Salary expectations in the data and analytics industry have been greatly inflated by the hype surrounding this sector, as anybody who has tried to fill a data scientist role will know all too well. PhD students who have not even finished their research believe they can command up to double what organisations think is appropriate, for example.
One reason is the seemingly bottomless appetite for these highly-skilled individuals from the giants of digital and mobile platforms and technology. Why bother spending another year to conclude a dissertation when you could get $250,000 working on an AI project at Google, Facebook, Amazon or Apple?
There are issues with pursuing that course, which include narrowing your field down to the optimal placement of an online ad or being a very small cog in a giant machine. If you have an ego or want to demonstrate greater impact on society, moving to Silicon Valley may not be rewarding.
But there are growing signs of backwash from these highly-spun roles into less glamorous, but business-critical positions. During a recent data conference run by a global insurance group, I heard about the results of a benchmarking exercise for an open position in the foundational data governance function.
Using identical descriptions of the role and the skills required, appeals were made for a data quality manager, data governance manager or data integrity manager. The difference between the lowest and highest salary asks was £35,000 - the golden word (as far as candidates were concerned) was “integrity” which was considered to require a premium.
Why should this be? Well, for one thing, integrity sounds more tech-oriented than either governance or quality. There is a long-standing use of the term within the IT and cyber-security realms that appears to have become burnished through contact with digital transformations and data-driven business initiatives.
Another factor which drives salary disparities is the lack of any standardisation of job titles and the underlying roles. In the example above, exactly the same tasks could be described three ways - and they routinely are across the industry, with recruiters and employers alike wrangling with how to make positions sound appealing, while also ensuring they attract the right skills set.
An outcome of that is often job specs that are lengthy pick lists of qualifications and technical abilities that are hard to meet. The now-fading description of “unicorns” for data scientists in the early days of their emergence (roughly speaking between 2012 and 2015) was accurate for the low likelihood of finding somebody able to meet every requirement listed. Organisations have since learned that they need a blend of skills across a data science team, not a sole genius.
Connected to this is the harder-to-confront reality that data and tech jobs are still fishing in a male-dominated pool of candidates. Just as AI solutions will reflect their training data, so data jobs can be described in a way that reflects the men who currently do them, rather than appealing in a gender-neutral way.
This can be over-stated as a reason for the empty positions - organisations do after all need people with specific abilities. But it is the elements within roles and workplaces that sit behind the technical skills that can make a difference. Women are still more likely to have responsibility for child or parental care or to be returning to work after maternity. Flexible or remote working have a stronger appeal - emphasising these options when recruiting can help to close the gap in the candidate gender balance.
But as the example of integrity demonstrates, picking the right term can have a big impact on your salary overheads. By all means find a way to add sparkle to roles that can appear routine. But remember that glitter also attracts prospectors drawn to the gold rush.