Ghost Engagement
Ghost engagement describes measured interactions such as page views, clicks or dwell time originating from non-human or non-buying actors, systematically distorting reach and performance metrics.
Ghost engagement refers to interactions that do not come from purchase-ready humans: agents, scrapers and automated previews generate clicks and views with no buying intent whatsoever.
Explanation
With research and shopping agents on the rise, the share of machine traffic grows sharply. In analytics these visits often look human: they load pages, follow links and spend measurable time on site. Yet they do not convert in the classic sense, because the purchase decision happens elsewhere or not at all. Teams that fail to separate this optimise campaigns toward an audience that never buys. What is needed is traffic segmentation by origin and behaviour, treating agent traffic as its own channel rather than noise, and shifting focus to downstream signals such as enquiries, quotes and closed deals instead of sessions.
Marketing Relevance
For performance marketing this is a budget question: counting ghost engagement overstates reach and substantially understates the true cost per real contact.
Example
A campaign reports 40 percent more sessions with unchanged enquiries. Segmentation shows the increase comes almost entirely from agent and preview traffic — cost per real lead has gone up.
Common Pitfalls
Blocking agent traffic outright is usually wrong, since it increasingly initiates genuine purchase journeys. Labelling it cleanly and evaluating it separately is the better approach.
Origin & History
Bot traffic is an old topic, but until 2024 it was mostly identifiable as crawlers. With user-delegated agents from 2025 a grey zone between human and machine emerged, for which the term ghost engagement took hold in 2026.
Marketing Use Cases
Segmenting agent, preview and human traffic across all reporting.
Evaluating campaigns by qualified enquiries rather than sessions.
Dedicated metrics for agent traffic as a potential precursor to real purchases.
Adjusted cost-per-lead targets after removing machine interactions.
Frequently Asked Questions
Is ghost engagement the same as click fraud?
No. Click fraud is deliberate manipulation for profit. Ghost engagement usually arises without malicious intent — from agents, preview services or monitoring — yet it distorts measurement just as much.
How do you identify machine interactions?
Through signals such as absent mouse movement, highly regular request patterns, unusual times of day, known agent identifiers and sessions with no interaction with dynamic elements.
Should agent traffic be excluded?
From performance evaluation yes, from the website no. Agents frequently research on behalf of real buyers, so their visits should stay served but be assessed separately.
Related Services
Go deeper: AI Search & GEO hub · AEO hub