March 2020. I moved my family to the US the same week COVID shut everything down. Within days, our DTC baby formula orders went 20x. No new inventory, no playbook, just a category behaving in a way none of our models expected.

That stretch taught me something I have been repeating in boardrooms ever since. It is also the thing I still watch most CPG teams get wrong.

We were all trained to think in funnels. Fill the top, push people down, measure conversion at the bottom. It is a tidy way to run a quarter. It is a poor way to run a category like formula.

Because formula, like a lot of CPG, is not a funnel. It is a flow.

A parent is relevant to a diaper brand for maybe 12 to 24 months. Then they age out. New parents arrive every day. Others leave every day. The category itself barely grows. So the real question was never "how do we convert more of the top." It was "who is in our audience right now, and how long do we have them."

Manage a funnel and you optimise for this quarter's conversion. Manage an audience and you optimise for three different things: who enters, how long they stay, and when they leave. You build the handover to the next category before you lose them, instead of mourning the drop-off after it happens.

The metric that captures this is not conversion rate. It is New-to-Brand: the absolute count of genuinely new people entering the brand over time. Growing that number is growing market share. Optimising conversion on a shrinking audience just means getting more efficient at emptying a smaller room.

Two very different corners of the market are circling the same idea right now. Nik Sharma, writing for the DTC crowd, put it plainly last week: the audience is the asset, and the 2026 winners "make things people want to watch, not just buy" (Limited Supply, June 28 2026). On the performance side, he also walked through why the gap between a $20M brand and a $100M brand is rarely first-mover luck. It is a deliberately built middle funnel. Eight Sleep spent roughly seven years building its reviews-and-comparison "Wall of Love" before it ever scaled ad spend (The Checkout Pod, June 25 2026).

Scrappy DTC language. Same lesson I learned selling formula. The asset is the audience, and the audience moves.

Here is where it hits the P&L. Most teams I work with still budget as if the audience sits still. It doesn't. If your media plan assumes a fixed pool you convert against, you are pricing relevance as a constant when it has a clock on it. That shows up as rising CAC you cannot fully explain, retention targets that quietly slip, and a "loyal" base that is really just aging in place.

Manage for flow and the calls change. You spend against entry, not only conversion. You measure dwell time inside the relevance window instead of only this quarter's cart. You engineer the handover, the diaper brand walking a parent toward the next thing the same company sells, before the clock runs out. And the audience you build compounds, because the reviews and comparison content that recruit humans this year is the same corpus the LLMs cite when the next shopper asks an AI what to buy.

One thing to do this week. Pull your category's actual relevance window, how long a customer is genuinely in-market for you, then map the flow against it: who is entering, how long they stay, when they leave. Not the funnel. The flow. If you cannot answer those three questions with a number, that is the meeting to book.

Most teams optimise the funnel because it is measurable this quarter. The brands that win are the ones who can see the clock.

-- Imteaz

Claude can be your analyst, editor, and strategist.
But most professionals are using it to fix grammar.

These 200+ Claude prompts take it from grammar tool to your most powerful AI work assistant.

Sign up for Superhuman AI and get:

Claim your free prompts