Applying a single attribution model, like last-touch, to all channels is a mistake. It undervalues top-of-funnel activities and can lead to budget cuts that starve the pipeline. Instead, measure each channel based on its intended outcome and funnel stage.

Since platforms like Google and Facebook have a vested interest in overstating their impact within their "walled gardens," a simple, qualitative approach can be more revealing. Adding a "How did you hear about us?" field to your forms provides direct, self-reported data from customers, helping you identify influential channels that complex models might miss.

By measuring success on 'last lead source,' the company was incentivized to pour money into paid search for product trials—a clear final touchpoint. This model blinded them to the higher value of other lead types and actively discouraged investment in demand creation activities that build brand and generate higher-quality leads.

Relying on last-touch attribution creates a feedback loop that over-invests in bottom-of-funnel channels like branded Google search. This model fails to account for the preceding marketing actions that prompted the search, misallocating budget away from crucial brand discovery activities.

The question modern attribution should answer is not "Which channel gets credit for this dollar?" but "What are the commonalities across our most successful buying journeys, and how can we replicate them?" This moves from a simplistic, linear view to a more holistic, pattern-based understanding of customer acquisition.

Go beyond standard W-shaped or last-touch attribution models. Create "influence reports" that measure the sheer frequency a channel appears in any revenue-generating journey. This provides a different lens, showing which channels are consistently present and influential, even if they don't get direct attribution credit.

Don't abandon attribution; evolve it. The old model of single-touch software attribution is outdated. A modern approach triangulates data from software (GA4), self-reported forms ("How did you hear about us?"), and conversational intelligence tools, using AI to identify common buying journey patterns.

CloudPay stopped attributing opportunities to single sources like "marketing" or "sales." Analysis showed multiple departments influenced every deal, rendering attribution a source of pointless internal arguments. They still use multi-touch attribution at the campaign level, but not to assign inter-departmental credit.

Relying on a single data point like "first touch" to explain pipeline creation is flawed. It ignores the complex buyer journey and inevitably leads to a blame game—marketing providing "shitty leads" versus sales doing "poor follow-up"—instead of a systematic analysis of what is truly broken in the process.