How online purchase decisions actually work
Online purchase decisions are not the linear funnel most models assume. They are iterative, chaotic processes where customers jump between channels, compare, abandon and return — often without being aware of the pattern themselves.
The funnel model has dominated marketing for decades: awareness, consideration, purchase. It is intuitive and easy to communicate. The problem is that it does not match how people actually behave online.
Google's "messy middle" research showed that the purchase process is more of a loop between exploratory and evaluative behavior. But even that model simplifies. In reality we see behaviors that don't fit any model: customers who start with price comparison before knowing what they want, who switch channels three times in a session, or who make decisions based on a single review after weeks of research.
Reflect's Journey Decision Engine captures these behaviors in real time. Instead of asking customers how they bought, we observe what they actually do: which sites they visit, in what order, what they compare, where they get stuck and what ultimately triggers the purchase. The patterns that emerge are almost always different from what customers themselves would have reported.
The implication for marketing is fundamental: if the purchase journey is not a funnel, we cannot optimize it as a funnel.
Key takeaways
- The funnel model rarely matches actual online behavior
- Purchase processes are iterative with frequent channel switches
- Self-reported journeys omit critical decision points
- Real-time observation reveals patterns interviews miss
- Optimization must start from actual decision points, not assumed ones