I have been running a product development agency for a dozen years. We have two kinds of customers, those who come to us with a tech stack, and those who trust us to make the decision. Sometimes the customer has a legacy application, and we need to work with it and improve it. Sometimes its a new application.

There are some applications where Elixir is such a great match that using anything else would be silly. I think that’s true for chat, e,g, WhatsApp, XMPP (ejabberd/MongooseIM), Discord, and custom web/mobile chat things. It’s good for high performance, concurrent network services in general. I think it’s a good match for things that need to be very reliable, and Nerves is a great way to make embedded systems in a post-scarcity world. These are all the things that make people excited about Elixir, and we don’t have to make any excuses.

A similar story can be told for health care and financial services applications, though there are other popular platforms, e.g. Java. There are other apps where it doesn’t make as much difference, e.g. CRUD apps. Elixir/Phoenix does perfectly fine at building these things. It doesn’t have as many libraries as platforms like Rails, but in four years of using Elixir/Phoenix, it hasn’t caused me significant problems. You need to have enough budget to do some new things, but it can be managed.

I talk about my thought process when determining if Elixir was ready in this post:

As a professional, I am not going to recommend technology to a customer that is not right for them. But I don’t think that simply choosing a relatively niche language is a big problem. There are plenty of web frameworks that get abandoned. The half-life of JavaScript frameworks is about a year. That’s probably a worse problem.

We are the long term partners for our customers, and that’s the most important thing for them. We are still supporting PHP, Rails and Python applications from 10 years ago, and we will do the same for Elixir. I think that Elixir is going to be around for a long time, just like Erlang.

Addressing your questions more directly, I think that good developers don’t have a problem with Elixir and Phoenix. It generally helps us with recruiting. We have been able to transition projects to customers staff and train/mentor them on Elixir. If you open up to remote, then finding developers is not a problem. We work remote from our customers, and we hire remote developers. And we are happy to work with agencies that want to ensure that they can deliver with Elixir.

In Scotland Elixir is not having too much adoption for now, thus developers from any level that want to work full time with the awesome Elixir will have a very hard time making their dreams become true, because of the cultural resistance of companies in recruiting remote developers and/or cross training them into Elixir, not to mention that a lot of them already want they come with a lot of experience in Elixir.

I would like to see this trend to change and see companies to not have afraid of:

• hire more remote developers.
• cross training developers from other languages into Elixir.
• giving opportunity to entry level developers looking for their first professional job.
• dismiss the academic background in favour of a self learner and passionate developer.

Million-dollar question: why do you need people on site? What’s wrong with remote? Unless you have a very small budget for a minimally viable product and want it done in a month (which implies a constant back-and-forth communication in an office space), there’s really no reason to require people be physically present.

Why limit yourself like that?

Elixir being young is misleading. Erlang, the original language of the BEAM VM (in which Elixir also lives and runs its code) has been around for 30 years. Countless telecom equipment and internet backbone tech run Erlang around the world. Elixir itself might be young but the tech behind it is as mature as you can get.

If you are gonna write proposals and will advocate for Elixir (not sure if I understood that part of your post correctly) then you absolutely want to include Erlang and its mature background. If you want to advocate, here’s a link to a huge thread I made not long ago: Please help me persuade a Java CTO that Elixir is the right choice. If you wanna read, you can find a few gems in there, shared by several people.