How purchase decisions are actually made - not how customers say
Traditional methods ask customers how they buy. The problem is that people don't remember, rationalize after the fact and omit key steps. Reflect observes actual online purchase behavior in real time, classifies decision narratives with AI and uses experimental design to identify what actually drives decisions.
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.
Why funnel models fail
The funnel model assumes customers move linearly from awareness to purchase. In reality they jump back and forth, leave the funnel, return via a different channel and make decisions based on factors the funnel doesn't capture.
AI classification of purchase narratives - how and why
Open-ended questions about purchase decisions yield rich data but are time-consuming to analyze manually. AI classification makes it possible to identify decision themes, motivation and barriers in thousands of narratives with consistency that manual coding cannot match.
Experimental design in journey research
Observation shows what customers do, but not why. By combining observation with experimental manipulation - changing price, channel experience or information availability - we can isolate what actually drives decisions from what merely correlates.
Reflect's journey analysis framework
Reflect's customer journey framework combines three methods: real-time observation of online behavior, AI classification of decision narratives and experimental design. Together they give a picture of how customers actually buy - not how they say they do.
Optics chain mapped the digital purchase decision
A JDE study for a Swedish optics chain revealed fundamentally different purchase processes for contact lenses and eyewear, and identified conversion barriers that never appeared in traditional surveys.