“I don't look for the channel first; I look for where your customer already is and whether you can measure what that channel actually does. At Cash App we built attribution and an MMM before we scaled, which let us compare paid search, connected TV, and podcast sponsorships on incrementality, not just last-click cost.”
“I don't start with channels; I start with the customer and the stage. At Cubii we proved unit economics on one paid channel before we built the DTC engine, and at CodaPet eighty percent of demand now comes from owned because we built the engine there deliberately. The right channel is the one your economics can support right now, not the one everyone else is buying.”
“I spent years at Airbnb pushing Facebook, Google, and every ad network to give us real lift-test tools, because without incrementality you're just reshuffling budget between channels that all claim the same conversion.”
“At Roku, the channel that changed our trajectory was not the one with the lowest CAC. It was the one that matched the exact moment someone decided to cut the cord. I always start there: where is the customer already making the decision your product solves, and which channel reaches them at that moment with the right signal.”
“I start every channel decision with a simple linear regression against weekly spend and new customers because it surfaces what the platform dashboards hide, including whether linear TV is actually moving the needle. The real work then becomes designing incrementality tests that match the account structure and mix you actually run, not the one where you measured it last year.”
“I stopped guessing which channel would work once I built a Marketing Mix Model at 7Mind: the podcast funnel alone was driving over €1M in annual revenue, and we had almost killed it because last-click made it look invisible. The real job is measuring the full journey, not chasing the channel that takes credit at the finish line.”
“I don't pick channels from a list; I run small, measured tests where the ICP already spends attention, then kill anything that can't show a path to a defensible CAC and payback. At Flanks, that discipline took organic traffic up 340% and cut blended CAC 47% by reallocating spend, not adding more.”
“I spent too long treating channels as a checklist, SEO, ads, email, social, and watched budget leak into places where our buyer never paused. Now I start from the buyer journey: which stage they're in, what question they're asking, and where they go to answer it. The right channel is the one that carries the specific message to the specific mind at the specific moment, not the one with the lowest CPC.”
“I start every channel test with paid because cold traffic tells you whether your message actually converts before you chase reach. At Wispr we built our whole program around that: 500-700 distinct creator ads a month, each filling a real gap, because Meta clusters anything too similar and one talking-head video becomes invisible next to a thousand copies.”