Data practitioners may be in a hot spot in terms of the interest in what they do from the commercial world. But as David Reed finds out, that does not necessarily translate into well-paid jobs or a steady supply of graduates looking to get a feel of big data.
If your business relies on data for key business processes or uses analysis and insight to make decisions, chances are you are busy recruiting. Demand for resource around data management and analytics continues to grow as the value of this function is increasingly recognised.
Having a job to offer in a down economy ought to make the task of recruitment easy. But this is far from the case as both employers and potential candidates fight shy of agreeing terms. As a result, many vacant posts are going unfilled and practitioners ready to take the next step are actually staying put.
A number of factors have combined to glue up the data job market in this way - weak internal pressure for career progression, strong supply from the consulting world and heavy competition from candidates from other sectors. “Gaming and social media are seen as the exciting options,” points out Mark Dexter, managing director at KDR Recruitment, which specialises in the business intelligence and data warehousing sector.
That reflects one of the major changes he has seen since setting up his business eight years ago. “The way companies are using data has changed - data is now a business-facing function, although there will always be a place for a technical worker to build reports and models. But they are in the minority now,” he says.
The sort of analysts which companies are looking for need to combine that technical expertise with a real understanding of business. Dexter says an ideal candidate would be, “a qualified accountant who knows about MIS systems. That is a very valuable person, one who understands how the business ticks at a financial level and can also design the system that reports on it.”
Unsurprisingly, this type of skilled practitioner is relatively rare, as it is to find individuals who combine business nous with an ability to run web analytics or statistical models. The path through higher education in the UK tends to push students towards increasingly specialised areas, rather than creating hybrid skills sets. So computer scientists are often unlikely to look for positions in the front office where they will learn how companies operate, rather than migrating to back office positions in public sector, pharmaceutical or defence industries.
“That type of person has got to be built - no-one is going to come out of University with that combination,” notes Dexter. Higher education has got a lot better at training graduates in core tools, such as SAS, but it is still rare to see that expertise applied to business goals, such as marketing or finance, as part of courses. “The industry needs to help with that education process. It needs to be made more sexy,” he argues.
The timing for such an initiative could not be more appropriate. Early in January, Michael Gove announced a major overhaul of ICT courses in schools to make them more appropriate and interesting. Many currently do little more than explain how to manage an Excel spreadsheet or prepare a presentation in Powerpoint - basic skills that most students could pick up for themselves.
The government wants to start building a pool of knowledge workers better equipped to move into those fast-growing industries like social networks and online gaming. At the same time, it is also proposing that every schoolgoer should study maths up to the age of 18 alongside whatever other courses they take.
Numeracy is often quoted as a major failing in potential candidates and it is certainly an area in which the marketing function lags, having developed from an arts, rather than science background. Yet all of the channels used by marketing and the tools which manage and measure them demand increasing levels of mathematical ability.
What this serves to do, however, is put marketing and the data industry into direct competition for candidates. Figures from the Higher Education Statistics Agency do show a steady rise in the number of full-time students taking maths courses between 2006 and 2011. But at the same time, full-time computer science courses have seen flat demand, while part-time courses at undergraduate level are actually declining in appeal.
Some of this is down to excessive specialisation. Computer sciences graduates have some of the highest level of unemployment because they have taken courses oriented towards specific gaming platforms, for example. These are often not very portable skills. Equally, such candidates might might be capable of retraining to work in data analytics, where visualisation is becoming one of the major tools, but they might simply not be aware of the opportunity.
Several years into a career, analysts and data managers have started to build that hybrid of knowledge and skills and ought to be looking for their next move. Yet it is notable that job tenure continues to grow among data practitioners, probably as a result of risk aversion on the part of both employers and employees.
“Companies do have budgets and want people, but they want the perfect candidate,” says Ian Thomas, director of Jobsworth Recruitment. What might start out as a straightforward brief for a role is increasingly being rewritten by multiple stakeholders, including the line of business where the post exists through to human resources. “You end up with one brief looking for three skills sets. So which is the most important? HR doesn’t want to make a mistake, so nearly every brief is over-written.”
This hesitancy in appointing new staff means opportunities are not opening up to replace staff as they move on. Salaries are being kept down as a result, which creates little incentive for those candidates with the right qualities to move on. “If somebody is working for employer X on £50,000 and employer Y is offering £50,000, why move? Often, that new job doesn’t have the same notice period or pension rights,” says Thomas. He is often advising candidates not to take up a position because they will lose out on the overall package as a result of employers looking to trim back pay and conditions.
Evidence for the flatness of pay on offer can be found in the annual tracking study carried out by Ball and Hoolahan. Among DM, CRM and insight managers - all core jobs for the data-literate marketer - average salary in 2007 was £39,981, rising to £42,920 on average in 2009. But in 2011, the average slipped to £42,646. If the positions on offer are at pay parity, but with less job protection and benefits, it is little wonder that skilled practitioners are not tempted to move.
“People do want to move,” adds Thomas. “Companies are getting more out of their staff because times are tough, so they are being worked harder and are not getting bonuses. Pressure is building up to move, especially in the US. Once confidence returns, lots of people will be leaving those employers who haven’t respected their staff or worked them too hard.”
One curious factor limiting job mobility is the way many client-side employers are choosing to look for candidates. For every company with an over-written job brief, there is one which decides to go around the established recruitment processes and do it for themselves. Putting a job ad up on a social network like LinkedIn has become commonplace in the expectation that potential staff are following the right groups.
While there is some merit in this approach - the idea that like-follows-like is after all at the heart of social networks - it is probably not optimal in the current climate. In particular, it does not draw in those data-literate and mathematically-skilled candidates who see their career path as progressing in those classic routes in the public sector or financial services. Such workers are unlikely to see a job ad in a marketing data group online.
The rise of outsourced service providers is also holding back the development of client-side skills. Getting access to cutting-edge solutions is quick and easy using this route and the number of providers has grown. At the same time, practitioners working for MSPs tend to operate at the same level year-in, year-out.
Thomas argues that moving client-side can be more appealling. “We have roles like head of data analytics in an organisation that doesn’t have those skills. Compare that with being one of 50 people in an agency team. To make a difference, you have to go client side and wow them, be the hero,” he says.
Even with this demand for outsourced services, there are agency practitioners who have lost their jobs because of the downturn and been found out as lacking in skills. “An analyst might be paid £40,000 to £50,000, but when you look at what they are doing, 90 per cent is creating graphics and Powerpoint, very little is doing data modelling,” says Huw Davis, founder of The Huw Davis Partnership.
Echoing Gove’s view of ICT courses, he believes analytics has become too simplistic and focused on presentation, rather than content. His business was set up to work around this problem by having the data modelling done offshore in Singapore, where data analysts cost half the UK rate to employ, and the delivery done domestically.
To really change the status and value of insight and analytics, however, he believes a skills gap on the client side needs to be closed. “A lot of insight teams often get briefed to do lowest common denominator work because their brand managers do not understand the technology and how data can be used to identify trends,” says Davis.
Data mining should be focused on value-adding tasks such as product purchase correlations, K-rates and buying trends. Instead, there is often a focus on generating reports for weekly sales and marketing meetings. “My theory is that 25 per cent of the insight team should be working on research and development just looking at the data,” argues Davis.
That kind of activity is increasingly what analysts want to be doing, especially in the context of big data. It is also what big data holders need in order to understand drivers of their business which are typically hidden. One social gaming network is loading three billion items of data per hour from its platform, for example. Simple reporting on that flow will not identify whether gamers are buying more or less, using combinations of products, playing multi-player versions or upgrading their services.
A growing number of organisations get that there are valuable insights to be gained from such vast operational and behavioural data sets. To extract them, they need skilled data practitioners and are competing in the jobs market for them. If they are fortunate to have a “sexy” proposition (social, digital), they may be lucky and find their hero. For the rest, a lot more training and development is required - and perhaps a bit more cash.