's OK. I'll fight those fights later.
Corey, is there anything else I should have asked you?
That depends on you. One of the fundamental mistakes that people make is they ask me questions they don't really want the answers to. They just want me to agree with them. Honestly, I view this from a perspective of, I don't take the positions that I do just to argue with people. I do it because I believe there's something that's being overlooked. I do speak the general case most of the time, consulting engagements are very different. That is what specific looks like. On the whole, I think that there's a lot of stuff that people are doing just because they think everyone else is doing them too.
Again, high school was entirely about being a popularity contest, and that is something people never outgrow. The industry is just like that only instead of voting people for Homecoming King, we wind up putting them on the Gartner Magic Quadrant instead, and that reassures us that we're doing the right thing. Maybe, maybe not.
I think academia has a lot of problems, but I do like the core idea that if you get a doctorate, it should be because you've pushed the boundaries of human knowledge somewhere, however small. Because I think that kind of applies to everyone in this economy. As we become more and more a knowledge economy, it is literally figuring out the thing that you can do that other people can't do and make that your career.
Absolutely. And look for the between spaces, try different things. I am not the best software engineer in the world. If you believe that you are. You probably work at Google because those people tend to. I am also not the best financial analysis person that you're going to meet, but if you combine the two of those things, it becomes a much narrower niche. There are fewer people there, and suddenly it is much easier for me to differentiate myself in that universe.
I think the same could be said for most folks. Figure out what eclectic group of skills you have that can be combined to form something resembling a superpower and then double down on that on some level. It's way easier said than done. Let's be clear on that. Instead of trying to do what everyone else is doing on some level, figure out what have your experiences led you to that empowers you to see things a little bit differently. Maybe there's something there. And if people are curious about this sort of thing and want to get more personalized advice, please reach out to me. I'm easy to find. I love having conversations like this.
It's fascinating because so much of the work that I do with corporations in the accelerator programs we run is literally sitting people down and saying, What is it that you uniquely can do? How can you build something that is a fundamentally unique selling proposition?
We're going to build a service that is exactly like one of AWS’ core services, and we're going to compete on having a better interface and being less expensive,
But it'll be a bit slower and not quite as reliable.
They will run you. If you can be one feature release away from not having a business model anymore, it's not a very defensible position. People say, Oh, Amazon’s product strategy is Yes. Because it is, and they're going to outcompete us. No, they're not. They're terrible at a lot of those things. You're not competing against the full might of Amazon. You're competing against the two pizza team run by some overworked product manager. It is easy to wind up out thinking and outmaneuvering the large corporate behemoths of our industry, but it takes a little bit of thought of, What if Amazon moves into your market is a great question when you're talking to investors, and if the answer is is what we could do a price cut, you have problems. The right answer to that is you're right.
Take mine. For example, what am I going to do if Amazon decides suddenly to cohesively talk between hundreds of different product teams simultaneously and then communicate with customers in a unified voice that takes all those perspectives in person? We're in trouble and their responses. Great answer. Here's another $500 million. Have fun because it's never going to happen. The thing that makes you rock also makes you suck. And Amazon's microservice org chart is one of the things that empowers them to do wonderful things and makes baseline things like billing, the console, customer support, et cetera, et cetera, et cetera, kind of sucked across the board. I love the people who work there, but they have challenges that are direct results of doing the things that got them where they are. There are no easy answers here. What tradeoffs are you willing to make as a company and everything is a trade off?
Automation is for computers and complicated, subtle trade offs are the work of human beings.
Corey, it's been an absolute joy to have you on the show. Thank you so much for your time.
"There are going to be a lot more remote-only startups that are going to form with some of the best people all around the world." - Adam Odessky
"Organizations and individuals being more entrepreneurial will help us get to better solutions on climate change, climate justice, diversity, equity, inclusion and taking advantage of the wisdom of more voices." - Diana Joseph
"If you really want to think about innovation and technology, start by thinking about what it is that you want to bring into the world." - Brad Henrickson