How a Taiwan Ban Became Free Advertising for Amap’s Map App

Incident Overview

In recent days Taiwan’s authorities warned users not to download the mainland navigation app Amap, citing "security concerns." The warning itself generated massive curiosity, propelling Amap to the top of iOS and Android charts in Taiwan as users rushed to try the app.

Traffic‑Light Countdown: How Amap "Guesses" the Signal

Amap’s ability to display remaining seconds for traffic lights is not based on an official traffic‑signal API. According to patent CN114463969A (2022), the method reverses engineers the signal cycle from massive "stop‑start" user trajectories.

Step 1: Extract sample trajectories – Identify vehicle traces that pause near a stop line with near‑zero speed.

Step 2: Set a reference start point – Within a target time window (e.g., weekday morning peak), pick the first vehicle that accelerates and treat its start time as zero.

Step 3: Compute offset distribution – For all other samples in the window, calculate the time difference between their start moments and the reference.

Step 4: Find periodic peaks – Aggregate offsets into a histogram; evenly spaced peaks correspond to the traffic‑light cycle, and the interval between peaks gives the signal period.

Because traffic lights are periodic, vehicle starts cluster a few seconds after the green phase begins, producing evenly spaced peaks. By averaging data across multiple days, Amap reduces noise and keeps countdown error within about 1 second .

Data Density Determines Accuracy

The approach relies on sufficient sample density: busy intersections with many users yield accurate estimates, while sparsely trafficked or newly opened roads lack enough data, causing the feature to fail.

Network Effects of Data Scale

Amap processes over 30 billion navigation requests daily, achieving 92 % prediction accuracy for 30‑minute traffic conditions. This massive volume creates a "data wall" that competitors cannot easily breach.

Google Maps, despite a large global user base, has a highly dispersed user distribution; in Taiwan its active users are only a few million, far fewer than Amap’s mainland daily active users, leading to a lower data density and less precise features.

Amap also integrates real‑time data from 30 major cities’ traffic‑management systems, reinforcing its network effect: more users generate more data, which improves the model, which in turn attracts more users.

Reverse‑Propagation Model of the Ban

The ban acted as a media shock that multiplied curiosity‑driven downloads. Modeling it as a simple contagion:

Let the pre‑ban daily new‑user count be U₀ (low).

The ban creates exposure factor k , instantly raising the download rate to k·U₀ .

Each new user influences R₀ potential users on social media; if R₀ > 1 and word‑of‑mouth is positive, exponential growth follows until market saturation.

The combined effect of the ban’s initial shock and Amap’s strong product‑level advantage makes the surge almost inevitable.

Conclusions

Data‑scale advantage : Features such as traffic‑light countdown, lane‑level navigation, and precise traffic prediction stem from massive user‑behavior data; the accuracy gap scales with data volume and cannot be closed quickly by engineering alone.

Google’s relative lag : Effective crowdsourced navigation requires a critical mass of real‑time user data, which Google lacks in Taiwan, leading to a “good‑enough” stance without competitive pressure.

Local alternatives : Taiwan’s paid “navigation king” relies on traffic‑camera data and serves a niche need (speed‑ticket avoidance) but cannot match the free, data‑rich Amap in overall quality.

Ban as acceleration : Administrative attempts to suppress software can backfire; the ban became news, the news became advertising, and the advertising drove downloads, illustrating a classic reverse‑propagation dynamic.

References: Amap patent CN114463969A; Medium technical analysis (ReturnZeroBeing, 2024); Henfridsson et al., 2020 (Warwick University data‑network‑effects study); trafficlair.com traffic‑light countdown overview; JustOneAPI Amap technical report (2025).