How to Measure ‘Attention’ When Defining Digital Markets

When courts consider whether different digital services “compete for attention,” they need concrete, legally defensible ways to measure attention. Below are pragmatic metrics and methods that antitrust lawyers, economists, and factfinders can use to assess substitutability and market boundaries based on users’ allocation of attention.

1. Core attention metrics (what to measure)

Time spent / dwell time: average session length and cumulative daily or weekly minutes per user on the service; useful to show the resource a service consumes relative to alternatives.

Active attention seconds: seconds of verified, on‑screen attention (e.g., viewability with active tab/gaze verification) rather than passive impressions.

Return frequency and inter‑visit interval: how often users return and how quickly they substitute away after disruption (captures habitual vs. peripheral uses).

Engagement actions: clicks, likes, comments, shares, messages sent — indicators of cognitive involvement and purpose (social interaction vs. passive viewing).

Scroll depth and completion/retention curves: where and when users drop off in feeds or videos; helps compare content formats and attention persistence.

Switching behaviors and costs: friction measured by account creation rates, data/identity portability, content migration effort, and multi‑account prevalence.

2. Data sources and collection methods

Combine multiple approaches for robust inference:

Server logs and telemetry: high‑scale, platform‑provided time and event data (subject to access and authentication).

Panel and device measurement: panels that track app usage across devices (helps when users split time across competing services).

Eye‑tracking and lab studies: small‑scale but precise evidence of gaze and attention for disputed content types or formats.

Surveys and stated‑preference tests: ask users where they would go if a service worsened — useful to show substitution intent when behavior data are ambiguous.

Ad exposure and advertising pricing: increases in CPM/CPC for advertisers after changes in a service can imply relative scarcity of attention.

3. Analytical approaches courts can use

Cross‑elasticity of attention: estimate how much attention for Service A shifts to B when A’s quality or availability changes (time‑series shocks, A/B experiments, or natural experiments work well).

Cluster analysis of usage baskets: identify groups of services commonly used in the same sessions or days to reveal functional substitutability (e.g., messaging + social feed vs. video app clusters).

Event studies: examine attention flows around discrete changes (feature removals, outages, policy shifts) to observe where users go.

Counterfactual modeling: simulate attention allocation under hypothetical degradations (useful where direct experimentation is infeasible).

4. Evidence standards and courtroom considerations

Triangulation: no single metric settles substitution — combine behavioral telemetry, panels, surveys, and experiments to meet the court’s burden.

Granularity: show both aggregate attention (total minutes) and user‑level patterns (how many users would switch) — antitrust analysis often requires both.

Temporal scope: analyze short‑run (session‑level) and long‑run (habit formation, switching costs) attention patterns; dominance can manifest differently across horizons.

Privacy and representativeness: document how measurement methods handle cross‑device identity, sampling bias, and consent so the court can assess reliability.

5. Common pitfalls to avoid

Relying solely on impressions or raw reach; ignoring multi‑tasking and background usage; failing to account for advertising‑driven attention versus user‑directed attention; and treating all attention seconds as equivalent across formats.

Using attention metrics lets courts capture competition that is not price‑based and reflects the real resource firms monetize. Properly measured and corroborated, attention evidence can show whether services are reasonable substitutes for users and whether a defendant exercises market power over the scarce resource of user time.

Sources

• How to measure and regulate the attention costs of consumer technology (Brookings Institution; 2021-11-04)
• Measuring User Engagement (book excerpt) (ResearchGate / academic authors; 2014-01-01)