Also, many startups strive for amazing customer service, which usually means fixing issues on the spot when a customer calls in, American business hours. (I realize many remote developers are in similar time zones, but it is worth noting.)
This one person (or small local team) can then 1) drive the product 2) act as a project manager with a big workforce and 3) stop coding, start reviewing code written by the minions and 4) get them where they need to be on new tech stack etc.
Timezone issues should be minor. Even people in the US should only be in the office 1/3 of the day. You need to rethink your whole management style, prioritization-wise if you run into problems like commonly needing stuff during off peak times. (Or, just hire a rotating shift so you have a 24 hr staff in Asia).
A Silicon Valley startup /generally/ has most of their customers in the US. Both US businesses and US consumers have a cultural expectation that issues be resolved quickly within business hours, but are usually more forgiving at 9 PM at night. (For example, if I call my bank at 3 PM, I expect a resolution darn near immediately; if I called at 11 PM I would expect to leave a voicemail. In both scenarios I am equally satisfied, even though in the first I received an immediate solution and in the second I did not.)
It's not a matter of having developers on call 24/7, but in having the 8 hours they are online overlap the 8 hours that the customer base expects prompt service.
If you notified your bank of some technical issue. They would probably respond thanking you and telling you that they would look into it. I would be very surprised if they followed up with you to tell you when it was fixed (although, that would be great service)
Edit: just to add, the software dev on site can review customer services logs asynchronously, he shouldn't need to be the one actually working the customer service queue
So, if you finding someone like that, how are you going to convince them to work on your problem?
most of the times, I work _with_ the client's in house dev teams so that they can take over when we are done. I _love_ this model as everybody wins.
Key is to staff with the right people offshore.
As for the quality of developers, you will find everything across the board. The best shops in India are process driven and serious about continuous improvement and can be a joy to work with. The dirty little secret of IBM Watson and other commercial NLP systems is that it takes a small army to code those kind of things and you can get the manpower there.
I think the #1 fear about a tech company outsourcing is that you don't really control the intellectual property. This might not matter so much if you are running a two-sided market (Uber could outsource development because anybody who wants to rip them off has to go fight city hall in thousands of places) but if you really are a "tech" company you want to keep your knowledge close.
Your employees might take home your entire code base in the US, while working on-site on an air-gapped network. They can certainly do it when you're set up for remote work, almost by definition. In the US, the payoff for stealing your employer's code is low, and the risk of punishment high, but it still happens. I can't speak for non-US countries, but I would guess that most have a bit more relaxed attitude toward enforcement, especially when it would be on behalf of a US company.
If you think your IP is secure from the software developers who wrote it, you do not understand how they work. If you do not hold the loyalty of your developers, you do not control the IP, regardless of what the contracts say.
(since I didn't know)
Regarding item 3, another perspective is that a lot of the outsourcing/off-shoring happens because it is something that is not a core competency of the company. Think any ERP system (SAP, Oracle, etc.) which is basically a CRUD app on steroids. If your company doesn't make ERP systems then you want the lowest skilled bidder doing the work. Is that the case for a tech startup?
It also requires almost the same count of people to offshore the work as it does to do it yourself. You need someone who can interface with the offshore team, answer questions about requirements (if they are even asked), etc. And deal with turnover.
You might be asking why any sane company would do this then? Simply cost but also accounting rules. When a company buys a service from another vendor (like buying bandwidth) I think it shows up differently on the books than when they have an entire department of people doing that work. That entire department needs to have retirement plans, benefits, vacation time, etc. and they show up in the employee count reported to shareholders. This nuance is probably why the counter-intuitive stock price increase happens after a massive layoff.
Ratio between salary ($150K vs $12K ) are too high comparing to ratio of skillset, 10x low salary doesnt mean 10x low skillset, thats just financial state in countries. A lot of guys are working at Amazon, Google, Microsoft and other big corps who took experience here, developers are not bad, they might not have enough experience.
~$300K is avg. spending for 10 developers with office and perks included annually. Compare it to $750K spending for 5developers where each of them has 2x skillset of outsorced developer just for salary and other spendings annually, which I guess will be around $1M/annually.
That's great (no snark intended) if you have a CTO who wants to move to Central Asia. Most startups probably don't.
1) Most companies are not institutionally comfortable with remote working, to say nothing of remote working internationally.
2) There exist pervasive skill gaps between talented engineers available to work at $150k per year and talented engineers available to work for $12k per year. I respect that many people wished this was not the case.
3) Many investors in the Valley, whose opinions are quite relevant given that one cannot self-fund employee salaries, are of the opinion that outsourcing development work suggests a tech company which will never achieve the path to $100 million in revenue followed by a $N billion valuation, which is what their business model counsels them to search for. They would point out "I cannot think of a single company whose product was not produced in-house in history which achieved that" and they would not be far from wrong.
4) Managing outsourced development is a skill, and its a skill that most companies and individuals suck at, and do not particularly wish to become great at.
5) It is the widespread perception in the American tech industry that BPO companies produce software whose quality is abominable and that the process of dealing with them is pathological, as evidenced by the fact that only companies which institutionally hate software and software people are willing to work with BPO companies. Many in the American tech industry do not hate software or software people, and thus would have a very high bar to be convinced to use a BPO company.