There is a kind of strategist who presents beautifully and cannot open a browser's developer tools. I have hired against that profile for years, and the reason is simple: in digital, the strategy and the plumbing are the same object. A media plan is a claim about what can be targeted, measured, and proven. Whether the claim is true is decided in the dataLayer, the consent state, the schema markup, and the API, and a strategist who cannot inspect those layers has to take everyone else's word for it. Vendors know exactly who in the room has to take their word for it.
That last sentence deserves expansion, because it is the economic engine of the whole problem. Ad tech is an industry where the seller's margin lives in the buyer's inability to verify, and every sales deck is calibrated to the least technical person with signing authority. The non-technical strategist is not merely at a disadvantage in those meetings; they are the product the meeting was designed for. I have watched the same vendor quote two clients materially different terms for the same product in the same quarter, and the difference between the quotes was precisely the difference in who on each side could read a log file. Technical literacy is not a nice-to-have on a strategy team. It is the negotiating position.
I minored in computer science at NYU alongside journalism, which at the time looked like a hedge and turned out to be the whole career. Not because I write production code for clients. Because I can read what the systems are actually doing, and reading is where strategy either survives contact with reality or quietly becomes fiction.
Three places it shows up
Measurement is the obvious one. Every number on a dashboard was produced by a chain of tags, parameters, and matching logic, and that chain fails silently and plausibly. Bad plumbing produces numbers that look slightly off rather than obviously wrong; the strategist who can trace an event from click to warehouse catches that in an afternoon. The one who cannot presents a quarter of fiction with confidence, and the confidence is the dangerous part, because it transfers. A room believes a presenter who believes the number, and nobody in the chain ever checked whether the number believed itself.
AI search is the newer one, and it is even less forgiving. Getting cited by ChatGPT, Gemini, or Claude is substantially a structure problem: schema, entity consistency, content the retrieval layer can parse and attribute. Most GEO wins live in the invisible layer of the page, which means the strategist who cannot read that layer literally cannot see the work, cannot audit it, and cannot tell a good GEO vendor from a rebranded SEO retainer, a distinction currently worth a lot of money in exactly the wrong direction. When I built the GEO practice for healthcare clients, the deliverables were markup audits and information architecture as often as they were content. Calling that an engineering task to be delegated is how strategy loses custody of the outcome, because the delegated layer is where the outcome is decided.
The third place is vendor negotiation. Ad tech pricing hides in technical language on purpose; the ISBA supply chain work showing a third of programmatic spend disappearing between advertiser and publisher was only possible because someone technical followed the log files. Every fee I have ever negotiated out of a stack was found by reading, not by relationship. The relationship got the meeting. The reading got the money.
The bar, and why the excuse expired
The bar is not becoming an engineer. The bar is technical literacy: comfortable in dev tools, able to read a tag container and a JSON payload, able to ask a vendor the second question. The gap between the first question and the second question is where the margin hides, because everyone asks the first question, the sales deck answers it beautifully, and the deck's authors are betting the room stops there. "What's your match rate?" is the first question. "Deterministic or probabilistic, measured against which identifier, on whose panel?" is the second, and the meeting changes temperature when someone asks it.
Here is the test I would give any strategy hire, and it takes ten minutes: open a site you know in dev tools, watch the network tab, and narrate what is firing and to whom. Not code. Narration. The candidates who can do it own their recommendations end to end. The ones who cannot are, structurally, presenters of other people's claims, however good the presenting.
And the excuse for not clearing the bar has never been thinner, because the AI tools I used to build this site will explain any stack to you, patiently, at two in the morning, for twenty dollars. The tag container you have been afraid of for five years will be explained line by line by a model that never judges the question. What used to require a CS minor now requires curiosity and a subscription, which means technical literacy has quietly moved from a specialty into a choice. The strategists who use the tools that way are pulling away from the ones who do not, and the distance compounds, because every stack they learn to read makes the next one faster, while the presenters-of-claims stay exactly where the vendors want them.