A discussion with Dr Ali Goode a cognitive scientist, interrogating contemporary notions of attention and engagement in advertising and operational contexts.
The discussion differentiates looking from cognitive attention, criticises overreliance on proxy measures (eg, webcam eye-tracking), and sets out practical principles — fluent processing, single-message design, heuristic use, rapid initial engagement and sequenced communications — for motivating behavioural responses.
The conversation explores the implications for credit risk and collections, the limits of AI and synthetic data, and the need to codify experimentally validated communication patterns.
Find out more about Dr Ali Goode -> Here.
Key Take Aways
- Attention is multifaceted: visual fixation (looking) is not equivalent to cognitive processing (attention).
- Attentional load is shared across senses; auditory and internal cognitive demands reduce visual processing.
- Webcam eye-tracking often lacks the accuracy to infer processing or comprehension.
- Engagement should be defined as two outcomes: mental engagement (processing) and behavioural engagement (action).
- People are predictably irrational; do not assume rational, deliberative responses to communications.
- Simple, fluent messages increase ease of processing, liking, belief and uptake.
- Limit messaging: a single clear proposition outperforms multiple competing points.
- Heuristics (herding, loss aversion/FOMO) are reliable levers to prompt behaviour when used ethically.
- Novelty attracts attention by forcing sense-making, but must remain easy to understand to avoid friction.
- Initial commitment is time-sensitive — secure engagement within c. two seconds, then cascade information.
- AI delivers where rules and clear structures exist; synthetic personalisation risks mischaracterising chaotic human behaviour.
- Measure outcomes by behaviour (actions) not proxies (gaze); codify effective patterns via controlled experiments.
Innovation
- Fluent processing as a design principle: construct messages that are immediately easy to understand to increase acceptance and recall.
- Behavioural sequencing: design communications as a gated cascade — capture attention, simplify processing, then motivate a single action.
- Rule-based AI application: deploy AI where decision rules exist (emotion, memory models, attention scoring) rather than to simulate full human chaos.
- Controlled codification: systematically test and codify psychological principles that explain why specific communications succeed in operational contexts.
- Ethical use of heuristics: employ herding and loss-aversion nudges in ways that reduce avoidance and prompt constructive customer behaviours.
Key Statistics
- c. 2 seconds available to secure initial engagement before drop-off.
- 400 hours of ethnography cited in a McDonald’s creative example.
- The brain predicts the next 2–3 seconds as a fundamental timescale for perception.
- Three people walking onto a platform triggered broader commuter herding in an informal test.
- Brands have relied on FOMO over the last five to ten years.
- A historical exercise asked experts to predict the next 25 years and the panel were “all wrong.”
Key Discussion Points
- Distinction between perceptual attention (looking) and internal attention (processing).
- Shortcomings of eye-tracking, particularly webcam trackers, for inferring comprehension.
- Attentional load and executive function: multisensory competition for limited cognitive resources.
- Operational definitions of engagement: mental versus behavioural metrics.
- Human irrationality illustrated by the “drunk elephant” analogy and implications for messaging.
- Importance of message fluency and cognitive ease for rapid uptake.
- The power and limits of heuristics: herding, loss aversion and FOMO in prompting action.
- Role and risk of novelty: attraction vs. comprehension trade-off.
- Two-second rule and cascade strategy for communicating complex information across multiple touchpoints.
- Personalisation and synthetic data: benefits, overreach and AI limitations in modelling human behaviour.
- Memory as schematic binding rather than literal episodic recording; implications for recall and messaging design.
- Recommendation to prioritise behavioural experimentation and codify what works in sector-specific contexts (eg, credit risk).
#AliGoode
RO-AR insider newsletter
Receive notifications of new RO-AR content notifications: Also subscribe here - unsubscribe anytime
