Ad Microtargeting: Precision Propaganda vs Broadcast Reach

Campaign Controls

Targeting precision 70%

Campaign budget 60%

Message intensity 1.0×

Ad-library disclosure (counter-messaging)

Live Readouts

Avg. belief shift

0.000

Targeting ROI

0.000

Reach (segment %)

Polarization σ

0.000

How it works

Each of the 1,600 grid points is a simulated person placed on two psychographic axes: political lean (x) and media trust (y). Every person has a fixed susceptibility s(x,y) — how easily their belief moves — modelled as a Gaussian bump centred on a "persuadable" region of that space:

s(x,y) = exp( -((x-x0)² + (y-y0)²) / (2σs²) )

A campaign pulse delivers exposure e(x,y) to each person. Broadcast (targeting = 0%) spreads the budget B uniformly: e = B / N. Targeted (targeting = 100%) concentrates it with a narrow Gaussian kernel centred on the most-susceptible cluster, so the same budget buys far higher exposure to the people most likely to move:

e_targeted(x,y) = B · exp( -((x-x0)² + (y-y0)²) / (2σt²) ) / Z
targeting mixes broadcast and targeted: e = (1-T)·e_broadcast + T·e_targeted

Belief updates with a bounded-confidence rule (each person moves a fraction of the way toward the campaign's target belief, scaled by their own susceptibility and the exposure they received), then decays slowly back toward neutral between pulses:

belief += η · e(x,y) · s(x,y) · intensity · (target − belief)
belief *= (1 − decay)   between pulses

Ad-library disclosure models platform transparency rules that force microtargeted ads into a public archive, enabling counter-messaging: it multiplies targeted exposure by (1 − disclosure_penalty), sharply reducing the ROI advantage that precision targeting normally has over broadcast — the same dynamic behind real ad-transparency and "why am I seeing this ad" mandates used against coordinated influence campaigns.

Targeting ROI = average belief shift ÷ budget spent, so it directly compares how efficiently a targeted vs. broadcast strategy converts spend into persuasion. Polarization σ is the population's belief standard deviation — a targeted campaign raises it (a persuaded cluster vs. an untouched majority) far more than a broadcast one.

x → political lean · y → media trust · height/color → belief

Unpersuaded (belief ≈ 0) Persuaded (belief → target)

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.