From Marketing to Product Manager: The Route AI Makes Easier

I started in Corporate Comms and Marketing at Nokia, then ran the web shop at easyJet, where the shop was managed as a product and Marketing was a close partner. Two more marketing roles followed, at MORE TH>N and EDF, both as the Digital Marketing leader, before I set up BlueLab at EDF, the first time I worked somewhere that was product-first. That is one route from marketing to product manager, among many.

I wrote in 2023 that marketers were moving into product management. The move is easier to make now, because AI is automating the execution side of marketing. Marketing Week reported that 18% of the 1,300 marketing leaders Gartner surveyed have already eliminated early-stage marketing roles, and the AMA’s 2026 career report found senior and strategic roles are holding steady while execution-focused roles decline. A role that is only a channel for something else is the one at risk, which is the question I put to learning leaders in The Travel Agent Test.

What carries over

The marketers I have seen become product managers had an appreciation of the market (the clue is in the name). They brought an understanding of the realities of selling a product that people who have only built them did not have. Add the habit of reading a funnel and running a test, and you have the depth that product work needs next to it: a T-shape that runs in this direction.

The arm’s-length problem

Many marketers work at arm’s length from the execution: managing agencies, writing briefs, and blaming the agency when things don’t pan out. Product is the opposite. You do the work, you sit beside the other people who do as a team, and the outcome is yours. Cagan has argued since 2007 that product management and product marketing are different jobs, and that one person who can do both is hard to find. So the product half cannot be assumed to carry over. In practice that means turning up for the unglamorous parts: the trade-off conversation with engineering, the no to a stakeholder with a roadmap, the bug report on something you shipped. You have to show it.

The evidence a hiring manager can check

Five things, in the order I would do them:

• Pick a problem you already own. A real customer problem, where you can see the data and nobody else holds the pen. A hypothetical does not count, and neither does one an agency owns.
• Make a decision with a trade-off. Say what you chose, what you gave up, and what would make you stop. A decision with no cost is a preference.
• Build a prototype with AI. It takes a day now, not a quarter, so a working version is a fair thing to ask of a marketer.
• Put it in front of real users. A handful of conversations or a small test is enough. A prototype nobody used shows you can build, not that you can decide.
• Write down what you learned and what you would change.

Nikhyl Singhal writes that interviewers now probe what failed, which trade-offs you made, and what you would change, and that current hands-on experience is the credential. The prototype is the cheap part: the written decision around it is the evidence.

There is no single path

An internal move, a hybrid role such as product marketing with real ownership, or an application from outside: there are many variations, and I do not think any one of them is the path. Take the one where your evidence is already in the room, because the market is hard. Mind the Product reported that product manager headcount at US tech companies is still about 28% below its 2022 peak. Nobody publishes a clean figure for how many product managers began in marketing, so I am going on the people I have seen make the move.

If you are hiring

Judge a marketer for a product role on one small piece of real work, not the CV: the same argument I made in Pass the Test or Master the Craft. Ask three questions:

• What product decision did you make, and what did you give up?
• Show me what you built, and what broke.
• What would you change?

Go through the portfolio and ask why each piece is the way it is. The answers show how the person thinks and how they decide: with data, or on a vibe. A working prototype tells you little; the reasoning around it tells you most. If the candidate cannot show a decision yet, an internal move or a hybrid role is a cheaper way to find out.

Pick one customer problem you already own, and ship the smallest decision you can defend.