Google, Alibaba, Slack, Skype and AppSumo all outsourced software development. Two of them are among the largest companies on earth, one was bought by Microsoft for 8.5 billion dollars, and one was built for fifty dollars.
That is worth saying first, because the assumption underneath most of these conversations is that outsourcing is what you do when you cannot afford to build properly. The record does not support it. A good deal of the most valuable software ever written was produced, at least in part, by people who did not work for the company that owned it.
Here is the second thing, and it is the one that actually matters. Outsourcing is not what made those five succeed. Plenty of companies have outsourced software development and lost a year, a budget and a codebase in the process. What separated the five below was how they did it, and that is a far more useful thing to learn than a list of names.
What is an outsourced software development company?
An outsourced software development company builds software for other organizations under contract, rather than employing the developers directly. You bring a problem or a product. They bring the engineers, and usually the project management, QA and infrastructure that go with them.
The term gets used loosely, and five different arrangements are routinely called the same thing.
- Outsourcing is the general case. Work that could be done in-house is contracted out to another company.
- Offshoring is outsourcing to a company in a distant country, usually chosen on cost.
- Nearshoring is the same idea in a nearby country or a compatible time zone.
- Staff augmentation is renting individual developers who sit inside your team and report to your managers. You still direct the work. You have simply rented the hands.
- A development partner is different in kind, not degree. You are buying judgment as well as execution: a team that will tell you when the thing you asked for is the wrong thing to build.
Those distinctions matter more than they look, because they decide who is responsible when something goes wrong. With staff augmentation, that is you. With a genuine partner, it is shared. A large share of the horror stories in this category come from people who bought one and believed they were buying the other.
| Model | Who directs the work | Who owns the outcome | What you are buying |
|---|---|---|---|
| Outsourcing | The vendor's managers | Defined by the contract | A delivered project |
| Offshoring | The vendor's managers | Defined by the contract | A delivered project, at a distant-country rate |
| Nearshoring | The vendor's managers | Defined by the contract | A delivered project, in a compatible time zone |
| Staff augmentation | You | You | Hours from rented developers |
| Development partner | Shared | Shared | Judgment as well as execution |
Yes, that Google. If you want proof that companies of any size and stature can benefit from contracting out development work, start at the top.
In 2019, The New York Times obtained an internal Google document showing the company employed roughly 121,000 temporary workers and contractors worldwide, against 102,000 full-time employees. The contractors outnumbered the staff. Google has been contracting out work since at least 2011, when it began sending AdWords support to call centres worldwide.
The reason is not money. Google can afford anybody. It is flexibility. A permanent workforce at that scale cannot expand and contract with the work, and a deep third-party bench is what lets a Google engineer take parental leave, medical leave, or an actual holiday without a project stopping dead.
There is a less flattering reading, and the Times made it: contractors at Google were paid less, given fewer benefits, and largely excluded from the company's internal life. Both things are true at once. Contracting gives an organisation flexibility, and how it treats the people providing that flexibility is a separate decision entirely.
Alibaba
Alibaba is the largest online retailer in China and, by some measures, the largest in the world. It exists because Jack Ma outsourced.
When Ma founded the company in Hangzhou in the late 1990s, China had barely any internet connections and Hangzhou had almost no programming talent. So he had the original site built by developers in the United States.
Not because it was cheap. Because it was the only way to get it built at all.
That is the cleanest version of the argument in this entire article. Outsourcing works best when it buys you a capability you do not have and cannot quickly acquire.
Slack
Slack was an accident, and outsourcing is a large part of why the accident turned into a company.
The people who built it were making a video game at a studio called Tiny Speck. Slack was the internal chat tool they wrote to make their own working lives easier. When they realised the tool was better than the game, they took the prototype to MetaLab, a design company, and handed over the interface.
Andrew Wilkinson, MetaLab's founder, has written about the engagement: "In late 2013, Slack hired us to help them turn their early prototype into a polished product. We did the logo, the marketing site, and the web and mobile apps, all in just six weeks from start to finish."
Six weeks. That is the detail worth holding onto, because it is only possible when the brief is small enough to describe in a sentence. MetaLab was not asked to build Slack. It was asked to make an existing prototype feel like something people would want to open.
Slack did not outsource to save money. It outsourced to get a perspective that did not exist inside the building, and that perspective is most of what made the product spread.
Skype
Skype was founded by a Swede and a Dane. It was written by three Estonians who answered a newspaper advertisement.
In 1999, the telecoms company Tele2 needed programmers and could not find them in Sweden, so its Estonian office placed a classified ad in a Tallinn newspaper. The work went to three schoolmates: Ahti Heinla, Priit Kasesalu and Jaan Tallinn. Niklas Zennström and Janus Friis worked with them on the Everyday portal, then on the file-sharing tool Kazaa, and then, in 2003, on Skype.
eBay bought Skype in 2005 for around 2.6 billion dollars. Microsoft bought it in 2011 for 8.5 billion.
Now hold that against the prejudice sitting underneath this entire topic. The cheap foreign contractors, hired through a newspaper ad because the company could not find talent at home, built a product that two of the largest technology companies in the world queued up to buy.
And the story does not stop there. The Estonians had equity. When Skype sold, the payout seeded what Estonians now call the Skype Mafia: more than thirty companies founded by Skype alumni, including Wise and Bolt. Jaan Tallinn became one of Europe's most active technology investors. A country of 1.3 million people built a startup ecosystem on the back of a contract that a Swedish telco placed in a newspaper because it could not find anyone closer to home.
Contractors, if you treat them as people rather than a line item, have a way of becoming something else.
AppSumo
AppSumo is the best story on this list, and it cost fifty dollars.
Noah Kagan wanted to know whether his idea would work before he spent real money finding out. By his own account, he found some registration code online, hired a developer at twelve dollars an hour to wire it up to PayPal, and pointed some Reddit ads at the result. His first customer was Imgur.
Fifty dollars bought a working MVP and, more to the point, an answer. AppSumo now does around eighty million dollars a year.
Sometimes the answer is all you need. And sometimes the cheapest way to get it is to pay somebody else to build the smallest possible version of the thing.
What the five have in common
Line them up and the pattern is not subtle.
They outsourced a defined problem, not an ambition
Alibaba needed a website built. Slack needed an interface designed, and got it in six weeks. AppSumo needed PayPal wired to a page. Not one of them handed over a vague brief and hoped. Each scope was small enough to describe in a sentence, which meant it was small enough to check.
They bought a capability, not hours
Every one of these was about getting something the company could not produce itself. Talent that did not exist locally. A design instinct that was not in the room. A working prototype before the money ran out. Not one of them was primarily about paying less per hour for the same work.
They stayed involved
Nobody on this list wrote a cheque and walked away. Ma was in the room. Slack's founders had a specific and strongly held opinion about how the interface should feel. Kagan knew exactly what he was testing and why.
The good ones stopped being transactions
Skype's contractors ended up with equity and went on to found an ecosystem. Google's contractor bench became structural. The relationships that worked did not stay at arm's length, and the ones that stayed at arm's length are not on this list.
When outsourcing software development fails
Almost every page written on this topic is written by somebody selling outsourcing, which is why almost none of them tell you when it goes wrong. Here is the honest version.
You hand over a product you have not specified
This is the big one, and it accounts for more failed engagements than every other cause combined. If you cannot describe what you want in enough detail to know whether you got it, no partner on earth can rescue you. You will get something. It will not be the thing.
You buy on hourly rate
The cheapest developer is almost never the cheapest project. Rework, missed requirements and a codebase nobody else can maintain cost far more than the rate you saved. We spend a meaningful share of our time rebuilding applications that were built cheaply somewhere else, and the second build always costs more than the first one would have.
Nobody technical is minding the store on your side
Someone on your team has to be able to tell whether the work is good. If nobody can, you are asking a supplier to grade its own homework, and most suppliers will award themselves an A. A fractional CTO exists for precisely this gap, and it is far cheaper than discovering the gap eighteen months later.
The partner disappears at go-live
Launch is the beginning of a piece of software, not the end of it. A shop optimised for delivery rather than maintenance will hand you a working application and a slow-motion problem, and you will not find out which you bought until about month four.
You do not own the code
Read the contract. If the intellectual property does not transfer, or if the application depends on a proprietary framework only they can maintain, you have not bought software. You have rented it, and the rent goes up.
Every one of those is a contract problem or a communication problem. Not one of them is caused by the developers being somewhere else.
How to evaluate a development partner
Six questions. The answers will tell you more than any portfolio.
Who exactly will do the work, and will I meet them? If the people in the sales meeting are not the people who will write the code, ask why not. At JetRockets a co-founder leads every engagement, which is a deliberate answer to a common failure rather than a slogan.
What happens when somebody rolls off my project? People move between teams at every agency, ours included. The ones worth hiring have a real answer for what that costs you and how they contain it. If they tell you it never happens, they are not being straight with you.
Who owns the code? The only acceptable answer is "you do," in writing, with no framework lock-in.
What is the pricing model, and what does it include? Ask whether QA and DevOps are billed separately. Ask for the minimum engagement. If either answer requires a discovery call, that is information too.
Will you tell me no? The most valuable thing a good partner does is talk you out of things. Ask for an example of a time they told a client not to build something. A shop that has never done that is selling hours, not judgment.
Can I hear about one that went badly? Everybody has a portfolio. Ask about the project that was hard, and listen for whether the answer includes any of their own mistakes.
If you want a low-risk way to see how a partner thinks before committing to anything, ask for a code audit. Ours is free, covers architecture, performance and security, and comes back as a written report in two to three days.
Onshore, offshore, or somewhere in between?
You are going to ask, so we will answer.
JetRockets is headquartered in Brooklyn and the company is run from New York. Our engineering team is global. We do not use the words offshore or nearshore to describe ourselves, and the reason is not marketing.
After the invasion of Ukraine, our CEO Natalie Kaminski relocated 46 engineers and their families out of Russia at the company's expense. She then hired Ukrainian engineers who had been left without work by the war. She has talked about it on NPR and elsewhere, and it is not a story we tell for its sales value, because it does not have any. It cost a great deal of money that would have been considerably easier not to spend.
But it is the reason the team looks the way it does, and it is why "offshore" is the wrong word. Offshore describes a procurement decision, taken to reduce a cost. This was the opposite of one.
Look at the Skype story again. Three Estonians answered a newspaper ad because a Swedish telco could not find engineers at home. They were, by any definition anyone would have used at the time, cheap foreign contractors. They were also given equity, treated as principals, and went on to build a national economy. The word "offshore" would have described that arrangement accurately and told you nothing useful about it whatsoever.
The practical question you actually care about is time zone overlap, communication and accountability. Your engagement is led from New York, by a co-founder, on your working day, and the accountability sits here. Where the rest of the team sleeps is our problem to solve, not yours.
Working with JetRockets
We are a Ruby on Rails agency in New York. Rails-only for over fifteen years, women-owned, and every engagement is led by a co-founder rather than handed to a junior team once the contract is signed.
Most of the founders who come to us are somewhere inside the decision this article is about. A good number have been burned once already. If you want to think it through with somebody who is not trying to sell you a headcount, that is a conversation we are glad to have.
And if you are earlier than that, still working out how to choose, our guide, The Non-Technical Founder's Guide to Choosing the Right Software Development Partner, covers the ground in more depth than a blog post can.
Frequently Asked Questions
What is an outsourced software development company? An outsourced software development company builds software for other organisations under contract rather than employing the developers directly. You bring the problem or the product; they bring the engineers, and usually the project management, QA and infrastructure that go with them. The category covers several different arrangements, from renting individual developers who report to your managers, through to a partner who takes shared responsibility for the outcome.
Which big companies have outsourced software development? Google, Alibaba, Slack, Skype and AppSumo have all outsourced software development. Alibaba's original website was built by developers in the United States. Slack's interface was designed by an outside agency in six weeks. Skype was written by three Estonian contractors who answered a newspaper advertisement. AppSumo's first version was built by a contractor for fifty dollars. As of 2019, Google employed more temporary workers and contractors than full-time staff.
What is the difference between outsourcing, offshoring and nearshoring? Outsourcing is the general case: contracting work out to another company rather than doing it in-house. Offshoring is outsourcing to a company in a distant country, usually chosen on cost. Nearshoring is the same thing in a nearby country or a compatible time zone. All three describe where the work happens and what it costs. None of them tells you anything about whether the work will be good.
Is outsourcing software development cheaper than hiring in-house? Sometimes, and that is usually the wrong reason to do it. Hiring in-house carries recruitment costs, salaries, benefits, equipment and months of ramp-up, so contracting can come out cheaper. But the projects that fail are almost always the ones bought on price. The cheapest developer is rarely the cheapest project once rework and unmaintainable code are counted. Outsource to get a capability you do not have, and treat any saving as a bonus rather than the objective.
How much does outsourced software development cost? Rates vary enormously by region and seniority, from under 25 dollars an hour to well over 200. JetRockets charges 100 dollars per hour per developer, inclusive of QA and DevOps, with a minimum engagement of 25,600 dollars. We publish both figures, which most companies in this category do not. A simple MVP typically runs six to ten weeks.
Do you own the code if you outsource development? Only if the contract says so. This is the single most important clause to read and the one most often skipped. With JetRockets you own all of it: the intellectual property, the source code and the repositories, with no lock-in and no proprietary framework you have to keep paying to maintain. If a prospective partner cannot give you that in writing, walk away.
When does outsourcing software development fail? Almost always for one of five reasons: you handed over a product you had not specified clearly enough to check; you bought on hourly rate; nobody technical on your side was able to judge whether the work was good; the partner disappeared at launch; or you did not own the code. Every one of those is a contract or communication problem. None is caused by the developers being somewhere else.
How do you choose a software development partner? Ask six questions. Who exactly will do the work, and will you meet them? What happens when somebody rolls off your project? Who owns the code? What is the pricing model and what does it include? Will they tell you no? And can they describe a project that went badly, including their own mistakes in it? A partner who answers all six straight is rarer than you would think, and worth considerably more than the cheapest quote.
Sources
- The New York Times, "Google's Shadow Work Force: Temps Who Outnumber Full-Time Employees" (Daisuke Wakabayashi, 28 May 2019)
- MetaLab, Slack case study and Andrew Wilkinson's first-hand account of the six-week engagement
- Estonian World, "Skype and the Estonian start-up ecosystem" and Skype Technologies
- Noah Kagan, "How I Built AppSumo for $50", AppSumo blog