It's amazing how much better we Software folks have it when it comes to starting a business.
Last I checked, the profit margins for my single-player SaaS business were somewhere around 90%. And the interesting thing about that isn't even the number. It's that SaaS is so profitable that you don't even have to calculate your margins. To an order-of magnitude, every dollar a customer pays for the service can be considered profit.
Real Businesses have expensive office or retail space. We have "wherever we happen to be living at the moment".
Real Businesses have employee salaries. We have an industry where a single person can plausibly run every aspect of the business from writing the code to marketing to racking servers to high-touch Enterprise sales. That single "employee" can have his "salary" set to (Total Profit) / 1.
Real Businesses have equipment and other recurring costs. We have those too, but they're tiny compared to other types of business. Like, single-digit-thousands per year tiny. All in, for servers, software, dev hardware, etc.
It's almost unfair, how Software wins in pretty much every category against pretty much everything else.
This is the back side of much lower success rates. Somewhere like 10-20% of offline businesses survive and thrive for decades, for online startups success rate is much lower. Say, i am 38 and have been coding, and had a large network since i was 18, and i have never seen a person who 'hit it' - even in the most limited sense, meaning consistently made more from his own product(s) than he could do on the Upwork. I mean, i know a ton of such people - some of them make up to $2M a month - but i know them since after, and _because_ they made it (so it can't be used to estimate success rate). Never someone who i knew _before_. And nearly everyone of my coding friends tried. Most tried seriously - the more seriously they tried, the more they lost actually, sometimes being forced to leave the profession completely for the lack of valuable skills (spent 5 years working on own startup with a limited stack and didn't catch up with what the freelancers code with), and need to make money.
Online startups are infinitely scalable, meaning something which is just barely worse than the best probably doesn't survive (see 90%/9%/1% rule), but if it does, it hits big. Neither is true for a restaurant.
It's sort of apples and oranges though, because it's much harder to liquidate and reuse assets from a brick and mortar than it is a software concern.
Every thing I've ever built, in a very tangible way, becomes a stepping stone for me with software. Be it businesses I create or people I work for. Also I can start multiple concerns in parallel and traditional business cant.
Most people, even if they are good at something, cannot start a business of it. That still includes software. There are so many things about running a business that don't directly reflect being good or not. That's excluding people that just aren't any good.
Wanting to be the owner of a business you designed yourself is a crazy persons dream. Software seems to make it more accessible because you can fuck off and get to death much quicker without the same cost, but I bet if you averaged it, what comes out in the wash is the same. A balance of having enough delusion, skill, stickiness, and luck.
I've got a mentor who subletted me some office space for a small thing I had once that I needed a few desks for. Recently he told me how close he was at the time to losing everything and us being there was a desperately needed life raft. Until he told me that, I always thought he was doing me a favor. His business is now at the top of his field. I still can't see the difference. Everything is bullshit until it's real and nothing that is real is permanent. Is success a marker of an IPO or a thing that allows you to take your children on neat vacations and affords time to paint? People only call it a failure when you exclaim you quit or someone else certifies that you've been removed.
This is so true. In the world of online businesses, it does often seem like "winner takes all". This might be because online business can compete globally, so the "winner" can literally take over the world (except China) like facebook did. A restaurant cannot do the same - the closest would be a big franchise like McDonalds.
Fixed costs vs. variable. Man it's insane. I was just talking to a friend about what it takes to run his high end construction company. He does very well and has 20 some odd years in the biz. The amount of work and plate spining. Literally having to take physical possession of sometimes hundreds of thousands of dollars of materials and move and assemble them. Even with insurance, an accident quickly followed by a second could destroy the whole business because of a premium hike.
I can test ideas with zero risk and inventory and make that a mediation to build out my business. Pivoting, changing, blending operations. I'm not as successful as he is by any means, but I can't imagine myself having to shoulder that burden to get where I want to go.
Dude maintains a fleet of trucks. Physically, like with the wrenches.
Do you think the insurance risk to your friend's company would be mitigated by automating parts of the labor? I've always wondered where a construction company would stand on this topic. On the one hand, there's a fear of replacing the human labor, especially when you have close relationships with your employees and they have a specialized skillset or trade. On the other hand, it seems like workers compensation and accident insurance are really substantial costs and appreciable risks for these companies.
The amount and skill that exists in high end construction is crazy. Not just in the individuals but also how they work as a team. Also how few people are qualified and want to do the work. Half of his team are undocumented Mexicans that have been with him for over a decade. I won't even get into the risks and costs associated with that.
The work is hard, physical, requires precision and attention to all sorts of details while being creative and getting up really early in the morning.
Just to give you an example of how not automatable it is. $37,000,000 home that the entire building is clad in a bespoke red aluminum paneling that the architect designed. It doesn't fit in a repeatable pattern and each panel is between 3x4 and 12x7 feet. They need to be hung with a tolerance of less than an eighth of an inch on a cantilevered split level structure with some odd pitched faces. Any drift in the assembly will mean a later piece will not fit correctly and you might not find out until the end. If say a single piece got maybe dropped or run over by a work truck it will have to be custom replaced at a huge cost because the panels were made in batch, also this would effect the overall timeline of completion. The contracts you take with a job like this sometimes come with bonuses for finishing early and penalties for delays. The process to even get to build them is highly competitive as well.
Just to get this on the wall involves coordination to make sure everything is true and to spec as it gets to the skinning. Every variance effects your adjustments for hanging the clad. Sometimes with your own guys involved in the internals and sometimes you depend on workers you have nearly no control over. Either way you have to deal with and adjust.
All of this has to be done without trampling the feeder roots just under the soil of the 250 year old oak tree in the middle of the job site because it's irreplaceable and a couple desire paths can kill it inside a year after the job is over, cutting off 80% of the work area you have access to.
I'm not saying you can't automate a job like this, but we are so far away from being able to. It is possibly on an infinite horizon in skilled human labor.
I would agree that the finish work, cladding, drywall, etc are definitely the hardest part of the construction automation problem. And maybe they never get solved entirely, especially things like the red aluminum panel example. In some sense, the uniqueness of the project is central to its value as an architectural statement. And uniqueness doesn't jibe well with our current approach to robotic automation.
There are a few intermediate tasks that may be doable with autonomous machines. Framing and concrete come to mind because they are tasks where lots of heavy stuff has to go to the right place in a repetitive arrangement.
So maybe a highly bespoke job like a multi-million-dollar house doesn't have any tractable tasks for automation. But a skyscraper might. Or at least, I'd like to think so.
Thanks for taking the time to share your perspective and awareness of the topic with such a substantive and vivid example.
I think there's definitely a mint to be made in making the hardest of it easier. I think that's the only way we keep building higher and more interesting.
If you make that tool though, it just becomes another hammer and the guys like my friend will still sit at the same risk level to use the new hammer in ways people haven't seen. He definitely wants your automation, but only so they can increase dynamism. I suspect that his job will always hold a static high risk and I find that kind of awe inspiring as a bit jockey.
Nothing about his work force is "illegal", they are undocumented. Anyone who works in the field also works with a considerable amount of people in a similar position in the same roles.
The "house" I described is an amalgamation of several.
I guess I would say to that "bored" person, go fuck yourself. There are plenty of other people to track down, harass, and get your target completely wrong on the internet. /b/ is thattaway.
This is the thing that so boggles everyone in "normal" business. They can't understand how I make any financial plans or decisions like investing resources in new capabilities or hiring consultants. I've got stable, long-term customers on annual maintenance contracts. Draw down substantially slower than it fills up. Done.
It's trivially easy IFF you can control yourself. I want to stay in business, so I do.
Bookkeeping is a similar non-event. Pay for everything on one separate credit card and blip the CSV of the annual statement to my accountant. Done.
Edit > Reading through the comments here reminds me of the hyper-importance of finding a viable niche, then occupying, preserving and defending it (_barriers to entry_), plus the high value of: a) "good" versus "wacko" clientele, b) loyal repeat customers that you can move to auto-renewing maintenance contracts, and d) frugality.
But, unless you found a niche or have a really strong brand, ten other competitors can pop up at any time, using the same cloud infrastructure you do and undercut your prices. The worst case is that GoogleMicrosoftOracleAdobe will just put a small team together and outcompete you in no-time. It's almost unfair, how Software wins...
This seemed like a plausible fear when I first started 10 years ago. People have indeed cloned both of my rent-paying products over the years (possibly helped by the fact that I've released step-by-step instructions for building one of them).
But then, as you say, software is easy to write. Businesses are less easy to duplicate.
Is the kid who saw your Show HN and cloned your thing "in a weekend" going to stick out the years of work it will take to build a business off of it? Is Google really going to decide to focus on executing in the "Hook Twilio into a Calendar Scheduler and have it automatically robocall people so they remember to be home when the plumber arrives" space so that they can smoosh your little $10k MRR business?
I hear your fear expressed a lot from people cautioning against building software businesses. But I'm skeptical that we should really be scared of it.
Some MBAs caution against it because they BUY software businesses or programming services. If you know how to write software you can start them and fail quick with so little risk.
I have worked on about 20 projects between $100,000 and $3,000,000 where the primary stakeholders felt that writing a check was enough to run a software company. Like somehow buying bespoke software creation was like buying a used Honda and the key will just start it later if you ever get it out of the garage. A buddy of mine describes the trap as "It's easy to imagine, so I imagine it's easy."
By contrast I've seen non-technical guys with $30,000 take an interest in the work and not just listen to the creators, but inspire creativity, turning that to millions.
I've seen some developers deploy a beloved idea how like OP sold $20 chicken and are surprised when no one gives a shit. If you polish your turd like Duke Nukem Forever or Chinese Democracy, you might not get the response you want and it misses the utility we have in software to engage quickly while correcting course to get the thing on it's feet.
The comment I remember from Joel Spolsky when he released CoPilot(?), the product built by interns over the summer was that he would be thrilled if it made $500,000 over its lifetime. But if it was built by Microsoft and made $500k, they would shut it down in an instant. Even if it made $5M, they'd probably still shut it down since it wouldn't be worth the time.
You worry about Google, Microsoft, etc when you're building a $50M business. Below that, it won't probably even register on their radar.
One advantage of a legacy software / SaaS business is vendor lock-in and the conveniency hurdle. So it is easy to retain existing customers. And if you are the first in the market, you might have enough customers so that your business remains profitable for a long time, even when your growth is stalling because of new competition.
But I don't want to be pessimistic. Because there is no excuse for not having tried to build that business.
I write code for Arduinos in my spare time. There's no shortage of online competition. But yet, even though clients keep reminding me that I cost more than everyone else, they still keep sending me work...
The same things that make it great to be in the software business make it extra-difficult. All of your competitors have the same advantages you do, and pretty-much anyone in the world can become a competitor to your business!
One advantage that local businesses have is a proximity lock-in. A restaurant in Portland just needs to be good enough relative to other restaurants in its neightborhood to survive. A SaaS headquartered in Portland needs to be globally good enough.
Personally I think a skilled plumber/electrician/tradesman would have an easier time starting a successful business than a software developer would. It's hard to generate demand for software, but those skilled trades are things that everyone definitely needs.
I think a skilled plumber/electrician/tradesman would have an easier time starting a successful business than a software developer would. It's hard to generate demand for software
Me, I'd think I was going about it wrong if that turned out to be the case.
A very, very large flip side to that, is there are extremely few winners in any given segment in software / online services.
There are 150,000 convenience stores in the US.
How many successful online auction sites are there? At least dozens!
How many CRM companies that successfully compete with Oracle and Salesforce? At least several!
How many successful competitors to Slack? A couple!
How many successful competitors to Github? At least a few!
Your odds of building just a million dollar sales business in online services or software, is extraordinary small. And if you're in a successful niche, your niche is either likely to disappear entirely in a short amount of time, or otherwise be consumed into a larger company's product.
Meanwhile every single McDonald's ($2.5m avg) location does over a million in sales. Most Burger King ($1.3m avg) and Wendy's locations do as well. That's 27,000 franchise locations in the US just between those three.
Zoom in a bit and things look better. You don't need a million dollars a year to raise a family on a single-person software business with 90% margins.
I know of at least four successful businesses that started as S3stat clones. No idea why they chose my particular niche either, as there are tons of better things to build.
You don't need to build a Slack. You can live quite nicely on the profits of "Basecamp for X", for hundreds of as-yet unaddressed X's.
- Developing can take a lot of time, especially in a market where a lot of players are already active and have had years to come where they are (time unpaid, without guarantees), and
- When growing past what one person can do, your margins will decrease, because you'll need office space and employees.
"the profit margins for my single-player SaaS business were somewhere around 90%. "
Thos are not 'profit margins' - not even remotely.
Maybe you mean 'gross margins'.
But even then, software is an inherently IP-based business, it's about as appropriate as saying: "Hey, this $25 book only costs $1 to make".
That 'software' is open to anyone, anywhere, means it's just that much more competitive, and margins are not 90%, on average they are probably negative.
Twitter has yet to make a dime. They are a few billion in the hole still. I think the same with Square.
Has Uber turned a dime in profit yet?
It's only a very hardened and entrenched few that get into solid, positive margin space.
It's still Income minus Expense, right? For a single founder, it's exactly as though the $25 book costs $1 to make, because you are the printing press, binder, paper supplier, ink supplier, distributor and author.
Yes, you can arbitrarily drop that number by having the business pay you a "salary". But the end result is the same. The money going out the door (to people other than you) is a tiny fraction of that going in.
The 90% figure refers to 'gross margins' and really means 'income minus cost of goods sold'.
In software, usually 'gross margins' are meaningless - it's a term used in most other industries and from classical economics wherein most of costs were actually COGS.
"Yes, you can arbitrarily drop that number by having the business pay you a "salary"."
No, it's not 'arbitrary' at all salaries and operating costs are the relevant costs to the business.
Software (and books) are IP businesses - the 'cost to make the book' or to 'run a website' are not useful in terms of understanding the business or profitability.
FYI - for most 'software businesses', net margins (i.e. after 'salaries and operating costs') are negative.
... so if salary is zero, gross margin and profit margin are the same. Which is the case for a single person business.
I understand the point you're trying to make, but it's still a bit silly to suggest that my business is not profitable because I may eventually spend some of the money it makes.
one interesting thing is how people are trying to operate a business and hire local people to do things are assaulted by taxes out the wazoo, and software companies like apple hide their profits overseas and avoid taxes almost entirely.
Yes, well, fortunately, the Important people in this story (The landlord) have plenty of tax loopholes they can take advantage of - which they will hold over the tenant's neck when it comes time to renew their lease.
Our tax and financial system has incredible benefits for landowners, at the expense of everyone who actually does useful work.
I'm not sure a comparison between a multi national and a local business is apt. Granted they could probably be treated more similarly but multinational software and multinational hard goods seem to get similar tax treatment...No?
Last I checked, the profit margins for my single-player SaaS business were somewhere around 90%. And the interesting thing about that isn't even the number. It's that SaaS is so profitable that you don't even have to calculate your margins. To an order-of magnitude, every dollar a customer pays for the service can be considered profit.
Real Businesses have expensive office or retail space. We have "wherever we happen to be living at the moment".
Real Businesses have employee salaries. We have an industry where a single person can plausibly run every aspect of the business from writing the code to marketing to racking servers to high-touch Enterprise sales. That single "employee" can have his "salary" set to (Total Profit) / 1.
Real Businesses have equipment and other recurring costs. We have those too, but they're tiny compared to other types of business. Like, single-digit-thousands per year tiny. All in, for servers, software, dev hardware, etc.
It's almost unfair, how Software wins in pretty much every category against pretty much everything else.
Nobody tell anybody!