Every marketer says the funnel is dead. Every ad platform still bills you by funnel stage. A new explainer from PPC Land lays out that contradiction in detail, and it is worth a hard read if you own budget.
The funnel is arithmetic, not a story
Strip away the awareness-consideration-intent poetry and a conversion funnel is one thing: a count of how many people survive each ordered step toward a defined action.
Its value is in decomposition. PPC Land uses a clean example: 100,000 sessions produce 4,000 product page views, 1,200 add-to-carts and 360 purchases. Aggregate conversion rate is 0.36 percent, which tells you nothing. Stage rates of 4 percent, 30 percent and 30 percent tell you the leak is at the top, not in checkout.
That distinction is a budget decision. A top-of-funnel failure is a targeting and creative problem. A bottom-of-funnel failure is a site, price or fulfilment problem. Different owners, different money.
The setting that silently breaks your comparison
One configuration choice matters more than the rest: closed versus open funnels. A closed funnel only counts people who entered at step one. An open funnel counts anyone joining midway.
Google Analytics 4 supports both, defaults to closed, and allows up to 10 steps and four segments per funnel exploration. Open funnels report higher completions. Compare an open funnel this quarter against a closed one last quarter and you have manufactured a trend that exists only in your settings.
Event naming is the other silent killer. Google’s recommended ecommerce sequence runs view_item_list, select_item, view_item, add_to_cart, view_cart, begin_checkout, add_shipping_info, add_payment_info and purchase. Custom names do not feed the built-in checkout report, and a missing step shows up as silence rather than an error. Two GA4 and server-side GTM defects documented in September 2025 inflated counts via service worker duplication, corrupting exactly the ratios funnels depend on.
Why the linearity critique keeps growing
The challenge is not new. McKinsey’s June 2009 research across roughly 20,000 consumers found buyers adding brands during evaluation rather than steadily narrowing, and proposed a circular decision journey with a loyalty loop. Google’s 2020 Decoding Decisions report described the messy middle, where shoppers cycle between exploration and evaluation under six cognitive biases.
Then the platforms joined in. At Google’s NewFronts on 23 May 2025, executives pitched streaming, scrolling, searching and shopping as a replacement for stage thinking, citing roughly 80 percent of consumers skipping paths. MiQ’s April 2026 study across 53 million households found 72 percent of under-34s completing entire purchase journeys inside social apps.
And yet the buying interface never changed
TikTok launched Brand Consideration Ads on 13 May 2026 as an explicit mid-funnel objective. IAB Tech Lab’s ECAPI 1.0 classifies page_view, ad_impression and search as upper-funnel events. Google’s Journey Aware Bidding feeds intermediate events into models when purchases take seven to fourteen days. Amazon extended conversion path reporting worldwide on 13 November 2025.
Meanwhile upper-funnel proof is getting harder. Meta removed 7-day and 28-day view-through windows from the Ads Insights API effective 12 January 2026. Google dropped Display and Video support from Performance Planner on 9 March 2026.
What to actually do
- Lock your funnel config. Document closed vs open and step definitions before any period-over-period comparison.
- Audit event names against Google’s recommended ecommerce sequence. Silence is not zero.
- Fix the checkout before buying more traffic. Baymard puts average cart abandonment at 70.22 percent. DHL research in August 2026 found 67 percent of shoppers abandoning over delivery terms while only 52 percent of sellers spotted that cause, with the payment page the top drop-off point at 32 percent.
- Check your recovery emails. Hesse’s data protection authority ruled abandoned-cart emails unlawful in June 2025, treating them as advertising requiring GDPR consent.
- Pair funnels with incrementality. Stage rates are correlational; they record what happened, not what your ads caused.
One more structural warning: agentic commerce. When an AI agent negotiates checkout through backend interfaces, merchants see completed orders or hard failures, not partial progress. Browse-to-cart measurement loses its subject entirely.
Treat the funnel as what it is: a diagnostic ledger and a buying interface, not a description of how humans shop.
Source: PPC Land
