How to Hire a Head of Data: 2026 Guide

Last updated: June 13, 2026 | By Tom Kenaley

A head of data in 2026 costs $180K to $300K in base salary and closes in six to ten weeks. The hire lives or dies on one early decision: are you hiring a player-coach to build your first data team, or an operator to run one that already exists? Those are different people. The job posting almost never says which.

The title is doing a lot of quiet work in that sentence. “Head of data” describes a player-coach at a sixty-person startup and a forty-person org leader inside a bank, and the same req gets copied for both. That is most of the reason these hires fail.

Last year a logistics company in Austin promoted their strongest data engineer into the head-of-data seat and handed him a team of one. Himself. Eight months later he had stood up a gorgeous Snowflake warehouse, shipped exactly none of the executive reporting the CFO had been asking for, and left for an individual-contributor job that paid less than what they were giving him. He was a builder. They had quietly written a role that was mostly stakeholder management and politics. Nobody told him going in. Nobody on their side knew to.

Where I sit in all this: KORE1 runs these searches and staffs the data science and engineering teams that report up to these leaders, and we earn a fee when you hire through us. So read me with that in mind. I will point out the spots where you should not call anyone at all, because a fair amount of this you can run on your own. We have been placing data and technology talent since 2005, our 12-month retention rate sits at 92%, and the recruiters who work these roles average more than fifteen years doing it.

The Job Underneath the Title

A head of data is the senior-most person accountable for what your data function actually produces. The pipelines, the reporting, the first models, the people who build all of it. They own delivery. In a company without a chief data officer, they are also the face of data to the executive team, which is a second job wearing the same badge.

Here is the line that matters, and it is the one buyers blur. A chief data officer is accountable for what data contributes to the company. A head of data is accountable for what the data team contributes. One is a board-facing mandate about governance, risk, budget, and strategy. The other is a build-and-run mandate about shipping. If you want the strategist with the board seat and the half-million-dollar package, you want a different hire, and we wrote a whole guide to hiring a chief data officer for exactly that. Most companies reading this do not need that person yet. They need the one who builds.

The U.S. Bureau of Labor Statistics does not track “head of data” as an occupation, which tells you how new and how slippery the title still is. The closest proxies it does track are revealing. Computer and information systems managers earned a median of $171,200 in May 2024, with the field projected to grow 15% through 2034. The people your head of data will be hiring, data scientists, are growing at 34% over the same decade, the fourth-fastest of any job the BLS measures. You are hiring a manager for a function that is doubling under them. That pressure is the whole story of the role.

It is also a near-universal hire now. The 2026 AI and Data Leadership Executive Benchmark Survey found roughly nine in ten large firms have appointed a chief data officer, and nearly every one of them has a head of data or a VP of data running the team a level below. Five years ago boards were still arguing about whether the function needed an owner. That argument is over.

Player-Coach or Org-Builder? Decide That First

This is the fork. Get it right and the rest of the search is mechanical. Get it wrong and you will be back here in eighteen months.

The player-coach is your first or second data leader. They still write SQL. They will personally build the dbt models for your first quarter, hire two or three people behind them, and translate “why did churn spike” into a query and then into an answer the CEO can use. You want recent hands-on work. A player-coach who has not shipped in three years is just a manager who interviews well. Pass on them.

The org-builder is a different animal. They have not touched production code in years and that is fine, because the job is twelve to forty people, a multi-team roadmap, a platform budget, and the politics of getting finance and product to agree on what “revenue” means. Hire one of these into a four-person startup and they will spend six months designing an org that does not exist yet while your pipelines rot.

A blunt test before you write the req: how many people report to this seat on day one, and how many on day three hundred? If the answer is “one, then maybe four,” you want the coach. If it is “eleven, then twenty,” you want the builder. Pricing, sourcing, and the interview loop all change based on that one answer. So answer it in the room, out loud, before HR opens a single profile.

What a Head of Data Costs in 2026

Compensation for this role is all over the map, and the spread itself is worth understanding before you set a number. Pull the aggregators and you will see why hiring managers get whiplash. ZipRecruiter put the average head-of-data salary at $176,687 in early 2026, with most landing between $155,500 and $190,000. Glassdoor reported an average closer to $230,600, with its upper quartile north of $312,000. That is a $54,000 gap in the average for the same title.

The gap is not noise. ZipRecruiter is heavier on base salary at smaller companies. Glassdoor’s figure pulls in total pay at funded and public ones, where equity and bonus do most of the work. The title is identical. The job is not. A head of data at a Series A startup and one at a 4,000-person enterprise share four words and almost nothing else.

Here is how it actually breaks down by company stage. These are base-salary bands for U.S. hires in major markets, with total-comp ranges that fold in bonus and equity. Adjust down 10 to 15% outside the top metros.

Two cautions on these numbers. First, they assume direct hire. If you are testing the role or the person, a contract-to-hire data leader costs more per hour and less in regret, and for a first-ever data hire that trade is often the right one. Second, the spread inside a single city is enormous, so do not anchor on a national average. If you want a live read for your exact market and stack, our salary benchmark assistant will pull a current band in about a minute.

Where the Role Should Report

Reporting line is not an org-chart detail. It is the single best predictor of whether this hire stays. The Austin engineer I mentioned earlier did not leave over money. He left because the role had no budget and no air cover, and a head of data with neither is a senior person being set up to fail in slow motion.

Three common reporting lines, and what each one signals to the candidate you are trying to land:

• Up to the CTO. The default at most tech companies. Works when the CTO genuinely values data and gives the role real budget. Falls apart when the CTO treats data as a feature the product team requests.
• Up to the CEO. Rare for a head of data, common for a CDO. If you are offering it, you are signaling that data is a first-class function. Strong candidates notice. So does everyone else in the building, which is its own kind of pressure.
• Buried under the CIO or a VP of Engineering three levels down. This is the configuration that kills the role. The best people read it instantly in the interview and pass. If this is your org chart, fix it before you post, or hire a contractor and skip the title entirely.

One more thing the candidate is measuring, even if they never say it out loud. Do they own a budget, or do they have to beg engineering for headcount every quarter? A head of data without hiring authority is a glorified analyst with a nicer title. They know it. You should too.

Running the Search Without Wasting a Quarter

Strong data leaders are rarely on the job market, because good ones get promoted or recruited before they ever update a profile. So the search is either proactive or it is slow, and there is no comfortable middle option where you post a public job, sit back, and wait for a great data leader to wander in on their own.

Where they actually are: running data teams at companies one size below yours, presenting at meetups and conferences like Coalesce or the local Snowflake and Databricks user groups, and writing the kind of thoughtful LinkedIn posts that no recruiter keyword search will ever surface. The best signal is not a resume. It is a person whose former reports keep following them to their next company. That loyalty is the whole job, demonstrated. Chase that one.

For the interview loop, throw out the algorithm puzzles. You are not hiring a senior engineer. Test the four things the job is actually made of:

• Give them a real, messy business question your company has right now, like “is our new pricing working,” and watch them break it into data they would need, caveats they would flag, and the answer they would give a nervous CFO. This one conversation tells you more than the rest of the loop combined.
• Ask them to design the first ninety days for a team your size. Vague answers mean they have not done it. Specific ones, who they hire first, what they delete, what they refuse to promise, mean they have.
• A short, honest team-building conversation. How do they hire, how do they handle a strong IC who will not mentor, what made their last great hire great.
• Then references. Not the ones they hand you. The skip-level ones, the former report who now works somewhere else and has no reason to flatter.

If you are building your first data org and have never hired this seat, this is a moment where a recruiting partner earns the fee, because you do not yet have the internal calibration to know strong from merely confident. If you have a seasoned CTO who has built data teams before, you may not need us at all. I would rather tell you that than sell you a search you can run yourself.

The Offer and the First Ninety Days

Close fast. Data leaders who are any good are usually talking to two or three other companies, and the offer that lands is rarely the highest one. It is the clearest one. Spell out the mandate, the budget, the reporting line, and the first three problems you want solved. Ambiguity in an offer letter reads as risk to a senior candidate, and risk loses to a competitor who made the job feel real.

Then protect the first ninety days, because that window decides the next three years. Give them one early, visible win to ship, the reporting the executives actually look at or the data quality fix everyone complains about. Resist the urge to drown them in your backlog on week one. A head of data who delivers one thing leadership can see in the first quarter buys the political capital to fix the deeper problems that take a year. A head of data who spends ninety days “getting up to speed” with nothing to show has already started losing the room.

Where Hiring Managers Get This Wrong

Most of the failed searches I see trace back to the same handful of mistakes, and almost none of them are about finding talent, because the talent is usually out there if you know where to look. The failures are about how the job was written, not about who answered the ad.

The big one is hiring the builder you need but writing the req for the strategist you imagine. You picture board decks and data strategy, so you screen for gravitas and polish, and you pass on the operator who would have actually shipped your pipelines while you waited for someone who looked the part in a boardroom. Then you wonder why nothing got built. Predictable, really.

The second is comp anchoring on the wrong reference point. You see “head of data” averages near $230K on Glassdoor, you set the band there, and you do not realize that figure is loaded with equity from public companies your Series B cannot match. Pay the role you have, in the market you are in, for the person who fits the stage. Not the title’s national average.

The third is the quiet one. You hire a real leader and then refuse to give them real authority, and you lose them within two years to a company that will. We see it constantly, and it is the most expensive mistake on this list, because by the time they leave they have hired the team, and the team often follows. For when a quiet, low-drama exit matters, a direct hire search done right protects against most of it. The authority problem, though, only you can fix.

What Founders and CTOs Ask Us Before This Hire

Head of data or chief data officer, which one does my company actually need?

If your data function needs to be built and run, you want a head of data. If it needs to be governed, funded, and represented to the board, you want a CDO. Most companies under a thousand people need the builder first. The strategist comes later, once there is something to govern. Our chief data officer hiring guide covers the board-facing version in full.

Can we promote our lead data engineer into this seat?

Sometimes, and it is often the best hire you can make. The risk is real, though. A brilliant engineer is not automatically a manager, and the head-of-data job is 60 to 70% leadership, hiring, and stakeholder work. Promote the engineer who already mentors, already translates for executives, and actually wants to manage. Not the one who is simply your most technical person, because that path produced the unhappy Austin story at the top of this guide.

Where should a head of data sit in the org?

Under the CTO is the workable default, as long as the role carries its own budget and hiring authority. Reporting to the CEO signals data is a first-class priority and helps you land stronger candidates. The configuration to avoid is burying the seat three levels down under a CIO, which is the fastest way to make a good leader quit.

What’s a head of data really going to cost in total comp?

Plan on $200K to $360K total comp for most growth-stage companies, climbing past $450K at large enterprises where equity dominates. Base alone runs roughly $180K to $300K depending on stage and city. The aggregator averages swing by more than $50K because they mix startups and public companies under one title, so benchmark against your own stage, not a national number.

Do we hire the data leader first, or build the platform first?

Hire the leader first, in almost every case. The platform decisions, Snowflake versus Databricks, dbt versus a managed pipeline, what to centralize, are exactly the calls you are hiring this person to make. Build the warehouse before they arrive and you have just handed your new leader someone else’s architecture to either defend or rip out. Most rip it out.

How do we keep this person from leaving in eighteen months?

Give them budget, authority, and a reporting line that signals data matters, then get out of the way. Money is rarely why data leaders quit. They leave when the job described in the offer letter does not exist inside the building, when they cannot hire, or when no one above them treats data as more than a reporting chore. Fix the role and retention mostly takes care of itself.

One Decision Before You Open the Search

Everything in this guide bends back to the first question. Builder or strategist. Coach or org-leader. Answer it honestly, write the req for the answer, pay for the stage you are at, and give whoever you hire the authority to do the job you described. Do that and this is a straightforward search. Skip it and no amount of sourcing saves you.

If you are about to hire your first data leader and want a second read on which version you actually need, that is the conversation we have every week. Talk to a KORE1 recruiter and we will tell you straight, even if the answer is that you can run this one yourself.