It's scope that increases headcount, not users.
WhatsApp could scale to 900M users because the app they build didn't do a lot. It was for sharing plaintext messages between groups of people. That's quite hard at scale but you can solve it with 50 people as they demonstrated. When they started adding more features, that's when they needed more people.
If you keep the scope of your app small you don't need many people to build it.
100% And adding a way to accept payments (which WhatsApp never really did) and have functionality for businesses is an increase in scope. So, 50 engineers is terribly understaffed for an economically sustainable business.
For example, Facebook, Twitter, Uber, etc. have hundreds of engineers who literally just work on how the company gets paid. And no, this isn't just "well they can just insert Stripe.js into their app" - it's way more complicated than that.
if they hadn't sold to FB, FB or another company would have run them out of business by doing the same loss-leading advertising and subsidies for their own product that they did for WA.
free is not a guaranteed winner if you have three trillion dollar companies paying people to use their competitive products and they control the platforms that your app runs on.
Architecture increases headcount.
Bad quality of product and operations increase headcount.
Skill level of employees increase headcount.
It’s very similar to a proper algorithm cs some naive way of solving things. You can throw a few million on hardware to it, and still only get mediocre results, or solve something in a way that actually works
You’d think that teams closer to the companies core would be more competent and higher value… but often you find groups of thirty pushing inconsequential metrics. Alternately, you’ll find small groups on the fringe that appear to be doing nothing, but occasionally spit out the next huge product.
For these reasons, investors have historically penalized large conglomerates for inefficiency. The growth of big tech through the ‘10s may have been due in part to negative interest rates which valued discounted future free cash flow at approximately infinite dollars.
In a high interest rate environment, one would expect cash flow to be redirected to investors who will then invest the money. If big tech executive pay drops, then there is limited reason for tech executives not to move into startups.
Enterprise sales, personalized support, marketing programs, etc. Oh, and you probably have a legal team now. And you need to enable those sales reps and manage them. Oh, there's HR now too. And a docs team becomes more important...
That kind of transformation only comes from scope or the things the GP pointed.
It adds up. Sure they probably have fat to trim, but twitter gives us a great case study how often the massive headcount is focused on revenue generation activities the user isn’t privy to.
Does it? Twitter the product has been fine (barring clumsy design changes driven by Musk) after the massive downsizing, and they still make billions. I was under the impression a lot of the lost revenue was because of Musk's caustic image that made brands pull their campaigns from the platform.
Long story short they were bought by some holding company that completely neglected the platform and stopped any kind of moderation, DMCA notice or anything else. The result? It became a huge magnet for trading child porn.
While listening to this podcast I thought of how Elon gutted twitter to the bone and probably fired entire teams designed to deal with this kind of content…
Not that companies never have bloat, but it's obvious from your comment you don't know what DropBox actually does these days. Have you seen the admin pane for enterprise users? There is a huge amount there alone that goes into permissions and role management, auditing, compliance requirements, etc.
One simple example - how many engineers do you think it took to get all the compliance frameworks (e.g. SOC II, HIPAA, etc.) implemented and maintained? My guess is that that is one of a hundred requirements you didn't even consider.
I don't think any sane person would argue that DropBox going into the enterprise was a bad business decision.
In the small company I may something like "flip this windows setting and reboot the server". They do, we test it, and the call is done.
In enterprise the different levels of complexity spiral out of control very quickly. Is it the standard windows setting issue, or is it the enterprise endpoint protection causing the problem? Ok, lets follow the windows setting issue. Do we have access control the windows settings on this machine? No, we'll need admin access. Actually its controlled by a GPO. Now we need to raise a ticket with the active directory team and may need a custom change. Custom changes need a compliance and security officer to review. Because this change also requires a reboot a impact report needs filed before implementation and reboot. Do we have to test this in the customers Dev/UAT/QA system before implementation? And if this is a new issue will I have to write up the issue for the documentation team?
And after that I'll have spent 10 hours of on the call time with the customer, mostly waiting on them. And next year when sales comes to them with an even higher price tag for the same product they pay the bill because they know when they call we'll stay with them until the problem is solved.
I guess that theoretically it's possible to have a productive very large organisation but it's really hard to do that compared with the simplicity of a smaller team.
Fred Brooks - The Mythical Man-Month (1975)
There's traditional web search.
Maps.
Images.
Video (which is essentially YouTube integrated into Google Search).
Each of these products would be Fortune 250 companies on their own.
Obviously, it's all powered by the Ads team / business.
But that's two majorly different parts - Google's 1P ads (including video), and the display network (3P ads).
None of this is useful without a gigantic analytics org.
And that's just "search".
Cloud & G-suite would each also be Fortune 100 businesses.
maybe years ago, this is not the case now.
This is obviously untrue if you want to offer and kind of support at all.
I also strongly doubt it's true you can scale up your application arbitrarily without needing more people, though I'll readily believe that it's a limited increase compared to what will happen if you have 20 product managers and designers trying to find ways to justify their salaries.
People want to share videos and now even uncompressed photos. Considering Facebook paid a lot of $ for the service they can't let it slide into obscurity.
It lead to performance issues at almost all of our large customers.....
See, before that point customers would just export the XML statistics and generate their own reports, but now with the reports being good, they pushed report generation back to our product. Report generation had always been coupled with another service before that point so a team had to be split off to separate these services and API controllers so they could run standalone. This required a fast feedback loop with the customers to ensure we understood their storage requirements and the ability to access secondary SQL datastores instead of the primary. Now we have a much more scalable, but far more complicated product so we had to train the support team handle these issues.
Fallacy: https://wikipedia.org/wiki/Sunk_cost
It is hard to learn to avoid biased thinking in our own lives! I think that cutting one’s losses is a very important skill to learn.
This is 100% true. Where scope creeps in is someone realizes that adding feature X will increase usage by enough that it's ROI-positive, even with a team of 10 supporting it.