Peter Farago has rewritten his company’s website five times in one year. Not refreshed, not tweaked: rewritten, because the positioning kept evolving, because the category kept moving, and because roughly 25 new competitors appeared in his corner of the AI developer-tool market in the same stretch. That pace would have signaled dysfunction a few years ago. Today it signals a marketer paying attention.

Peter, a high-tech marketing executive in the San Francisco Bay Area, and Steve Haney, a fractional go-to-market consultant in Silicon Valley, spent a recent episode of Misadventures in Marketing, the AMA San Francisco podcast, arguing that the traditional marketing playbook has stopped being a document you execute. Their conclusion compresses into one line from Peter that every marketing leader should sit their board down and read: “It’s not about scaling what works in marketing. It’s about figuring out what’s going to work in the first place.”

Positioning is now a process, not a tablet

The old model treated positioning as scripture. You ran the research, locked the document, and cascaded it into every asset for the next two years. Steve’s update on the podcast was blunt: a positioning document is no longer the ten commandments chiseled in stone, and the marketer’s real job is managing a living process rather than bringing tablets down from the mountain.

The forces breaking the old model are structural. Categories reshuffle as competitors launch and reposition weekly. AI has blurred the lines of authority inside companies, putting, in Peter’s phrase, lots of cooks in lots of different kitchens, with engineers and CEOs now contributing directly to copy, sites, and strategy. His own team abandoned Webflow for a headless CMS because even a fast agency workflow couldn’t cycle at the speed the positioning demanded. When your infrastructure decisions are driven by how often the story changes, the story has become the operating system.

Discovery is most of the job now

Steve put a number on the change that deserves circulation. Customer discovery, the work of recording calls, mining conversations, and learning what the product should even be, used to consume 20 to 25 percent of a marketer’s time. He now estimates it at closer to 70 percent. Marketing at an early-stage company has become a discovery function running three loops at once, in Peter’s framing: the narrative, the customer, and the distribution, each feeding adjustments to the others.

The classic frameworks still hold, they just cycle faster. Peter kept returning to STP, segmenting, targeting, positioning, and to a reframe he credits to Emily Kramer of MKT1: it’s not really product marketing anymore, it’s audience marketing. His condo example makes the point without jargon. The same unit sells to a new graduate, a couple expecting their first child, and a retiree, and every one of those audiences requires different benefits, channels, and comparisons. Pick the audience first and everything downstream follows, including whether you run a product-led motion or a sales-led one.

Focus is the differentiator AI can’t fake

The sharpest story of the episode came from competitive research. A colleague sent Peter a rival’s website with the note that it looked good, and at first read it did: fourteen pages of competent, engineer-friendly copy. Then the seams started to show on closer reading. Two snappy taglines were doing the same job on different pages. No clear answer emerged to whether this was one product or several, and nowhere was there evidence of a strategic choice. The team had generated volume without ever deciding who they were for, proving the oldest rule in the discipline: market to everybody and you’re marketing to nobody.

Peter’s own site is three pages that took weeks of arguing with his CEO, and he left the comparison more confident, invoking the old line about writing a shorter letter if given more time. Steve’s version of the same advice goes to his consulting clients directly: pick the beachhead, even something as narrow as school districts in Colorado, and turn away the Fortune 500 buyer who doesn’t fit. Canva started as a yearbook layout tool. AI shines a light into the dark corners of fuzzy messaging, and what it exposes fastest is a company that never made a choice.

Experiments are the new marketing playbook

What replaces the playbook, practically, is experimental velocity. Peter’s advantage now comes from how fast his team runs good experiments, reads the signals, and iterates, which requires what he calls purposefully naive thinking, a willingness to retest assumptions that failed in a previous era because the tool set underneath has changed. Investors haven’t lowered the bar, they still pattern-match on pipeline, customers, and unit economics. What has changed is how a marketer earns those numbers: first-principles problem solving instead of executing best practices from the last company’s binder.

Steve tied the moment to a larger reversal. After decades of STEM evangelism, leaders like Reed Hastings are publicly betting that liberal arts skills, storytelling, communication, judgment, will define the next generation of managers, because the job is increasingly directing people and AI agents toward problems worth solving. Taste and judgment, the two words Peter and his PhD-engineer CEO lean on most, aren’t in any playbook. That absence is rather the whole point. The marketers who thrive over the next few years won’t be the ones who memorized what worked. They’ll be the ones who got comfortable finding out.