What is a data concierge, and when does a company need one?
A data concierge sits between a fractional CDO, a boutique agency, and a senior freelancer. Here's what the role actually does, who it's a good fit for, and how it compares to the alternatives.
Most people who land on this site haven't heard the phrase "data concierge" before. That's intentional, it's the name I picked for what I do because none of the existing labels (data consultant, fractional CDO, analytics agency, contract data engineer) actually describe the work honestly.
If you're trying to figure out whether a data concierge is what your company needs, this post walks through the alternatives, what each one fails at, and when the concierge model fits.
The four ways most companies bring data expertise in
Before we get to what a data concierge is, let's be honest about the existing options. When a company decides "we need help with our data", they almost always end up looking at one of these four:
- A full-time hire. Bring in a senior data engineer, analytics engineer or head of data. Slow to hire, expensive on payroll, requires a clear job description (which is hard if your data problems are ambiguous), and the wrong hire is painful to undo.
- A traditional consulting firm. Big-name strategy or analytics consultancy. Excellent slides, expensive, the people who win the contract aren't the people who do the work, and "implementation" tends to mean handing recommendations to your team for execution.
- A boutique data agency. A small team that builds for you end-to-end. Faster than a big firm, but you're typically getting a generalist team where the senior person is selling and the work is done by mid-level engineers who rotate across projects.
- A senior freelancer. A single experienced operator on a contract. Usually great execution, but the engagement is scoped narrowly to "build this thing" rather than "fix the broader picture", and there's no continuity once the contract ends.
What a data concierge does differently
A data concierge is the senior-freelancer model with three differences, borrowed from how a hotel concierge actually works:
- Personal accountability for the whole stay, not just one task. A concierge owns the outcome of your data foundation, warehouse, dashboards, governance, AI-readiness, even when individual pieces are delegated. The agency model fragments accountability across team members; the concierge model keeps it on one person.
- Discreet, opinionated, and senior. A good hotel concierge has strong opinions about which restaurant you should try and tells you, instead of handing you a menu. A good data concierge does the same with metric definitions, tooling choices, and architecture decisions, not a survey of options, a recommendation with the reasoning.
- A persistent relationship, not a transactional project. The concierge doesn't disappear after the warehouse is shipped. They stay on a thin retainer, available for the questions and decisions that come up as the company evolves. Continuity is the point, not an upsell.
When a data concierge is the right fit
The model isn't for everyone. Big enterprises with mature internal data teams don't need a concierge, they need either a strategic firm to argue with, or specialised contractors for narrow scope. Tiny startups pre-product-market-fit don't need one either, a senior freelancer for ten hours building a Stripe-to-Sheets pipeline is plenty.
The concierge model is right when all three of these are true:
- You're past "we just need a dashboard" and into "we need a foundation", typically 30+ employees, multiple data sources, executive demand for numbers that are starting to disagree with each other.
- You don't have a senior data leader yet, or your senior data leader needs a peer to bounce decisions off, not another report.
- You want one person on the hook for the whole picture, not a roster you have to coordinate.
How concierge work is structured in practice
For my engagements, the structure is usually two phases. A 1–2 week paid diagnostic to map the current state honestly, including the metrics that disagree, the orphan dashboards, the "we're not quite sure" pipelines, and propose a concrete sequence of fixes. Then a 6–12 week build phase where the high-priority items get fixed, followed by a thin monthly retainer for continuity, governance reviews, and the inevitable next questions.
The retainer is usually where the value compounds. Most of the costly data decisions a company makes happen quietly: a new tool gets adopted, a new metric definition gets invented, an analyst leaves and takes context with them. Having a concierge on retainer means there's someone whose job is to catch those decisions before they harden into another duplicate-metric problem.
A 30-minute discovery call costs nothing. You walk me through where you are; I tell you whether the concierge model would help and, if it wouldn't, who I'd refer you to instead.
See if the concierge model fitsRelated reading
If you're evaluating models for bringing data expertise in, two other posts on this blog are worth pairing with this one: duplicated metrics is the canonical reason most companies start looking for outside help, and fixing your data before AI is the strategic frame I use when companies arrive asking about AI adoption but actually need foundation work first.