The arithmetic of a field force

This post walks the published research on what one US field rep costs and can do, with each figure's year and source attached, and sets it against what a zenrep is structurally — properties of software, not measured performance. Zenreps is pre-launch and publishes no performance numbers it hasn't earned. And the frame matters: a zenrep is a coworker, not a substitute. This arithmetic is about the work software should carry so your reps can carry the relationships.

What one rep costs

Industry research puts the average yearly cost per rep at ~$260,000, up to $450,000 depending on geography, therapeutic area, and physician type ^[1]. A fleet car alone runs about $14,000 a year ^[7]. And rep pay alone works out to $20–57 per call — before the vehicle, samples, and management overhead ^[4].

What one rep can reach

As of 2016, 44% of US physicians were rep-accessible, down from roughly 80% in 2008; 18% were severely restricted ^[2]. Access is bounded by office hours and windshield time in one territory, one conversation at a time — the benchmark inside rep makes 46 dials for 5.8 quality conversations a day ^[3].

The hiring math

Ramp runs 3.3 months on average for inside reps ^[3], up to three months for pharma hires to get up to speed ^[1], and 5.42 months just to fill a vacant field role ^[5]. Turnover runs roughly 26% a year across sales forces, at $97,690 average cost to replace ^[5] — Gallup puts replacement at 0.5–2x annual salary ^[6].

The other column

Against each of those lines, a zenrep is a structural fact rather than a measured result. Software carries no salary, bonus, benefits, fleet, or T&E line items. It is available around the clock, including the after-clinic hours when HCPs actually have time. Capacity is not bounded by headcount — it scales with demand. Ramp is zero once the approved corpus is loaded; the "new rep" is a copy of the trained one. It doesn't resign, doesn't take the territory to a competitor, and carries no backfill cost.

And the record-keeping inverts: where call notes are self-reported after the fact, every zenrep conversation is logged verbatim, automatically, the moment it happens — into an append-only audit trail. Replies are assembled from your approved corpus and, on web and Telegram, verified claim by claim before send. Those are design properties of software, not performance claims.

Honesty notes

Figures are shown with their survey years and are not inflation-adjusted; several predate 2020 and are almost certainly higher in today's dollars. DePaul, Gallup, and Bridge Group figures are general sales-industry benchmarks, not pharma-specific. The ZS access data shown is from 2016, the most recent year we verified against a primary public source; access is widely reported to have tightened since. The MedCepts per-call figure is rep pay only. Everything claimed for a zenrep above is structural — a design property of software — because Zenreps is pre-launch and publishes no performance metrics of its own yet.