Surface lanes where the p25 has collapsed year-on-year. The earliest empirical signal of structural margin compression on any molecule-destination pair.
Vildagliptin India-into-Egypt p25 has collapsed 23% in a single year, from $90.51 to $69.67/kg. The lane held 41 shipments in 2024 versus 18 in 2023, a 128% volume increase. New entrant pressure compressed the bottom-quartile bid floor without yet pulling the median (still $70.50) down meaningfully. The radar fires before the median moves; by the time the median falls, the structural compression is already baked in.
Single molecule or full active-lane list. Default window is the last full year vs the prior year, configurable.
Year-on-year p25 change per lane. Lanes with drops above the user threshold (default -15%) are returned with the current median and the volume delta as causal context.
Citation-anchored brief with the affected lanes, the magnitude, the new-entrant count from lane competitiveness, and a recommended action: defend, exit, or restructure.
New entrants undercut at the bottom quartile first. The median holds for months. By tracking p25 directly, the radar fires before the volume-weighted price moves.
Set a threshold (default 15% YoY drop) and the radar emits a chip per affected lane. The brief surfaces the magnitude, the volume change, and the new-entrant count.
Each affected lane is tagged with a recommended posture: defend (raise minimum bid), exit (deprioritise capacity), or restructure (renegotiate the contract mix).
Trade aggregates refresh weekly. The radar reruns on every refresh; the brief carries the exact computation timestamp.
A real chat thread in PharmaDB. Type a question, the AI agent runs the tools, the answer lands as a saveable note.
pharmagraph_query Twelve lanes where p25 has fallen 15% or more, 2023 to 2024 · India-origin, USD/kg.
The margin pressure radar is the leading-indicator surface for lane-level price compression. PharmaDB self-joins molecule_country_flow year-on-year, computes the p25 delta per lane, and joins to lane_competitiveness for entrant context. Refresh cadence is weekly.
New entrants compete at the bottom quartile first. They undercut to win lane share; the median moves only after their volume catches up. Tracking p25 catches the erosion months before the volume-weighted price falls.
Default is a 15% YoY drop in p25 with at least 10 shipments in both years. Configurable per user. Higher thresholds (25% or 30%) surface only severe compression; lower thresholds (10%) surface early signal.
Shipment count is the proxy. Kg or pack-unit volume is captured where available. The radar emits volume delta alongside p25 delta so the user can distinguish volume-driven compression from oversupply compression.
Lanes with under 10 shipments in either year are excluded. The radar requires sustained, lane-level signal across multiple shipments; one outlier shipment does not move p25 in a 10+ shipment lane.
Defend: raise the minimum bid, lock contracts longer. Exit: deprioritise capacity allocation, divert to higher-paying lanes. Restructure: renegotiate customer mix, drop loss-making counterparties. The recommendation is computed against lane HHI and the volume delta.
Trade aggregates refresh weekly. The radar reruns on every refresh; the catalog row carries an as-of timestamp.
Bring the molecule, the lane, or the supplier you're sourcing this week. The AI agent runs it on PharmaDB in 30 minutes. You keep the brief.