How marketing can fill its data skills gap by rethinking its recruitment strategy
The marketing industry has declared itself to be lacking in people with data skills. In a field that generates a lot of data, it’s long past time for CMOs and Marketing Directors to address this shortfall, and make sure it has the people and expertise it needs to take advantage of the ever increasing amount of data it is producing; data it could be using to generate actionable insight, inform decision making and improve performance.
The data problem
I have often heard the marketing department referred to as ‘the colouring-in team’. The less than subtle suggestion being that marketing is all creative, fluffy, and about making pretty things, often while wearing overly fashionable spectacles and riding on the bleeding edge of the coffee zeitgeist; the department of form over function if you will.
For the agencies that pride themselves on making overly flowery infographics when a simple graph would convey the message much better, this may very well be true.
For the rest of us however, there is slightly more to it than that. From analysing landing page split test experiments using Chi-squared tests and using Monte-Carlo simulation for conversion rate optimisation, to identifying optimal pay-per-click campaign settings using support vector regression, it’s not all about making a shiny brochure with your felt-tips.
And this has created a problem for marketing departments. A recent study by Marketing Week and MiQ has revealed that around 40% of marketers believe that data, research and insight skills are their major knowledge gaps.
In an increasingly data-driven industry, there is competitive advantage to be gained by using these data to do things better. Or, at the very least, identifying what is not working, and stopping it, or doing more of what is working.
Of course, large sets of data going beyond what is often possible to ‘eyeball’. It may be that what you think isn’t working is actually driving increased performance of what you think actually is, and turning off the thing you think isn’t working might upend the entire channel into a metaphorical sales toilet reminiscent of the one proudly featured in Trainspotting.
For me, the most interesting point in the the Marketing Week article is a quote from Direct Line’s Director of Marketing, Mark Evans, when talking about a new hire who added value to the business, but
…would not have made it through a traditional recruitment process.
For teams who are used to hiring ‘traditional marketers’ — and I include the somewhat nebulous term ‘digital marketers’ in that — is there a skills, or process, gap in the recruitment process?
Yes, there is.
I think there must be. There are very skilled people out there. A lot of them. The problem may well be though, that the last position on their CV might be something like one of the following:
- Post-doctoral fellow in hydrodynamics of particle-laden flows
- PhD student studying the impact of transcultural migration on perceived intelligence
- Lead analyst: Cash crop micro-economics, Latin America
- Research associate in cellular immunology
What do all these positions have in common? They are all research-based, and they all involve data.
But they’re not marketers.
One of the the great things about data is that the statistical tests don’t really care where the numbers came from; just as long as you’re using the right one at the right time. Data is data. As with so many things, the real skill comes in knowing the right question to ask, and the way to go about using data and a sound research methodology to answer it.
Unfortunately, if your job advert specifies experience in Google Analytics, Bing Ads, Facebook Insights etc., you might have instantly ruled out the vast majority of those potential applicants.
The power of the multi-disciplinary team
If you are an SME with two people in the marketing team, hiring a ‘reformed academic’ just might not make sense given the lack of economy of scale. You might not have enough data, or sufficient opportunities to take advantage of the insight, to justify a full-time hire. In those cases, you need to find the skills internally, or consider outsourcing to an agency or freelancer.
If you have a large team though, and the regular marketing activity boxes are ticked with people that are already strong in ticking them, then adding someone who brings a completely different set of skills — and perspectives — to the table might really pay off.
Even without a marketing background, if you give a good researcher a problem to solve and some data to get stuck in to, you might be surprised at what they can find out. If you don’t frame a problem up-front, there’s a good chance they’ll still uncover some interesting findings from the data, but the real progress comes when they understand the job functions of the department, and what insight is needed to advance the business performance.
Once your data analyst has spent some time working in your team, getting to know what people do, how they do it, and what they feel they need to be able to do it better, they can get to work. Hopefully not just providing people with what they feel they need to do their jobs more effectively, but also discovering what they didn’t know they needed to do their jobs better.
When your email team is empowered with a classification model that segments customers into the right number of groups that balances personalisation with marketing resources, social has a network analysis map that identifies key influencers and PPC has a predictive model that allows the most effective interaction between search and display campaigns, that’s when you know you’re getting somewhere.
Finding the unicorn of insight in your recruitment process
Of course, first you have to find your (and I hate myself for saying this, but unicorns seem to be big in marketing right now) unicorn of insight. And that might take a significant rethink of how you recruit, and possibly a lot more work recruiting them.
The thing about required skills and essential criteria on job descriptions is that, with each one you put in, you limit your applicant pool. Let’s say 1 in 100 people can perform an analysis of variance, 1 in 100 can build advanced segments in Google Analytics and 1 in 100 people have experience with Mailchimp or Constant Contact. Let’s also assume that those three skills are independent and knowing one doesn’t affect the likelihood that you’ll know the other.
If you make all of those things essential criteria, you’ve taken your applicant pool down to 1 in 1,000,000. Add another criterion, and another, and before you know where you are your applicant pool could be down to 2 people in the country and, let’s be honest, just because they can demonstrate that they meet the criteria doesn’t mean that they’re people you’d want to hire...
Clearly, if you want to address your need for data analysis skills in the organisation in a marketplace where demand is outstripping supply, you’re going to have to do things differently.
Why not go with a kaggle-inspired approach? Make a dataset available (perhaps you could take a custom report from Analytics or Adwords, change the column headings and transform the data a bit to keep the data private) and ask potential applicants for a short report on what the data says to them? After all, that gets right at what you’re actually hiring them to do.
You might end up with a lot of CVs and reports to read, but hiring the best person up front is probably less work than managing the wrong one long-term…
You’re not hiring someone to sit on their behind all day providing convoluted answers to competency-based interview questions. You want to know what they can offer your organisation. You want to find someone who can find something in your data that you currently can’t find. So why not try to find that person by giving them just that challenge.
Take homes to address the skills shortage
A very wise man once said to me “good people are good people.” And that’s something that will stick with me. If you hire the best person available, with the core skills you need — the core skills that take a long time to learn and practise — you’ll do okay.
Don’t fixate on what they don’t know. If they’re driven, research-focussed and take a personal pride in being the first to make a new discovery, they’ll work out what they need to know to be more effective, and what they need to do and with whom they need to speak to get there.
Yes, there might be a learning curve and, for the first month or two their productivity may lag behind that of someone coming in with relevant industry experience, but, if you’ve hired ‘good people’ they will soon get to grips with the business need and vernacular, and may well soon outperform the other candidate. It’s all about the area under the curve.
Breaking the previous thousand words and change down into three key messages:
- A good analyst is a good analyst, the source of the data often doesn’t matter; don’t not hire someone for what they haven’t done, hire them for what they can do.
- People who chose research as a career will likely have a natural curiosity that will continue to drive them to learn new techniques and look for new insight, and quite often at unsavoury hours of the day or night; they just like to find out something new.
- Finding your analysis unicorn may take some work, a lot of time looking at a lot of CVs, and possibly some changes to the recruitment process.
Ultimately, it comes down to how you want to use your data, and whether you think making the most of that data will put your brand in a better place for the future. In the same way that a former fighter pilot led the team that got Apollo 13 home, could a former linguistics researcher empower your marketing team with insight?
