The Engineering Talent Bubble
blog.jayteebee.org
blog.jayteebee.org
A lot of people talk about top engineers being 10x (or even 100x) as productive as mediocre engineers. Assuming that's true (and I suspect it is, but admittedly I don't have any evidence to back that up):
Historically, those 10x engineers got paid maybe 2-3x.
Because it's become so much easier to start software companies recently (thanks in no small part to folks like Y Combinator, but largely due to cheap servers and FOSS), those 10x engineers can now easily go seek compensation that's more proportional to the value they actually create.
In turn, large software companies are finding they need to pay more to stay competitive with the alternative of striking out on your own.
What we're seeing here is pay becoming more proportional to value creation. Which I think is a very good thing.
As an aside, I think the rise in talent acquisitions is largely a backdoor for companies to hire top talent without needing to change their deeply entrenched pay-scales. Apparently Google is starting to figure out that it's more efficient to just pay people more instead of having them leave for a couple of years in order to take advantage of this backdoor.
People have been bandying the 10x-100x saying for years, ignoring the fact that perceived productivity differences between employees are hugely subjective, and nearly impossible to quantify, in any case.
As a constructive proof of this, consider that borderline prima-donna personality type we know and love: management (or those simply jealous of his stock options or reputation, but who don't know him well) might consider him a "rainmaker", the big kahuna, etc -- the classic "10x" employee. For a while, at least.
Whereas those who have worked under (or with) that person for a while might have a substantially different evaluation (more like -10x). Either because they've deal with his morale-squashing abrasiveness... or because they've seen his handiwork up close. Enough to detect, let's just say, a bit of a disconnect between their material output, and the superhuman gravitas projected onto them.
Also, while it may be arguably more true for tech that intrinsic employee values greater than other professions, there's not much basis for believing -- as many hackers seem to like to believe -- that this is something uniquely, or nearly uniquely true for tech people.
There are all kinds of professions (not even professions that have outlandishly high salaries) where it could easily be said that certain employees contribute greatly more than others -- bartenders, chefs, and (though some may laugh here) schoolteachers come immediately to mind.
Yet even though competition is sometimes quite fierce for these star performers[1], no one in those professions is in the habit of saying "so-and-so is 10x, 100x more productive than your typical bartender/chef."
So to a large extent, I think the whole 10x/100x thing has a lot to do with the current zeitgeist -- and the fact that a lot of tech people, frankly, have incredibly inflated egos, and easily imagine themselves in that rarefied 10x/100x category.
And not so much with the incredible, intrinsic business value of high-performing engineers.
I mean realy now: 10x? Okay, maybe. 100x? Come on.
[1] well, maybe not for schoolteachers, unfortunately.http://baseballanalysts.com/archives/2010/11/mlb_salaries_ov...
When you look at the linear graph it looks less like a bubble and more like settling in at a new level.
You can't have a bubble in something you can't resell.
This isn't an argument that salaries are at the correct long term level. I'm just saying that salaries can't have a bubble that "pops" and collapses like asset bubbles do.
A bubble is when you're paying for something based purely on speculation about the future value of that thing, rather than on a rational analysis of the expected future return.
I would wager that Google does not have an internal projection of the value of any engineer showing them to be worth $3.5m over four years; rather, I'd bet that comp decision was made based on an irrational fear of yet another defection from Google to Facebook.
If we were talking tulips, if you purchase at 100 florins, and a year later the bulbs are worth 400 florins in the market, you've made 300 florins profit.
With an employee, if you hire them at a salary of $100k, and a year later, they are worth $400k in the market, you do not receive any of that difference.
edit: A bubble in assets that can be resold can happen because it rational on the small scale to buy into the bubble. Everyone may know it will pop eventually, but people think they can make money and then get out. A bubble in unsellable things like employees can't happen because no one can expect that he will be able to "resell" an employee at a higher price.
When people see a valuation in the marketplace (eg, a talent acquisition that makes everyone extremely wealthy) and assume they can achieve the same kind of returns with minimal effort, that is the hallmark of a bubble. The same thing happened with day traders and real estate speculation.
From what I could find on the bls.gov site, in 2008 there were about 900,000 jobs in software engineering and 75,000 jobs in computer engineering. From what I can tell from searching, Google has 20,000 employees, but not all of them are necessarily engineers.
Even if all google employees were engineers, and they were going out and trying to create startups in order to be bought, that would be 2.0% of the US-wide engineer pool.
Bubbles end badly because of leveraged bets on overvalued assets.
Call me insensitive, but this doesn't strike me as so bad:
"they will find themselves unemployed and without savings."
I was with him until the last paragraph. I think a lot of really bright engineers who accept the premise are banking a large percentage of their salary, no?
99.9% of engineers get paid an upper middle class salary, with a few more getting these 7 figure bumps from acquisitions that aren't really.
I don't think something like 1/10 of 1% of engineers having something happen to them is something to worry about ending.
Or Ev Williams who actually created Twitter?
Or Dennis Crowley who created foursquare?
There is always a shortage for great engineers and they generally have great jobs, so acquiring the top tier will always be competitive as in any other industry. For most companies, the only way to gain access to this tier continues to be personal networks or if they are in the founding team. For other organizations, growing this talent from promise is the most pragmatic approach assuming have some really good talent to mentor them.
The middle and lower tiers are being paid upper middle class wages commensurate with education, experience, and economic contribution.
Things are more competitive in hot geographies like Silicon Valley and Seattle, but that is nothing new.
So called bubbles always exist - they're hot sectors - and then there's a self correction of varying severity.
Seven-figure salaries for developers? Are you sure you put the decimal point in the right spot and counted the digits correctly? I find it very difficult to believe that any developer (not a manager, not a c-suite executive, etc.) is making in excess of $1,000,000. Name one. Otherwise, I call bullshit.