“Deus ex machina” —Menander

Data science, machine learning, Ai, and predictive analytics are the current ‘rock star’ *shudder*, skillsets of digital marketing, and here at Distributed we provide our clients with access to all of them, either to work on in-house proprietary technology, or to augment their teams and help deliver first rate client work .

As we have quite a bit of experience in hiring awesome data scientists I thought it would be helpful to benchmark what costs you should be anticipating when building your own data science team. Below are a few examples which should shed some light on the discipline and what it takes to bring these skillsets in-house.

Firstly, let’s make sure that it’s a data scientist that you need.

Here’s a great graphic to help visualise that we’re looking for in a data scientist.

Jack of all

So what we’re looking for in a data scientist is someone who is capable of overseeing complex data projects from beginning to execution. In addition to having great technical skills, they need to be able to effectively communicate their findings to others in the organisation. They should also be able to manage a team. They should be able to query databases like an analyst, but also able to perform much more sophisticated analysis using statistical techniques and machine learning, depending on the task at hand.

Like most of us that work in specialisms, each data scientist will lean towards a particular specialism, here’s a quick look at the different specialisms.

The number cruncher:

This specialism is the modern/coding equivalent of a data analyst and is most often found building systems that automate reporting.

The modeller:

Mathematics forms the basis of this data scientists work. They’re usually found modelling complex phenomena or analysing stocks.

The engineer:

Commonly found hacking on pre-written code and pre-trained algorithms, this role is very much about taking data and ‘making it work’, using available technologies.

The machinist:

Need an algorithm written, or need an existing algorithm improved? This is the data scientist you need, they’ll run an optimise any piece of intelligence you need.

The Deep learner:

This scientist needs a GPU and lots of well tagged data, don’t expect any feature work out of this role, but do expect some incredible results from your data sets.

Now I may have generalised a bit there for the sake of reading time, so if you’re a data scientist and you’re reading this, please forgive me if I’ve left out some granular stuff.

Now that we’ve covered what a bit about the thought process behind which type of data scientist you may want to hire, let’s break down the cost of hiring one.

Here’s a snapshot of the average data scientist salary across a few major cities. To get this data we scraped a couple of recruitment sites and Glassdoor, what you see below is the aggregated average salary figure:


  • Over 1,400 unique data scientist positions advertised
  • Average salary offered £78,000pa

New York:

  • Over 2,500 unique data scientist positions advertised
  • Average salary offered $95,000pa

San Francisco:

  • Over 1,800 unique data scientist positions advertised
  • Average salary offered $105,000pa

With this level of competition for the best candidates you can be sure that the ‘A-players’, in this field are agreeing salaries above the averages generated by our research, and a data scientist’s salary is the only cost you should be factoring into your decision making. Let’s look at the cost of the recruitment process next.



Now, there are many industry leaders that believe hiring a single data scientist to be pointless, so you’re looking at multiples of the above figure if you want to establish a permanent data science team within your organisation. Before you begin your data scientist recruitment process make sure you are building a team with a purpose that will have legitimate and sustained impact on your business for the next decade, and if within that scoping process you discover that you do need some data science resources, but not enough for a permanent team, then please get in touch with us at Distributed, we offer fully managed data science teams on a pay-as -you-go basis to all of our clients.

If anyone has any comments/suggestions/additions please reply or tweet me @callumadamson


Dealers of Lightning: Xerox Parc and the Dawn of the Computer Age by Michael Hiltzik

Tesla, Google, Apple, Microsoft, Facebook — none of these companies would exist without the creations of the team at Xerox Parc. From the laptop computer to Ethernet, they invented it all, in an insanely short period of time, and we’ve been using their inventions ever since. This is the story of Xerox Parc and how they invented whole industries with a team of less than 25 people.

Cal on LinkedIn
Cal on Twitter

Join thousands of subscribers building the future of work.