The instinct is to hire an analyst. It is often the wrong first move, and it ends with your whole capability resting on one person who can resign. A better way.

Do not make your first data hire an analyst

When a business decides it is serious about data, the instinct is to hire an analyst. It feels like the obvious move, and it is often the wrong one. The work that creates a trustworthy analytics capability spans three different jobs, getting data in and clean, modelling it, and turning it into insight, and one person rarely does all three well. Hire a single analyst into a business with messy data and no model, and you have bought insight with no foundation to stand on.

For more on this, see our data strategy and leadership work. It is also worth reading alongside Hire For The Next Decade, Not The Last One if the topic is new to you.

The lone-analyst trap

The common first hire is one capable, enthusiastic analyst. For a while it works. They build things, people are pleased, demand grows. Then the trap closes. They are doing the engineering, the modelling and the analysis at once, stretched across all three and excellent at none. Everything runs through them, undocumented, in their head. They become impossible to give leave, dangerous to lose, and the whole capability is one resignation away from vanishing. The problem is not the person. It is asking one person to be a department.

So work out which part is actually missing before you write a job advert. Most mid-market businesses are not short of questions, or of people to ask them. They are short of clean, modelled, trusted data to answer from. Their real gap is engineering and modelling, not another analyst. Hiring the role that sounds most appealing, rather than the one that fills the hole, is how you end up with a brilliant analyst sitting idle on top of data nobody has prepared.

When a partner beats a hire, and when it does not

Early on, you need a range of skills you cannot justify employing full-time. The initial build wants an architect's judgement, an engineer's plumbing and a developer's modelling, for a few months, not forever. Hire one generalist and you get one person's range. Engage a partner and you get a team's range for exactly the phase that needs it, and then it tapers. A partner also has properties a single hire does not: it does not take leave at month-end, it does not resign, and it has seen the problem before.

Once the platform is built and the work settles into running, extending and maintaining what exists, the balance flips. An employee is usually better and cheaper than a partner on the clock, and you want that knowledge living inside the business. A partner who is still indispensable two years in has failed at the part that mattered, which was reducing your dependence on them.

The hybrid that usually wins

For most mid-market businesses the right answer is not one or the other. A partner builds the foundations and does the genuinely hard initial work, while your first hire, often a BI developer rather than an analyst, learns to run and extend it alongside them. You get capability now and knowledge in the building for later, and the partner deliberately tapers as your own person grows into the role. You end up with both the result and the independence.

The one rule underneath all of it

Whatever you choose, design out the single-person risk from the start. Document the model and the pipelines. Certify what is trusted. Cross-train, so at least two people understand anything that matters. The lone-analyst trap and the indispensable-partner trap are the same mistake wearing different clothes, and the cure for both is shared, documented, owned capability.

So before you post that first job advert, ask which of the three jobs you are actually short of, and whether one person can really cover them. If the honest answer is no, you have just avoided the most common and expensive first move in data. Building the foundations and handing them to your own people is the whole shape of our Analytics Acceleration Programme, and where a fractional data lead fits before you hire. The full guide to hire, partner or both is in Your First Data Hire.

If any of this sounds familiar, talk to us about your data.

Related reading

• Today Versus Tomorrow: The Data Leader's Toggle
• Your Data Strategy Should Fit On One Page
• Work From Happy, And Why We Put A Number On It