Job-seekers hold the upper hand in the job market. Are you doing everything you can to attract the right candidates?

Between my consulting work and my professional network, I get to hear from both sides of the hiring desk. On one side, employers complain that the job market for analytics -- data scientists, data engineers, team leads, and so on -- is a rough sea right now. There are apparently more available data jobs than people qualified to fill them.

There may be some truth to that, but it doesn't explain the whole picture. On the other side, job-seekers have told me what they've seen in the hiring process and it's not all flattering. Employers are unwittingly sabotaging their own hiring efforts, dissuading people from accepting job offers or even applying to interview.

If you're having trouble building your analytics team, then, there may be more at work than a mismatch between supply and demand. Here I'll explain steps you can take to improve your chances of attracting the right candidates and growing your team. The first step is to have a plan.

Develop a strategy

The most common complaint I hear from job-seekers is that companies don't know what they're looking for. They usually say that's because those companies are in a real hurry to build an analytics team before they are sure what to do with data. That makes for frustrating interviews and, worse yet, early departures.

To remedy this, you need to develop a data strategy: a road map of what to do and when to do it, based on your business model, present state, and desired future state. A data strategy will connect data analysis to your business goals. This includes lining up the first few projects to undertake and all of the technical skills they'll require. Part of developing that strategy is to peer into the future and see your business through the eyes of your first data scientist: you uncover relevant use cases, figure out what data you'll need and how to store it, and so on. (If you don't already have these skills in-house, you'd do well to engage an outside consultant to help you bridge the gap.)

Companies often skip over the strategy part. They either don't know that it's an option, or they don't feel it's "real" enough compared to people sitting down at a computer and cranking away at hands-on analysis work. That's unfortunate, because experienced candidates can tell when you don't have a plan. They'll pass you by for someone who takes data seriously.

Do you even need someone right now?

A realistic strategy will also tell you whether you're even ready to hire into your data team right now. You'll first need to get your house in order: acquiring and preparing data, building out infrastructure, or otherwise making sure those initial data hires can hit the ground running. If you plan to collect data from your home-grown applications or your website, for example, you may need to do that for weeks or months before there's enough for meaningful analysis.

(On a related note, data pipelines and cleaning are why your first hire will be more of a data engineer than a data scientist. My post on the roles on your analytics team explains this in greater detail.)

Consider outsiders

Assuming you're ready to start work on your projects, do you really have enough work, and the right kind of work, to merit a set of full-time hires? You may be better off to engage outside help. Consider bringing in some staff-augmentation contractors or outsourcing the work to a larger firm. Not only can these groups tackle your projects, but in doing so they can smooth the on-ramp for your full-time hires once you're ready to build an internal team.

I've encountered companies that are allergic to engaging contractors and consultants. Some of those same companies spend months with an open seat on the team, when a contractor could have launched (and possibly completed!) a project in the time they spent waiting for a full-timer. Ask yourself which is more important: having someone to whom you issue a paycheck and a W-2 form, or completing your project in a timely fashion?

If you're on the fence, work through the following questions:

• Do you foresee a series of one-off projects, with lots of time between them?

• Do you require lots of manpower now to build something that will then move into maintenance mode?

• Similarly, does your project require specialized skills to build but less so to support?

Any yes here is a vote in favor of finding outside help, and to postpone building an internal analytics team. A set of no answers means you're probably ready to build an internal analytics team of full-time hires.

Skip the unicorn hunt

Once you're ready to build your team, it's not as easy as asking the marketplace for a data scientist. That term covers a lot of ground so you will need to clarify your job posting by listing the skills you're after. (You'll also need more than data scientists.)

It's tempting to ask for Every Skill Under the Sun, "just in case." That would be a mistake. Prospective hires will reason you're casting a wide net because you don't know what you're looking for. Instead, use your data strategy (you did define one, didn't you?) to guide you and narrow the scope. Note the skills you'll need for the first few projects and have the discipline to include just those in the job posting. People scanning your list will determine whether they fit the bill and, hopefully, suitable candidates will respond. You're now a step closer to building your data team.

If you don't already have experienced data talent in-house, ask for outside help to write the job description and interview candidates. This will signal to prospecive hires that you're serious about the work they'll do. Furthermore, this extra step will smooth out problems that turn away suitable candidates.

One such problem is the unicorn hunt, a job posting that asks for an extremely rare combination of skills. At best, such a posting will attract people who are overconfident. At worst, it will turn away people who would be a great fit for your company: they'll see that you're asking for multiple roles in one person, and look elsewhere. (If you are intent on finding a unicorn of a data hire, note that you will wait a long time to fill the role with a single person who fits the job description to a tee, which will delay your projects.)

Can't find a seasoned expert? Grow one.

It makes sense to build your first-round data team around experienced practitioners. Once they've set the shop in motion and you're ready for your second round of hires, why not widen your scope? You'll probably have some work that would be a distraction to your senior staff but very interesting to a less-experienced team member.

Consider the number of people who have the raw talent and need some on-the-job experience to shape and mature their data skills. Take your pick of software developers, economists, or anyone else who is interested in the work you do and has the potential to become a contributing team member. They will get a job in which they can grow, and you will develop an expert in working with your data.

The key here is to make sure you're prepared to bring on an entry-level team member. Your junior staff should be vastly outnumbered by the more experienced people so they can always turn to someone for help. If you flip the ratio -- many junior staff to a single senior team member -- you will set yourself up for a failure down the line. You risk building a mess that a lot of experienced people will have to clean up later.

Moving forward

Hiring into your data team will be a challenge for the foreseeable future. Don't make it any harder than it needs to be. Developing a strategic roadmap for your data efforts will help you understand whether and when to build a team, show you what you're looking for in a hire, and make it easier to attract your ideal candidates.

Are you looking to build or grow your data team? I can help you, from early-stage data assessment and strategy, all the way to serving as interim analytics lead. Contact me to learn more.