Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: At your company, what things do you want to outsource but can't?
53 points by stretchwithme on Dec 2, 2017 | hide | past | favorite | 52 comments



The problem with outsourcing isn't what we can or can't outsource. It's the constant friction of vendor relationships.

Outsource a function that requires a group of skilled people? The vendor you picked will gradually reduce the quality/availability of the staff you work with. Until eventually you're paying a vendor to train their new hires.

Outsource a software service? Expect them to use vendor lock-in to their advantage. The price will increase 10%+ annually. Changes will become exorbitant.

Outsource a department? Watch their scope become narrower over time, and their demands on anything outside the department increase. Eventually you need massive forms filled out 'just so' to get anything out of them. Errors become rife, and require a lot of effort for an outside party to fix. Emergencies for the business are not emergencies for the outsourced department.

As someone who makes these sorts of decisions regularly, it's rarely a question of what we want outsourced. It's a question of how we can package the function in a way that is conducive to a vendor relationship.

Not a rant against outsourcing. There are other structural risks with in-sourcing.


As someone who was worked both in and with an outsourcing firm here's my take on it - An outsourcing company wants to have a steady stream of income. So, they give you the best pricing for your project.

During the negotiations they also see if you have some form of in-house expertise. If yes, they will provide us the best available people. If no, they try hard to sell you few duds - people with bloated resumes but not enough knowledge. You end up paying for their training.

Then it comes to actual implementations. If you hired duds, good luck with that. You will be constantly be on the hook for some unreasonable demands. If your hires are good, then everything should be smooth sailing.

Once the project is over the outsourcing firm, wanting a steady stream, will try negotiate a support contract. If you do then the best guys in the outsourcing team will start leaving. Then you end up paying for a lot of duds.

My take is as a technology company, never outsource some new technology. Need someone to work on say Airflow? Build your team and train them. Yes there is a training cost required but would you rather train FTEs or pay an outsourcing cost and training fees as well.


I deal with a lot of investors who want to know insourcing vs outsourcing for their portfolio companies.

Here's the thing about outsourcing, the people (FTE's) managing the vendors have to be just as deadly or smarter than their outsourcing peers, otherwise service providers pretty much walk over them.

Marketing tech service companies are notorious for this. I'm looking at you PPC, SEO and Ad buying companies...you're the worst.


That is perfect description of what happens.

The other quote I had was every 5-10 years management gets the idea to either outsource or insource to save money. Just switch back and forth every time you get a new CEO.


I have heard of management consultances generating revenue by simply recommending as "best practise" as whatever the opposite of what their customer/victim is currently doing.


I wonder if that has to do with the scar tissue that arise around the interactions of a given system. Sometimes, abrupt change is the answer not because the people are the issue ("too expensive") but rather because the system doesn't work as a whole.


Every six months or so I rotate my room's furniture layout among a couple arrangements to increase efficient use of space and feel better.


It's a matter of alignment of incentives. Within the company Department X knows that the company's success is their success, one team, one dream. On the outside their success comes from extracting as much as they can from the host entity, without killing it. You can't fault people for acting as they are incentivised to do of course. You can only set up a system of incentives to shepherd people towards the outcomes you want.

Emergencies for the business are not emergencies for the outsourced department.

Well of course not. If someone wanted to work for company Y, it's because, for whatever reason, they wanted to work for Y. If Y stabs them in the back, TUPEs them over to an outsourcing company, why would they go the extra mile? Good people whose goodwill has been wantonly squandered, and who know where all the skeletons are buried...


Is there a way to solve these issues - perhaps even via a startup?


Wow. Great answer! Sounds like you also have experience with crappy outsourcing s well


I am not sure about "want to" but you can never outsource the "core" of your company. You build software as your core product ? Don't outsource it. You provide marketing as your core ? Don't outsource it. The only thing you can consider outsourcing is things that are tangential and will never directly impact the viability of your company.


What about whatsapp ? Didn't they outsource their product development to Russia?


What? I thought whatsapp was pretty renowned for doing their development in erlang with a very small core of engineers (18?).


This article started that rumour due to an error: https://elpais.com/tecnologia/2012/07/03/actualidad/13413401...

Towards the end, they have clarification that states: "Technological development is mainly done at the WhatsApp headquarters in Mountain View and not in Russia as the information suggests. There the company maintains a 'small technical team'."


Thank you, I stand corrected.


They've had development in the engineering department somewhere in Silicon Valley, no?


In Mountain View, just off Castro street, AFAIK. Their name isn't even on the building. Super low-key. An external shot of the building was used in the show Silicon Valley as a stand-in for the Pied Pier office in season 2 or 3 (where Jack Barker is the CEO)


My firm has been grappling with outsourcing the transcription of sensitive internal communications for years.

Our weekly sales calls are done over teleconference, but not everyone can attend. Obviously, a lot of information that we don't want to be public is shared during these calls, like sales figures and client status reports. The calls can last anywhere from 30 minutes to two hours.

Right now, our copywriters in marketing transcribe these calls, which they are not happy about. It can take half a day or more to transcribe one call, and it's not the best use of their time. Although they rotate each week, it still takes up a significant amount of time and is not really part of their job description.

Our COO does not want to use an external service to outsource this, as she is concerned about leaks. The fact that many transcription services exist where an NDA must be signed does not mitigate her concern. However, she also does not feel like hiring a full-time transcriber is worth the cost.

The copywriters have tried automated transcription tools but they are not good enough.

I have no idea how this can be solved except for hiring a full-time transcriber ...


Listening to a tape and writing down what you hear is entry level work. You’re wasting more money sitting on the fence about this than you would by just hiring a college student at $15/hr.


Look at your state’s procurement website and figure out who does their court transcription.


Hiring people as employees around the world without setting up an entity everywhere. Stripe for employment.


This is a good one. It never occurred to me the hoops one would have to jump through to hire someone in another country by the books.


The challenges I've had with outsourcing been due to outsourcing prematurely. Similar to premature optimization. Until I've spent significant time in a role, outsourcing it tends to create more work than it saves.


Everything which is not core but without managing the outsourced task.

I don't save money/time when i have to explain to much and when i have to finish it or checking the result takes to much time.


> I don't save money/time when i have to explain to much

Wonder how many tasks have "definition" part and "execution" part, so you can make "definition" yourself and let somebody else to do the "execution". Do you mean this approach should work?


When i have to define so much, that i could do it myself, i don't save anything much.

Defining only works when you outsource it to someone who knows the task better than you.


Hi, I'm founder of https://bitbank.nz I'd love to outsource the marketing, i created a referral program where you can earn .003btc per paying user referred.

But without traffic people don't learn about it.

We have a good conversion rate and good product but without the traffic its meaningless really, advertisers in the bitcoin space seem to accept bitcoin and send fake traffic that never converts into a real user at the rate of other sources (e.g. HN).

Also the way most advertising companies work is they want to see the money upfront, with startups you don't really have it until you get traffic, there's a chicken and egg problem there that makes it hard to get off the ground even when everything seems to be going great.

The advertising companies aren't really bothered about sending you real value just filling the numbers and they aren't aligned 100% with your success, i suppose there are a variety of models that have been tried out already e.g. some companies hire people and offer sales bonuses ect. some offer seo/content marketing ect but again arent incentivized to actually deliver just to take as much of your money as possible... :/

perhaps some kind of paid retweeting service like bitcoin birds (but one that actually works) could work well too. The mainstream advertisers are crazy expensive cpc too.


It is hard to outsource reports and mobile applications because setting up a sandboxes version of the database and APIs in order to protect sensitive data from vendors becomes harder as your system grows.

There is also the fact that they will need the same level of expertise with the database as you have and that already takes months each time we get a new recruit in the team.

It will just be impractical to hand over all the knowledge and the experience working with a MLOC ERP, so we write our own apps and reports.


Ad-hoc/per project PM/Ticketing. In my day job I've come to appreciate having a backlog of tickets to help me get over the mental hurdle of systems I have to build. I'd love if there was a service where I could chat with a PM for a day or so to explain my project, the vision, the features I want to build, and the structure of the thing; then have them take my assets and make a whole slew of tickets with the specific assets attached to each.


I have nothing to add, but as someone who had a business, I just want to say that that's an extremely smart question to ask.


Being that this is HN, it's pretty much assumed the person posting is fishing for business ideas.


My leadership team.

They are managing not for growth, but for their own survival until retirement. Totally devoid of innovation or initiative.


Architecture & project management. I was promoted from contractor to employee and was told basically not to spend any time on the project anymore because there are new (bright, capable) contractors to work on it.

But it's never in a contractor's best interest to take responsibility for the big picture. So I spend lots of time testing contributions (okay but not much fun/reward) and being a project manager (detestable and a black mark on my resume afaic).

Why I took responsibility when I was a contractor? Clearly I do not always act in my own best interest.


We would like to outsource backup of our KVM VPSes. We have a few requirements:

1. Access to snapshots of the last 30 days with single day granuality.

2. Backup has to be encrypted.

We have 7 full servers with over 100 vpses.


Not looking to do work for you, just a tip. Look into ZFS. Daily snapshot, incremental etc? No problem.

However something you'll need to think about if you're snapshotting - do you need to snapshot the memory state/disk buffers too? e.g. a database. Or anything where the total state of the VM is split between disk and CPU/memory.

If the answer is no this is an easy problem to solve.

If the answer is 'yes' life gets more complicated.


We manage this somehow. The issue here is that taking snaphshot is only a part of the problem. Then there is transfering data, managing it, retention, hard drive management and so on. And we don't see competitive advantage here. We just need this requirement to be fulfilled.


Marketing, but not just marketing.

I want marketing that demonstrates direct returns and validity through empirical testing.

Most marketing people and companies I've seen will happily take your money all day for no demonstrated return at all, and it's incumbent on you to hold them to account. Holding them to account is the expensive part, not doing the marketing.


PPA/CPA? There's a big industry built around letting others market for you based on results alone.


Speaking of outsourcing, does anyone have any advice on finding, evaluating and hiring contract dev teams?

I posted an “ask hn” a few days ago about this but it did not seem to make it to the ask page:

https://news.ycombinator.com/item?id=15816396


I (and many in my department) wish we could outsource some extract work that is weekly and takes about 2 to 10 hours of my time. Unfortunately, the state of Texas forbids any offshoring when it comes to their Medicare data so that throws a wrench in our plans to do that.


I'd love to outsource sales. I'd want them to work for a variety of similar shops, and work out deals that use the best services for each job... instead of trying to shoehorn the staff we have into whatever jobs they can get.


My own job. Trust is the key issue. Trusting someone to do an important job takes time.


If it’s something that always must be done exactly right, then it should be automated. Computers are good at that type of thing.


I'm a manager of humans and systems - I can automate parts of my job, but long term planning and strategy are very much human domains.


Build a Magento plugin, because PHP is shit and coding a Magento plugin is a nightmare. Can't do it because it may be too expensive ...


please drop me a few lines of spec, i'll send you an offer, you will like.


Management. And recruitment, so that I could have direct access to the CVs of those applying.


Audits. A lot of it is outsourced but I wish they would fill in forms for me as well.


Oncall.


For what types of systems? This is something I've worked on a little bit. How far would you want to outsourcee to go in fixing issues? (Ie how much would you trust them to touch/change?)


Everything, if you could, you would outsource everything, and just jump around one and it'd be so much fun.





Consider applying for YC's Summer 2025 batch! Applications are open till May 13

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: