Should a CMO know how to code? The literal question is the wrong one
Should a CMO know how to code, in the sense of writing and shipping software? No. A chief marketing officer who spends the week in a code editor has stopped doing the job they were hired for, which is owning the revenue number and the story that gets you there. Nobody needs their marketing leader optimising a render cycle.
Asked properly, though, the question is a good one, and it goes like this: can this person build the thing they just described to me? Founders who raise the coding question are almost never asking about syntax. They've sat through the deck. They've seen the growth diagram with the neat arrows between the boxes. What they want to know is whether anything happens after the meeting that doesn't come with a vendor, a statement of work and two quarters attached.
Technical fluency in a marketing leader means four concrete things. They can read a data schema and tell you what actually joins to what. They can specify a workflow precisely enough that it can be built from the description. They can assemble a working version themselves in the current stack: an automation platform, a database, a model, an API key. And they can find the broken step at six in the evening before a launch instead of filing a ticket. None of that needs a computer science degree. All of it needs having built something.
What changed is the price of execution, not the definition of the role
For twenty years the marketing org was split down a hard line. Strategy sat on one side, execution on the other, and the distance between them was measured in budget and calendar time. A CMO decided, a team or an agency or an engineering backlog delivered, and the lag between the decision and the working thing was the single largest tax on the function. That lag is what made "strategic" and "hands-on" feel like opposite ends of a spectrum instead of two halves of the same job.
That gap has narrowed sharply. Workflow automation, data enrichment, content pipelines, competitive monitoring, lead routing: the work that required a developer three years ago is now assemblable by someone who understands the process well enough to describe it. Per McKinsey's State of AI survey published in May 2024, 65% of organisations reported regularly using generative AI in at least one business function, roughly double the share reported in the previous survey ten months earlier. The tools arrived fast. The leadership definition has not caught up.
Meanwhile the money moved the other way. Gartner's 2024 CMO Spend Survey put marketing budgets at 7.7% of company revenue, down from 9.1% in 2023, with CMOs reporting they are expected to deliver the same or more on less. When budget shrinks and expectation holds, the only remaining lever is output per person. Systems are that lever, which puts the ability to build them somewhere near the centre of the job rather than out at the edge with the other nice-to-haves.
There's a related number worth sitting with. Gartner's marketing technology research found stack utilisation at 33% in 2023, down from 42% the year before, meaning roughly two thirds of what companies had already paid for went unused. That is not a purchasing problem. It's what happens when the person choosing the tools cannot picture how they connect, so they buy capability instead of building a working system.
The bar, level by level
"Technical" is doing too much work as a word here, so it helps to break it into levels. Most hiring conversations go wrong because the founder and the candidate are silently talking about two different rungs of this ladder.
| Level | What they can actually do | What it costs you | When it's enough |
|---|---|---|---|
| 0. Directs | Sets strategy, briefs the agency, approves the work. Talks about the stack in brand terms. | Can't tell an honest estimate from a padded one. Every build becomes a procurement cycle. | Large org with a real marketing ops function and dedicated engineering time. |
| 1. Specifies | Writes a requirements document a builder can work from. Knows what data exists and what it can support. | Still dependent on someone else's roadmap for anything new. Iteration speed matches their queue. | You already employ the builders and their time is genuinely available to marketing. |
| 2. Builds | Ships a working workflow themselves: automation platform, database, model, API. Prototypes in days, hands over what proves out. | Needs the discipline to stop building once something works and hand it to someone who maintains it. | Under about 200 people, no marketing ops team, and speed of learning matters more than elegance. |
| 3. Engineers | Writes and maintains production code, owns architecture, reviews other people's pull requests. | Marketing judgment usually thins out at this end. You are paying CMO money for engineering hours. | Rarely. This is an engineer with a marketing interest, and the org chart should say so. |
Level 2 is the new bar, and I'll defend that position rather than balance it. A leader who can prototype the thing themselves gets to test an idea in a week that a Level 1 leader has to argue into a roadmap. Over a year that difference compounds into a completely different volume of learning, which is the actual currency of a marketing function that isn't yet sure which channel works. Level 3 is a different hire wearing the same title, and hiring it as a CMO usually means the company gets excellent infrastructure and mediocre positioning.
What to test for, and how
Interviews reward people who talk fluently about systems. Building one and describing one sound almost identical across a table, which is why the standard question set fails here. Three tests that separate them, in the order I'd run them.
1. Ask what broke
Have them walk you through a system they built, then ask what went wrong with it. People who genuinely built things remember the failure in detail: the field that silently changed type, the rate limit nobody read about, the automation that ran twice on a Sunday and emailed the same list. People who supervised the build have a clean story with no texture in it. The specificity of the failure is the signal, not the polish of the success.
2. Hand them your actual stack
Show them what you really run, including the spreadsheet everybody pretends isn't load-bearing. Ask what they'd connect first and why. A useful answer names a sequence and a reason, usually starting with wherever the truth about customers currently lives. A weak answer names tools. If they reach immediately for a new platform before understanding what data you already hold, you have a shopper rather than a builder.
3. Ask what they'd refuse to automate
This is the one that tells you whether the technical fluency is attached to any judgment. The right answer has edges: positioning decisions, the messages that go to your ten most important accounts, anything where being wrong at scale is worse than being slow. A candidate who wants to automate everything has confused capability with strategy, and that combination builds fast machinery pointed in an unexamined direction. The rest of the hiring frame, including how to scope the engagement, sits in how to hire a fractional CMO.
Where the builder CMO is the wrong hire
Sometimes it is, and pretending otherwise would be selling. If your positioning isn't settled, if you can't yet describe who buys and why in a sentence your sales team would recognise, then a systems-first leader will build you something efficient aimed at the wrong market. Systems amplify whatever they're given. Clarity gets amplified. So does confusion, and it gets amplified faster because nobody has to type it twice.
The sequence matters. Positioning first, then the motion, then the system that runs it. A leader who can build should still spend the first stretch of an engagement not building, which is genuinely hard for people who like to build, myself included. If the company hasn't done that work, the honest answer is that the first marketing hire question comes before the technical-fluency question, and treating them as the same decision is how founders end up with a beautiful pipeline of the wrong leads.
It's also worth noting that at real scale, the case weakens. A 400-person company with a marketing operations team, an analytics function and engineering capacity ring-fenced for growth work does not need its CMO in the build tools. It needs a CMO who can specify precisely and hold the strategy while other people execute. The Level 2 argument is an argument about companies where the distance between deciding and building is currently measured in months.
Why this is a moat and not a trend
The market has plenty of marketers who talk about AI and a much smaller number who have wired anything together and watched it run in production. That asymmetry is the whole opportunity. A leader who understands buyers, positioning and pipeline, and who can also build the system that acts on that understanding, compresses a handoff that most companies still pay full price for twice: once to decide, once to have it built.
The framing I use for this is deliberately narrow. A CMO who codes the systems, not just the slides. Not an engineer who learned marketing, which is a different and much more common animal. Marketing depth first, with enough technical fluency to make it operational. The full architecture of what gets built sits in AI marketing systems for B2B, and if the vocabulary is new the definitions live in the glossary.
The deck with the neat arrows is easy. Anyone can draw the system. The question worth asking in the interview is who is going to build it, and whether they have ever had one break on them at six in the evening. Ask that one first.
Keep reading: AI marketing systems for B2B · How to hire a fractional CMO · The AI marketing stack for lean teams · Marketing automation vs AI marketing systems
Frequently asked questions
Should a CMO know how to code?
Not in the sense of writing production software. A CMO who spends the week in a code editor is not doing the job they were hired for. The useful version of the question is whether they can build the systems they specify: read a data schema, define an automation precisely, assemble a working version in a no-code or LLM-based stack, and find the broken step when it fails. That level of fluency does not require a computer science degree. It does require having built something rather than only having commissioned it.
What technical skills should a modern CMO have?
Four, in practice. Data literacy: knowing what is stored where, what joins to what, and what a given number can honestly support. Systems thinking: describing a workflow precisely enough that it can be built from the description. Hands-on build ability in the current generation of tools, which usually means an automation platform, a database and a model wired together. And enough debugging instinct to locate the broken step instead of declaring the whole system unreliable. Notably, none of those four is a programming language.
Can a marketing leader learn to build AI systems without an engineering background?
Yes, and this is the part that has genuinely changed. Workflow automation, data pipelines and model calls used to require a developer. They now have interfaces a determined marketer can learn in weeks rather than years. What cannot be shortcut is the reps. A leader who has built three real systems and watched two of them break makes better architecture decisions than one who has read about it. The learning curve is short. The experience curve is not.
Should you hire a technical CMO or a CMO plus an engineer?
It depends on the size of the gap between deciding and building. Below roughly 200 people, with no dedicated marketing operations or data function, a leader who can build closes that gap without a headcount request, and the difference in speed is measured in quarters. Above that, with an ops team and engineering time already allocated to marketing, a CMO who can specify precisely is usually enough, because someone else is paid to build. Hiring an engineer to sit under a CMO who cannot specify the work is the expensive version of this mistake.
Not sure which rung your marketing leadership is standing on?
Every Focus4ward engagement starts with an audit. Two weeks to map what your stack actually does versus what you pay for, where the deciding-to-building gap is costing you quarters, and which systems are worth building first. Diagnostic first, no pitch.
Book a call