The Curse of Success
How do you get a whole company back to work, when being asked to do work is not chill?
We tend to underestimate the impact of power laws on technology companies. For many companies, if you get one or two things right in the early days it could set up you up for success for a long time. The nihilistic take on this is that nothing you do afterward matters; the more optimistic take is that it gives you the runway to try and hit another home run. But what happens if you get carried away with the success your initial momentum gave you? As the economy takes a turn, we’re all about to find out.
Atlassian and Canva are in the news today for asking their employees to do some work. Michael Hutchens had some cheeky commentary on it from an investor perspective.
During a downturn, shitting on tech companies that drown workers in perks and have low standards for productivity is fun and easy.1 If you believe that tech companies should be profitable and not wasteful, then the dab writes itself.2
But too often, people over-simplify the argument. Read Hutchen’s Linkedin screed, and the comments on it, and you come away with the impression that Atlassian and Canva are doomed. Nobody there does any work. Nobody has ever done any work - those companies are zero interest rate phenomenons!
And yet… nobody I know uses Photoshop, while Fedexes around America have fliers showing you how to set up a Canva account. There’s a million bug trackers out there, but most people I know use Jira. Reluctantly (they say), but not so reluctantly that they don’t pay its exorbitant price.
Clearly those companies have done something right. Is this because of their workforce or despite it? Canva user growth since Covid is nearly vertical. Is this because of the work of their 3700 employees, or despite it?
Every founder has, at some point, whispered “we’ve hired more people, but we aren’t moving any faster”. Founders over-exaggerate this a little, as they tend to turn a blind eye to things that other people find important. But it’s not unreasonable to say that each incremental hire will not linearly add the same output.3
Despite this, big tech companies have added lots of people and still produce lots of output. This suggests one of two things to be true:
Either these companies were extraordinarily productive when they were smaller, and early employees had enormous output. They’ve been riding that momentum ever since.
Or they’ve found a way to buck the trend. When they make new hires, output grows linearly (or faster).
If #2 was true, you’d do anything to avoid downsizing and wouldn’t be very worried about individual performance, because it’s so high. And yet, as we enter a recession, many big tech companies everywhere have aggressively cutting headcounts and raising standards. As companies reach a certain scale they focus more on internal productivity, and that helps a little, but I think on aggregate we can agree that #2 is not the norm.
If #1 was true, then it suggests a few more things:
Breakthroughs early in a company’s life are disproportionately important. If a company doesn’t make some significant breakthrough early in its life that generates significant traction and momentum, it’s unlikely they’ll ever get there. This means that past a certain age, you are only left with companies with momentum, and dead companies.
If you find this momentum, there’s a strong argument for redirecting future investment to sales and marketing, so you can capitalise on the momentum you have. This is because breakthrough discoveries are rare and feel random, but incremental sales improvement feels doable.4
Decreasing headcount and raising standards should not have a proportionate impact on the company’s ability to produce output. If the impact of adding the next hire is less as you add more, then the inverse is true too.
This sounds closer to reality.
But all this tells us is that decreasing headcounts or raising standards won’t hurt much. It’s not clear if it will actually help.
The question for every company that is laying people off or managing people out is if this is a path to profitability, or to sustainable growth.
One school of thought is that just by cutting wages by a big enough number, profitability will come automatically. But this implies that everyone who was cut was doing nothing of value (or at least, nothing more valuable than their wage). Maybe this was true at Twitter, but do we really think it’s true across the entire industry?
An alternative answer is that by shrinking the companies they will be able to do things they weren’t previously able to do. For example, by decreasing the amount of bureaucracy required to launch a product, they’ll be able to launch new products that otherwise would never see the light of day.5
If you take this argument to its extremes, you could wind the company back to just its founding team and they could build anything they like. It’s not a given that they’d make money, but they would move fast, right? Maybe. This works if the founders are still dialled in and eager to work as hard as they did on day 1. At some companies this is true, but it’s not universal: Atlassian’s founders seem more interest in buying mansions and making hostile takeovers of utility companies than they do in improving their products.
Extrapolating this back in the other direction is insightful. No matter how much bureaucracy you remove, it won’t actually lead to new products, better services, more innovation, or any other breakthrough, unless everyone still at the company is culturally able and ready to do the work. If everyone is used to 90% of people getting “exceeds expectations”, then we’re talking about a big shift. The HR team telling everyone to work harder isn’t going to make it so.
Lots of big tech companies have great products and incredible user growth. But in many cases these products are old, and the company’s success is based on momentum created a long time ago - not on work being done today. Only a small percentage of the things a company builds end up really mattering, and the challenge during a downturn will be to create an environment where more things that matter can still get done.
I predict that the companies that do the best over the next few years will not be those that cut the most, or even necessarily those that cut the least. It will be those that are most culturally ready to work hard, delay gratification, and create great things.
It’s worth noting that Atlassian was bootstrapped for 8 years before raising any money. Presumably they were profitable - or at least, sustainable - for most of that time. That said, it’s been 13 years since they did this, so it’s plausible they may not have any bootstrapper DNA left.
This trajectory is pretty common, and tends to lead to laziness.
Apparently this was the big cultural change at Google after ChatGPT; suddenly you were allowed to pitch AI projects again. (It’s ironic that Google is used as an example of a highly productive company in the AFR piece. Do most employees there get a lot done, or do they just have momentum from owning the world’s greatest business model for decades?)