dorsal/arxiv
View SchemaConsumers don't play dice, influence of social networks and advertisements
| Authors | Robert D. Groot |
|---|---|
| Categories | |
| ArXiv ID | physics/0508065 |
| URL | https://arxiv.org/abs/physics/0508065 |
| DOI | 10.1016/j.physa.2005.08.023 |
Abstract
Empirical data of supermarket sales show stylised facts that are similar to stock markets, with a broad (truncated) Levy distribution of weekly sales differences in the baseline sales [R.D. Groot, Physica A 353 (2005) 501]. To investigate the cause of this, the influence of social interactions and advertisements are studied in an agent-based model of consumers in a social network. The influence of network topology was varied by using a small-world network, a random network and a Barabasi-Albert network. The degree to which consumers value the opinion of their peers was also varied. On a small-world and random network we find a phase-transition between an open market and a locked-in market that is similar to condensation in liquids. At the critical point, fluctuations become large and buying behaviour is strongly correlated. However, on the small world network the noise distribution at the critical point is Gaussian, and critical slowing down occurs which is not observed in supermarket sales. On a scale-free network, the model shows a transition between a gas-like phase and a glassy state, but at the transition point the noise amplitude is much larger than what is seen in supermarket sales. To explore the role of advertisements, a model is studied where imprints are placed on the minds of consumers that ripen when a decision for a product is made. The correct distribution of weekly sales returns follows naturally from this model, as well as the noise amplitude, the correlation time and cross-correlation of sales fluctuations. For particular parameter values, simulated sales correlation shows power law decay in time. The model predicts that social interaction helps to prevent aversion, and that products are viewed more positively when their consumption rate is higher.
{
"annotation_id": "99decb66-8bf8-41cd-a500-4eeb6bd0d36f",
"date_created": "2026-03-02T18:01:00.136000Z",
"date_modified": "2026-03-02T18:01:00.136000Z",
"file_hash": "ed0322d3fa557dde3822e8f1868be52b1ace6405e1f8dd8e53ad5489c0e4e733",
"private": false,
"record": {
"abstract": "Empirical data of supermarket sales show stylised facts that are similar to\nstock markets, with a broad (truncated) Levy distribution of weekly sales\ndifferences in the baseline sales [R.D. Groot, Physica A 353 (2005) 501]. To\ninvestigate the cause of this, the influence of social interactions and\nadvertisements are studied in an agent-based model of consumers in a social\nnetwork. The influence of network topology was varied by using a small-world\nnetwork, a random network and a Barabasi-Albert network. The degree to which\nconsumers value the opinion of their peers was also varied. On a small-world\nand random network we find a phase-transition between an open market and a\nlocked-in market that is similar to condensation in liquids. At the critical\npoint, fluctuations become large and buying behaviour is strongly correlated.\nHowever, on the small world network the noise distribution at the critical\npoint is Gaussian, and critical slowing down occurs which is not observed in\nsupermarket sales. On a scale-free network, the model shows a transition\nbetween a gas-like phase and a glassy state, but at the transition point the\nnoise amplitude is much larger than what is seen in supermarket sales. To\nexplore the role of advertisements, a model is studied where imprints are\nplaced on the minds of consumers that ripen when a decision for a product is\nmade. The correct distribution of weekly sales returns follows naturally from\nthis model, as well as the noise amplitude, the correlation time and\ncross-correlation of sales fluctuations. For particular parameter values,\nsimulated sales correlation shows power law decay in time. The model predicts\nthat social interaction helps to prevent aversion, and that products are viewed\nmore positively when their consumption rate is higher.",
"arxiv_id": "physics/0508065",
"authors": [
"Robert D. Groot"
],
"categories": [
"physics.soc-ph",
"q-fin.GN"
],
"doi": "10.1016/j.physa.2005.08.023",
"title": "Consumers don\u0027t play dice, influence of social networks and advertisements",
"url": "https://arxiv.org/abs/physics/0508065"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "3ad7266e-daed-42e1-ad43-0f0ee016c22c",
"id": "arXiv Dataset IDs",
"type": "Model",
"variant": "snapshot-2026-03-01",
"version": "0.1.0"
},
"user_id": 1000002
}