Kill The Funnel
- Andrew Nilsen
- May 25
- 2 min read
The marketing funnel has been a useful fiction for decades. But fiction doesn’t play nice with literal cultures.
It's a useful shorthand. "That feels too low funnel" or "we need to go further up funnel" is easily understood in conversation. Keep it there.
The trouble starts when literal-minded operators - a critical group in every org - treat it as established science and build entire team structures, budgets, KPIs and processes around it.
Here's what breaks:
Org structures built around funnel stages end up siloed and pitted against each other for budget and resources. Learning isn’t shared, channels aren’t optimized fully (I’m doing X in Meta for the Upper team, but Y in Meta for the Low team, instead of I’m doing Z to max the return on Meta).
Creative testing gets narrow. "That's a performance message, it can't be in a brand execution." I've watched companies fight tests that ultimately yielded strong performance because “it didn’t make sense” to put a performance-oriented message in an upper funnel campaign or vice versa.
Measurement gets fractured. Upper funnel gets evaluated on awareness metrics or, at best, traffic and nothing else. Lower funnel gets evaluated only on last-click conversions. The total impact of either - which almost always crosses those lines - is invisible.
The funnel replaced one failure mode (thinking only about "sales right now") with a new one: turning a rough conceptual sketch meant to inspire into an owners manual to be followed to the letter.
We're not blending art and science. We're trying to turn art into science - where it ceases to be art.
The fix isn't a better funnel. It's knowing when to use the metaphor loosely and when to put it down entirely. I’m seeing metaphor increasingly unable to maintain the correct elevation.
This is a big risk as marketing operations become more automated. Part of the funnel’s utility has been in dividing up work. In a world where AI has made the need of dividing up work less important, creating source prompts ladled with assumptions about funnel positions is a huge missed opportunity.
Comments