The Unlearning Imperative: What Some of the Sharpest GTM Operators Are Letting Go of in 2026
The frameworks you've defended in QBRs are becoming the ceiling keeping you from what's next.
The most dangerous thing in your head right now isn’t what you don’t know.
It’s what you’re sure you know that no longer applies.
Every operator I respect is going through some version of the same reckoning right now. The frameworks that got them here are quietly becoming the ceiling keeping them from what’s next. The playbooks they’ve defended in QBRs, the mental models they’ve built careers on and now, they’re stress-testing all of it.
I asked ten of them to name it out loud. Here’s what they said, and what it means.
1. Traffic is not the proxy anymore.
Kevin Indig, Growth Advisor, put it plainly: “In 2026, I’m unlearning that traffic is the best proxy for growth. As AI answers more questions before the click, visibility, trust, and pipeline matter more than pageviews.”
This is the first domino. If your north star metric is sessions or pageviews, you are optimizing for a world that is actively dissolving. AI search doesn’t send traffic…it absorbs intent and returns answers. The companies winning in this environment aren’t chasing clicks. They’re engineering the answer. Visibility in the model is the new first-page ranking, and it doesn’t show up in Google Analytics.
2. Software as the default is over.
Aakash Gupta, Founder of Product Growth, said something that should make every SaaS vendor nervous: “I’m unlearning using software and doing everything with Claude Code instead. I even run my edits through there now.”
The assumption baked into most GTM stacks is that tools are the answer. Buy the platform, configure the workflow, train the team. That assumption is cracking. When a single AI interface can replace five tools, the question stops being “which software should we use” and starts being “which software still deserves to exist in our stack.”
3. Delegation is being redesigned from the ground up.
Ethan Smith, CEO of Graphite, named something most leaders won’t admit: “I am unlearning some forms of delegation. I am able to ask AI to do so much of analysis and work that I want completed, that I can do much of my work by myself. I still delegate, but less than I did before.”
This is uncomfortable because delegation has been a leadership virtue for decades. The executive who “lets go” is celebrated. But if AI can do the analysis, draft the memo, model the scenario — the math on what requires a human changes. This isn’t about replacing people. It’s about being honest that some of what we delegated wasn’t leadership. It was just work we didn’t want to do.
4. The org chart is a design problem now.
George Bonaci, VP Growth at Ramp, reframed the entire people management conversation: “I’m unlearning ‘span of control’ as a people problem when in 2026 the real skill is knowing which nodes on your org chart should be humans and which should be agents.”
Span of control, how many people a manager can effectively oversee, has been an organizational design constraint for a century. Bonaci is saying that constraint no longer applies uniformly. Some nodes on the chart should be agents. Which ones? That’s the new design question, and most organizations haven’t started asking it.
5. The marketing team structure is ending.
Jonathan Martinez, Founder of GrowthPair, didn’t hedge: “Old structure of a marketing team. It’s coming to an end, and we’re going to see an advent of new roles/org structure. Growth marketers -> growth engineers, media buyers -> creative strategists, etc.”
The job titles we’ve hired for over the last decade were built around a specific division of labor that AI is collapsing. Growth marketers who can’t build are becoming less valuable. Media buyers who can’t think creatively are becoming redundant. The operators who will matter in two years are the ones who are already expanding into adjacent skill sets — not waiting for a job description to tell them to.
6. Attribution will never explain everything. Stop waiting.
Simon Heaton, Director of Growth Marketing at Buffer, named the belief that has quietly paralyzed more marketing teams than any other: “I built my career on data-informed growth, but I’m unlearning the belief that attribution will eventually explain everything. In today’s fragmented discovery landscape, conviction, customer closeness, and a deep understanding of your market are becoming a bigger part of the job again.”
We’ve been waiting for the attribution model that finally solves it. It isn’t coming. The customer journey is too fragmented, too nonlinear, too influenced by channels that don’t pass UTMs. What replaces perfect attribution isn’t guessing — it’s judgment. Which requires being close enough to your customer to have earned an informed opinion.
7. The tool-switching reflex is done.
Eoin Clancy, VP Growth at AirOps, captured something I’ve felt but haven’t named cleanly: “In 2026, I’m unlearning the desire to log into individual tools when there’s a job to be done. From strategy to reporting, MCPs are helping me make decisions faster, and with more depth.”
The habit of opening a specific tool for a specific task — CRM for contacts, BI tool for reporting, doc for strategy — was trained into us by a decade of SaaS. That habit is becoming a tax. The operators moving fastest right now are the ones who’ve broken it.
8. Staying a beginner is the strategy.
Pranav Piyush, CEO of Paramark, said the quietest thing on this list: “My priors about how marketing works were built in a different world. AI is changing a lot of things faster than I can track, while some fundamentals are not changing at all. The unlearning I’m focused on: staying a beginner.”
This is harder than it sounds. Expertise is identity. You get paid for what you know. Choosing to hold that knowledge loosely — to approach a domain you’ve worked in for a decade with genuine beginner’s mind — takes real intellectual courage. It’s also probably the most durable competitive advantage available right now.
9. More is not more.
Uzair Dada, CEO of Iron Horse, said what a lot of CMOs are thinking but won’t say in a board meeting: “Unlearning past marketing frameworks and best practices. Unlearning that more content and more paid media does not mean more results. Unlearning that the old funnel is still relevant. It’s simplifying what we do as marketers and focusing on Getting Discovered and Getting Chosen.”
The volume playbook — more content, more spend, more touchpoints — was always a blunt instrument. AI has made it obsolete faster than anyone expected. Getting Discovered and Getting Chosen is a cleaner frame. It forces the question: are we actually showing up where our buyers are forming opinions, and are we actually the obvious choice when they do?
10. Channel mix is the wrong question.
Everett Butler, Head of Marketing at Lindy, closed the list with the sharpest line: “Your channel mix doesn’t matter. What matters is which product moments make someone pull out their phone and text a friend.”
After nine operators dismantling the infrastructure of modern marketing, Butler lands on what actually endures. Not the channel. Not the campaign. The moment. The experience so good it creates organic word-of-mouth. That hasn’t changed. That won’t change. Everything else is up for renegotiation.
The pattern across all ten.
Nobody named a channel. Nobody pitched a tactic. Nobody said “learn this new playbook.”
Every single one of them is in the middle of demolishing something they used to believe was true — traffic metrics, org design assumptions, attribution models, delegation instincts, tool habits, volume plays.
The operators pulling ahead right now aren’t learning faster.
They’re deleting faster.
The question worth sitting with: what are you still carrying that no longer applies?