If you have the technical/management chops to know what work can be offshored successfully, package up the work, evaluate the shop, ask very specifically for what you want, assess the quality of what comes back and push back when it needs to be fixed, they can be a great option.

Don't know specific salaries but in my experience work is 30-50% cheaper to have done offshore on a project scale (imagine savings are better if you are paying salaries for full-time engineers). Those savings quickly evaporate if something goes wrong or takes longer than expected (very common due to communication issues), then rapidly become multiples of the original cost if you need to throw it away and have it re-written. This isn't necessarily because the offshore developers are bad, but because communicating software requirements is hard enough even without cultural/language/time barriers.

Offshore shops are a tool in a toolbox, one with a more specific use than most people think, that when deployed can be incredibly powerful but when deployed wrong can be catastrophic to a project.

As with all service relationships, building a partnership over time is the best way to get good work from vendors regardless of their physical location.

Never start with a large, mission critical project - give a small piece of work to a vendor and see how they do, and ramp up from there. Even consider giving the same small project to 1-3 vendors and see who does well - the redundancy is a sunk cost but a small price to pay for long term success.

As the owner of an offshore shop, I think you are right on the nose!

I run an offshore dev shop (web/mobile) with over 15 developers based out of my Ukraine office.

I don't think the base country will make too much difference. The biggest issues I had (that would not have occurred in the US) were with cultural barriers.

No matter the country, language can mostly be addressed with the right process and managers, but the attention and service that US customers expect is not something culturally that exists in many other places.

Getting that right, so that my team leads and PMs meet customer expectations was the hard part. I think that is where many offshore projects fail.

If your team is committed to delivering quality, and you, as the end client know how to communicate your needs and check the deliverables, things usually progress very well.

If your offshore team doesn't understand or care about this (many just don't get it, because customer service is NOT a thing in most countries outside of North America and Western Europe) then you will have frustration no matter where you go.

My clients are about 75% US based and I would be happy to talk with you directly if you have more questions or want to discuss a project. Reach out to me here: ethan [at] abovebits [dotcom]

Sure I will reach out for more questions. One important question for now: do you provide dedicated developers to a client or you prefer allocating 1 developer to multiple clients for risk purposes ?

Great, would love to hear from you.

In regards to full time, we do both. Typically we start small and build up to ensure we are both happy with the initial work and can work well together. Once a client is ready to commit to one more more full time developers, we go month to month and the only issue would be scheduling the start time. If a developer was shared with another client/project part time, we would want him/her to finish up that commitment before pulling them off.

What I've encountered - Poland teams are usually way bigger than they are required, not sure why though; e.g. a project that would require a scrum team of 6-7 people, will usually be done by 10-12 people. But Poland pricing is quite good, there are a lot of people, only downside that I encountered is the English language, it still seems that people are not that interested in it :). Ukraine on the other hand, seems more pricy; maybe due to multiple big USA corporations being opened there.. But it seems people are willing to work for international companies than to internal ones. Have you ever considered the Baltics (LT, LV, EE)? As they are way smaller, they have no choice, but to be international (when it comes to language and tech stack) as the internal market is too small. However, pricing is extremely different per companies, so probably your negotiator skills would be tested there.

It depends the field. Some companies in some locations will charge north of $250 an hour for some profiles, and more for an end to end product custom software. We shipped custom ML platforms for large organizations as part of our activities. The applications allowed their marketing and sales department to train models or load models on their data according to their workflow. Custom made. Some solutions allowed them to send personalized offers to millions of subscribers.

For some clients, on call 24/7. Going to the client. Meeting their people. Handling regulatory paperwork for some projects in regulated fields. Even taking care of importing the hardware for them. Interfacing with their data people and learning their third party vendor's billing systems.

As with everything, it depends on the domain, the service, the support, and generally speaking what they're going to do for you, and which part or parts they're tackling.

Repeat business is common as when these organizations like the outcome, they want more. So that's my two cents from the vendor perspective.

Are you saying that offshore shops in Ukraine/Poland charge $250/Hour for best candidates ? In that case, why would someone offshore it and not find devs in the US directly ? I am genuinely asking because the premise of my post was about finding the balance between cost savings and getting projects/work done effectively for companies that cannot afford to hire expensive devs in US.

No experience with Ukraine or Poland, we're based in Algiers and Paris and focused on ML. Our clients are in France and Algeria.

My point was that it's not automatically cheaper as many people I talk with in Algiers assume that our clients in France come to us because we're "cheap offshore services" and we're not.

Our clients come to us because we hold their hand from the moment they want to understand the value of this, to ideation on where in their business this can have an impact, to working with them to execute in that space learning their systems and all, to making sure they have good results. We've shipped products for banking, telcos, energy, rail transportation, and other domains.

We're a tiny company so we're very flexible and can make decisions very fast and clients seem to like that enough to come back and refer people to us.

So, that's the context for us. Not sure how it goes for dev or Web dev or mobile as it is something we did for our clients, but not something the company is focused on. As in, sure, we'll make a mobile application if the client needs it, but it is a vehicle of ML capability. I'm saying this to give more info from someone in such a company, on another continent, albeit in a field and a budget that may not be what you're talking about.

This makes sense. ML is also a high value specialty right now. I believe that general web/mobile development services are a totally different ballgame due to the proliferation of talent, free (and good) tools, etc.

We definitely don't charge anything close to that. Of course, I wish we could... :) I don't see the cost benefit to clients.

Mid-level developer (3-5 years of experience) in Poland makes probably up to $3000 net (after taxes) per month. Dunno about Ukraine.