This is an attempt to build a working model of human attention — not as a metaphor, but as a structure with measurable layers. How categories compete for a fixed budget. How brand share shifts within a category. And what actually converts attention into purchase intent.
Think of a person's total attention across a year as 100%. It doesn't expand. When one topic surges — say, cricket during IPL — it doesn't add to the total. It compresses everything else. The chart below shows this as a stacked area: the bands always sum to 100%.
Topics follow three distinct patterns. Some are dormant — always present at low intensity, like coffee or fountain pen inks for enthusiasts. Some are spot-to-dormant — they peak with an event or season, then settle back to a baseline (cricket during IPL; the monsoon). Others are spot-to-exit — they spike with a news event or launch, then permanently disappear from the person's roster (a geopolitical crisis; a viral trend).
Always occupies some share, never fully exits. Low intensity, high persistence.
Coffee · fountain pen inks · running
Peaks with events or seasons, then settles to a baseline rather than disappearing.
IPL cricket · monsoon · Diwali shopping
Bursts with a trigger — news, launch, trend — then drops to zero permanently.
Iran war · phone launch · viral moment
"When the Iran war story dominated attention in February–March, it wasn't that people stopped caring about coffee or cricket. Those topics compressed. They re-emerged the moment the crisis faded — because they were dormant, not gone."
A brand whose category is dormant still has an audience — it is a small, stable audience that is primed to re-engage when triggered. A brand whose category only shows up seasonally has a narrow window. A brand in a spot-to-exit category should not expect loyalty mechanics to hold; the interest itself has left.
The fixed budget applies not just to how much attention each topic receives, but to how many topics are actively held at once. Research and intuition converge on a similar number: roughly 30–40 open "attention relationships" at any given time, of which only 5–10 are foregrounded — actively receiving attention in the current week.
The outer ring of dots represents the dormant roster: topics the person genuinely cares about, which exist in memory but are not competing for attention right now. The inner named bubbles are the foreground — where active mental processing is happening. Switching the time snapshot shows how the roster rotates through the year without growing or shrinking.
A brand acquiring a new email subscriber is essentially asking for one of the person's 35 roster slots. That's a non-trivial ask — it competes with established topics. But a brand whose category is already in the roster just needs to be the best option within it.
Category attention tells you when a person is thinking about a topic. Purchase intent tells you when they are likely to act. The two are correlated — but not identical. The gap between them is created by financial and life triggers: an annual appraisal, a salary credit, a birthday, a festive season, a bull market.
The chart below overlays category attention (the shaded band) with purchase intent (the solid line). The dashed vertical lines mark when a trigger fires. Notice how intent can spike well above attention at trigger moments — and how in some categories (especially finance), an external condition like market sentiment can collapse intent entirely regardless of personal triggers.
The monthly pulse below applies universally: credit card statement pressure in days 1–5 suppresses readiness across all categories; salary credit around day 25–28 creates a reliable peak.
"A brand reaching someone on day 3 of the month vs day 26 is talking to the same person with the same category interest — but in completely different financial readiness states. The conversion difference can be substantial. The creative doesn't change. The channel doesn't change. The person does."
The three layers multiply. A person in peak category attention (layer 1) × whose favourite brand is salient (layer 2) × who just received their salary and got an annual raise (layer 3) is at maximum purchase readiness. Email engagement history, salary-cycle inference, and declared life events (birthday, anniversary) give a sender all three signals simultaneously. No paid channel sees this combination.
The marketing industry has spent decades building tools to solve this problem. Share of Voice measures how loud you are relative to competitors. Eye-tracking and attention metrics — the newer wave, popularised by firms like Lumen and Adelaide — measure whether anyone actually looked. Thales Teixeira at Harvard Business School went further, showing that the cost of consumer attention has risen sevenfold in two decades and arguing that attention itself should be treated as the currency brands are actually buying. McKinsey's attention quotient segments audiences by how willing they are to engage with media at all. All of it useful. None of it answers the question that actually determines whether a communication converts: was this person in-market for this category, right now, in this week of their month?
SOV is a market-level metric. Attention metrics are creative and placement metrics. McKinsey's quotient is a demographic segment. What they share is that they operate on the channel — the ad, the placement, the format. None of them operate on the individual's attention state at the moment of receive. That state is what this model is trying to map.
A predictable objection: the industry already knows attention peaks at certain moments. Payday treat campaigns. Diwali bursts. End of season sales timed to bonus cycles. True. But that knowledge is a population heuristic, not individual intelligence. It describes when the pond gets crowded — when enough people have enough disposable income at broadly the same time. It does not describe where any specific person sits in their own attention cycle on a specific Tuesday in February. Everyone shows up at Diwali. The media gets expensive. The noise floor rises. You are competing with every brand that read the same calendar.
The distinction matters. The industry has learned to read the collective clock. This model reads the individual one. Diwali is a hand on the collective clock. Priya's appraisal is a hand on hers. They do not always point in the same direction — and when they don't, the Diwali campaign lands in a dormant state regardless of how well-timed it looks from the outside. The calendar-based heuristic is a symptom of not having individual-level signals. If you knew who was in-market at any given moment, you would not need Diwali as a proxy. You would send to the ready and hold from the dormant — irrespective of the season.
Here is the problem made concrete. A brand sends one communication — a TV spot, a Meta ad, an email — to its entire audience on the same day. It lands simultaneously. But the three people below are in completely different places: different category interests, different salary dates, different life events this month.
This is why mass communication underperforms relative to its reach. Not because the creative is wrong or the placement is cheap. Because timing and readiness are invisible to the sender. Engagement history, salary-cycle inference, and declared life events make that invisible layer visible.
Loading...
Each of the four layers has a body of research behind it. This is not a complete literature review — it is the minimum set of citations needed to show that the model rests on established findings, not intuition.
The foundational framing comes from Herbert Simon (Nobel laureate, 1978), writing in 1971 — before the internet existed. He argued that a wealth of information creates a poverty of attention, and that the right problem to solve is not information scarcity but attention scarcity. Daniel Kahneman's resource model of attention provides the neural grounding: one central attentional pool, flexibly allocated, but hard-capped in total capacity.
The spot-to-dormant and spot-to-exit patterns in the model are measurable at population scale. A 2019 study in Nature Communications analysed Twitter, Google Books, Wikipedia, Reddit, and cinema data — and found that topic peaks are becoming steeper and shorter over time as information volume increases.
The 5–10 foreground / ~35 total roster structure is inferred from Dunbar's layered model, not directly measured. The inference runs as follows. Dunbar showed that the neocortex constrains our capacity to maintain ongoing cognitive relationships — tracking state, history, and salience of another entity — in a tiered structure: ~5 intimate, ~15 close, ~50 meaningful. An interest relationship makes the same cognitive demands as a social one: to actively follow cricket, you track current standings, recent results, upcoming fixtures, and emotional investment. The neocortex does not distinguish "tracking a person" from "tracking a topic." On this basis, the same tier structure likely applies: ~5 topics foregrounded at any moment, ~15 regularly checked, and ~35–50 in the full active roster. This is structural inference, not direct measurement — but it is grounded in the same neocortical capacity constraints Dunbar documented.
The monthly salary cycle is one of the most replicated findings in consumer behaviour research. Multiple studies using bank transaction data confirm that spending surges the day after payday, and that the type of spending also shifts — promotion-focused (aspirational) near payday, prevention-focused (maintenance) as the month ends. This is not a preference change; it is the same person in a different financial state.
The charts in this document are illustrative — directionally accurate, not empirically measured. The specific numbers (35 topics, brand share percentages) are models, not data. The research above supports the structure of the model: that attention is fixed and layered, that topic salience cycles predictably, and that purchase intent is modulated by financial timing independent of category interest. The precise calibration is an empirical question.