You hired a fractional CDO. 90 days later you own a new stack, a catalog, and a meeting you cannot skip. That is not leadership. That is a vendor with a better title.
A fractional CDO is part-time data leadership. You buy judgment, sequence, and an honest no. You do not buy a product under a new name.
This market exists because a full-time CDO seat is expensive, and most companies do not need that seat yet. They still need someone who can say what to do this quarter. That is a real job. Vendors noticed. They put CDO on the business card and kept the product roadmap.
I have been the buyer. I have been the hire. I was the founding data lead at ShiftKey. I later led data teams at Gallo and GoTu. Before that I was a product manager at HomeAway, part of Expedia, for homepage, traveler search, and geo. The title changes. The test does not.
If the person across from you needs you to buy their tool to succeed, they are not your CDO.
What you are buying
You are buying a leader who works a set number of days and reports to you.
They help you pick a stack that fits this stage of the company. They help you put people on work that has an owner. They help you write a plan that says what you will decide this quarter.
They sit on your side of the table when a vendor arrives.
If they cannot do that last part, stop. You hired a salesperson.
The first 90 days of real work look boring on a slide. They map the decisions the company actually makes. They find the 3 to 5 sources that feed those decisions. They put quality checks on those sources. They name an owner on your payroll. They write down what you will not buy yet.
If the first 90 days are a warehouse migration, a catalog program, and an AI pilot, you are not in a leadership engagement. You are in a delivery contract.
3 ways the pitch hides
1. The tool pitch
The first meeting is about your pain. The second meeting is a demo.
The stack they recommend is the stack they already implement. A warehouse. A catalog. An observability product. A governance suite. The names change. The pattern does not.
A real fractional CDO can name tools. They can also name the stage when you should not buy yet. Picking the right tool for the stage is the job. Selling one tool for every stage is a vendor motion.
Ask this: What will you refuse to sell me in the first 90 days?
If they cannot answer, you are not in a leadership conversation.
2. The catalog-first roadmap
A catalog is easy to show. Search. Lineage. A diagram that looks like control.
A catalog does not make the data correct. It makes bad data easier to find.
A catalog-first roadmap starts where the software lives. A leadership roadmap starts where the decision lives.
Which number does the CEO not trust? Which process still needs a spreadsheet? Which pipeline breaks on Monday and nobody notices until Friday?
Fix that first. The catalog can wait.
3. Dependency dressed up as advisory
The worst version is quiet. They do the work. They join every meeting. They become the only person who knows how the warehouse is wired.
You feel supported. You are stuck.
A fractional CDO who cannot point to someone on your payroll who will own the work is not advising. They are renting you a bottleneck.
Good advisory has an end in view. Not a dramatic exit. A handoff. Names. Runbooks. A week that still works when they are not on the call.
If the engagement only works while they stay, you did not buy leadership. You bought extra hands.
How to buy it without getting trapped
Do this before you sign.
Separate advice from implementation
Write 2 scopes. Advice is sequence, hiring, vendor selection, and what to stop. Implementation is pipelines, models, and licenses.
One firm can do both. Both should not hide in one line item.
If they implement the stack they also recommended, you review that recommendation yourself. Do not outsource the yes.
Demand a 90-day plan with nos in it
A useful plan names 3 outcomes you can check. It also names 3 things you will not start.
No new warehouse. No catalog program. No AI pilot. Those are valid nos if the company cannot trust last month's revenue number.
If every slide is a buy, it is a pitch deck.
Put an owner on your payroll
Every workstream needs a name that is not the fractional CDO. An analyst. An engineer. A product owner. Someone who will still be there on day 91.
If you have no such person, hire that person first, or admit you are buying hands, not leadership.
Time box the relationship
Set a review date. 90 days is a start. 6 months is plenty to see if the practice is real.
On that date you ask 3 questions.
- What decisions are better.
- Who on our team owns the work now.
- What we can stop paying for.
If the answers are that the roadmap is bigger, and that you still need them in the room, you have a vendor.
Watch the first 2 weeks
The first 2 weeks tell you the rest.
A leader maps the decisions, the data that feeds them, and the people. A vendor maps the products they know.
A leader asks who owns the metric. A vendor asks which cloud you prefer.
A leader will tell you to wait. A vendor will tell you to start.
What I would do this week
If you already have a fractional CDO, do not start with a feeling. Start with the contract and the calendar.
Read the statement of work. Circle every product name. Circle every task that only they can do. Those circles are your risk.
Then ask them to write, in one page, the 90-day nos and the names on your team who will own the work. If that page does not arrive, you already have your answer.
If you are about to hire one, use the same page as the interview.
A fractional CDO is not a vendor with a better title. Keep that test, or you will pay for both.