Taste is the most valuable skill in marketing

"Taste" has almost become a cliche at this point. Everyone keeps talking about "taste" being the skillset that we all need to develop. What on earth does that even mean?

Taste sounds soft. It is not. Taste is years of pattern recognition, the same kind of compounding judgment that April Dunford describes in her positioning work or that Elena Verna applies when she re-finds product-market fit every three months at Lovable. It is knowing which customer insight matters and which is noise. It is the instinct that tells you to zig when everyone else zags. It is looking at ten AI-generated headlines and picking the one that actually works.

AI commoditizes execution. Anyone can generate a blog post, design a landing page, write email copy, build a campaign brief. The tools are free or close to it. The output is competent. Sometimes it is good.

When everyone's output clears the bar, the questions change. "Should you produce this?" and "is this the right version?" AI can write a hundred positioning statements in ten minutes. It cannot tell you which one is right.

I have hired hundreds of marketers over my career. The best ones were never the fastest writers or the most technically proficient. They were the ones with judgment. They could look at a positioning statement and tell you what was wrong with it in ten seconds. They could read a landing page and feel that the story was off before they could articulate why.

Last month, I created my own agentic framework and my own version of Granola. I exhibited horribly poor taste in doing so. Not because the products aren't cool or beautiful. On the contrary, they're both wonderful. But I wasted time (and tokens) on both those projects when off the shelf systems were more than adequate for my purposes.

In a world where you can do anything in hours, soon to be minutes and seconds, the ability to be discerning about where you apply your energy and thought process is far more valuable.

As the execution gap closes, the judgment gap widens.

Has technology killed creative skills before?

Technology has democratized marketing execution in every major cycle. And every time, the people who predicted that the old skills would die were wrong. The skills got more valuable because there was more output that needed quality control. This pattern has repeated with desktop publishing, the web, social media, and now AI.

Desktop publishing in the late 1980s. Aldus PageMaker shipped and suddenly anyone with a Macintosh could lay out a brochure. Before that, you needed a typesetter. You needed a paste-up artist. The cost of a single marketing flyer ran into the hundreds of dollars.

PageMaker changed all of that. The receptionist could make a flyer. The sales manager could design a brochure. Everyone predicted the death of graphic designers.

You know what happened. Everyone made flyers. Most of them were terrible. Comic Sans on a teal background with clip art of a handshake. Twelve fonts on a single page. Text crammed into every available space because the person making it had no concept of white space or visual hierarchy.

Graphic designers did not get replaced. Bad designers got replaced. The designers with taste, the ones with an eye for composition, color, and typography, became more valuable. Now every company was producing ten times more printed material and somebody had to make sure it did not look like garbage.

Design spending increased after desktop publishing, not decreased. The bar rose. Companies that had never hired a designer before suddenly needed one.

Every time some Twitter rando with a blue checkmark prattles on about "designers are cooked, bro," I have to laugh. Sure, there are AIs that move pixels around. But who is going to tell you which pixels to move and to where and in what color? Come on. Let's give humanity a little more credit.

The web did the same thing in the 1990s. Everyone had a website. The companies that invested in great web experiences pulled away. Social media did it again in the 2010s. Everyone had a Twitter account. The companies that invested in authentic voice and real community built audiences the "just post regularly" crowd could never touch.

I wrote about this in good marketing in the AI era. Every technology cycle raises the floor. Every one also raises the ceiling higher. AI is this pattern at a scale we have never seen.

Why are "AI skills" not the answer?

AI skills are the most common career advice in marketing right now, and that advice is wrong. "Learn to prompt. Build an AI stack. Hire people who understand AI workflows." The hottest job titles are "Head of AI" and "AI Marketing Strategist." But AI tool skills commoditize almost immediately. Today's advanced prompting technique is tomorrow's default feature.

I am not saying AI skills are useless. You should know how to use AI in your marketing workflow. Your team should understand the tools. But the model gets smarter and the prompting gets simpler. Every release cycle erases the advantage of the people who spent months learning the previous version's quirks.

Taste does not commoditize. It compounds. Ten years of shipping campaigns and watching what works builds an instinct that no amount of AI training can shortcut.

60-70% of traditional growth tactics no longer apply for AI companies. Teams re-find product-market fit every three months. In that world, the person who can evaluate options fast and choose well is worth ten people who can generate options fast.

Knowing how to prompt well matters less than knowing which output to ship and which to throw away.

What does taste look like in practice?

Marketing taste shows up in concrete decisions every day.

It is the product marketing manager who reads a draft positioning statement and says: "This is accurate but it is boring. We sound like everyone else." And then rewrites the opening line.

It is the content lead who looks at ten blog post ideas generated by AI, kills eight of them, and explains why the remaining two are worth writing. Not because of keyword volume. Because they say something only this company can say.

It is the developer advocate who watches a demo recording and says: "Skip the first two minutes. Start with the moment where the query returns in 40 milliseconds. That is the hook."

It is the head of marketing who reads a competitive battlecard and says: "We are being too defensive. Reframe this around our strengths, not their weaknesses."

None of these decisions can be automated. All of them require experience. All of them make the difference between marketing that works and marketing that fills a content calendar.

How do you build taste?

Taste is built the same way a sommelier builds a palate: reps. A sommelier does not have a magical tongue. They have tasted thousands of wines and built a mental database of what good tastes like. Marketing taste works the same way.

Ship campaigns. Read the data. See what worked and what did not. Do it again. Read great marketing and bad marketing. Study the difference. Work with people who have more experience than you and watch how they make decisions.

Taste is not mysterious. It is not innate talent. It is developed through repeated exposure to quality and its absence.

The shortcut, if there is one, is to use AI to increase your reps. Automate the repetitive parts of your work so you have more time to evaluate, decide, and learn from outcomes. Use AI to generate ten versions so you can practice choosing the best one. Use AI to produce first drafts so you can practice editing.

The people who use AI to skip the work will not develop taste. The people who use AI to do more work, to see more options, to iterate more cycles, will develop taste faster than any generation before them.

What does this mean for hiring?

Hiring changes when taste is the bottleneck. AI skills become table stakes, like knowing how to use a spreadsheet. Everyone will have them. They will not differentiate.

Companies will start hiring for judgment, editorial instinct, and brand sensibility. They will look for marketers who have shipped enough campaigns to know what failure looks like before it happens. Who can evaluate AI output the way a senior editor evaluates a junior writer's draft.

This is uncomfortable for people who have invested heavily in AI expertise as their primary career differentiator. But the pattern is consistent. In every technology cycle, the tool-specific skills commoditize. The judgment skills compound.

What am I telling my marketing teams?

I run marketing organizations. This is what I tell my teams.

Everyone learns the tools. Non-negotiable. If you cannot use AI effectively to produce first drafts, research, and generate options, you are operating with one hand tied behind your back.

Nobody ships without editorial review. AI-generated content goes through the same review process as human-generated content. Actually, stricter. AI content that looks polished can hide shallow thinking. A human writer who does not understand the product will produce visibly bad content. AI will produce invisibly mediocre content. That is harder to catch.

We centralize strategy and decentralize execution. The positioning framework, the messaging architecture, the brand voice, the quality bar: those come from the center. The production of assets, the creation of variants, the localization, the adaptation to channels: that can happen anywhere, with AI, within the guardrails.

We invest in taste. I send my senior marketers to more customer conversations, not more AI workshops. I want them in the room with customers, hearing how they describe their problems, learning the language of the market. That is how you build the judgment to evaluate whether AI-generated content actually resonates.

What is the punchline on taste?

When I started in marketing, we had to harness fire, chisel our value props into stone, and distribute them by mammoth. Fine. We had FrontPage and a fax machine. But I have seen this movie four times now. The technology changes. The plot does not.

Invest in taste. Centralize your strategy. Let AI do the typing. And when the smart people in the room tell you that "AI strategy" is the future of marketing, smile politely and ask them: who decides what the AI should build?

That person is the one you cannot replace.