Drilling delivers one well at a time. The production engineer inherits all of them at once — and has to keep a few hundred of them alive, profitably, mostly through a screen.

If the reservoir engineer thinks in years and the drilling engineer thinks in hours, the production engineer thinks in barrels this morning versus barrels yesterday. Their domain is everything between the sandface and the stock tank: the well, the lift, the flowline, the separator. Their enemy is deferment — every barrel the asset could have made and didn't.

The morning call

The day starts on the operations call. Overnight, SCADA has been streaming surface data off every well — tubing and casing pressures, rates, run status — into a system like CygNet, XSPOC, Weatherford ForeSite, or a Peloton/PI historian. The production engineer comes in to a dashboard of exceptions: wells that went down, wells making less than they should, ratios drifting the wrong way.

The first job of the day is triage, and it's relentless because the population is large. A pumper has physically driven past some of these wells; most of them only exist to the engineer as a row on a screen and a trend line. Out of a few hundred wells, which five are costing the most barrels right now, and which of those is worth a truck roll, an intervention, or a phone call to a service company?

Artificial lift: the heart of the job

Almost no well flows to surface on its own forever. Sooner or later it needs artificial lift, and choosing and babysitting that lift is the production engineer's signature work. The method depends on the well:

  • Rod pump (the nodding "pumpjack") — the workhorse of stripper and onshore oil wells. The engineer reads dynamometer cards — pump-load-versus-position plots — the way a doctor reads an EKG. A fluid-pound, a gas-locked pump, a worn barrel, a parted rod: each has a signature on the card.
  • ESP (electric submersible pump) — high-rate wells. Here the daily worry is run life. An ESP failure means a workover rig and a long, expensive downtime, so the engineer watches motor temperature, intake pressure, and current trends looking for the slow drift that precedes a failure, and runs root-cause analysis on every pull.
  • Gas lift — inject gas to lighten the fluid column. The daily optimization is allocation: you have a finite amount of lift gas and a field full of wells competing for it, and putting the next thousand cubic feet where it makes the most oil is a genuine optimization problem people solve, in part, by feel.
  • Plunger lift, PCP, and others for the cases that fit them.

Nodal analysis and the well model

When a well underperforms, the production engineer reaches for nodal analysis — modeling the well as two curves that have to meet: inflow (what the reservoir will give up, the IPR) and outflow (what the tubing and lift will carry to surface, the VLP). Where they cross is the well's operating point. Move the crossing — bigger tubing, more lift gas, a different choke, a stimulation — and you move the rate. This is done in PROSPER / GAP / MBAL (Petex's IPM suite), PIPESIM (SLB), WellFlo, or equivalents, tuned against the latest flowing survey and well test so the model actually matches reality before you trust it to tell you what to change.

Well tests, failures, and interventions

Threaded through all of it: well tests to validate what each well is really making (and to keep the allocation honest), failure analysis when something breaks, and interventions to fix it — slickline for the light stuff (changing a gas-lift valve, a gradient survey, a cleanout), coiled tubing and workover rigs for the heavy stuff (stimulation, water shutoff, a pump change). And always flow assurance: the scale, paraffin, asphaltene, and hydrate problems that quietly choke a well to death if the chemical program lapses.

Underneath everything sits LOE — lease operating expense. The production engineer is the person most directly responsible for the denominator of every well's economics, the one who knows that a chemical truck, a slickline run, and an electricity bill are the difference between a marginal well that's worth keeping and one that gets plugged.

The screen-fatigue problem

Here's the structural tension of the job: the production engineer is asked to watch more wells than any human can actually watch, through a stack of systems that don't quite talk to each other. SCADA in one tool. Allocation in another. Well models in a third. Economics in a fourth. The "single pane of glass" is, in practice, a row of browser tabs and an Excel workbook that stitches them together. Alarm fatigue is real — when everything pages you, nothing does — and the genuinely important signal (the ESP that's three days from failing) hides inside a flood of nuisance trips.

The best production engineers aren't the ones who stare hardest at the dashboard. They're the ones who've figured out how to make the dashboard tell them only the things worth acting on.

Where this connects to what we build

This is the discipline where Workollab's thesis is most obviously true: the bottleneck isn't a missing algorithm, it's integration and signal. A small operator already has the SCADA data, the well tests, the failure history. What they usually lack is a way to fuse it into a surveillance view that surfaces the few wells worth acting on today — without renting a seven-figure enterprise platform aimed at a super-major. Practical, local-first tooling that runs on the operator's own data and respects the engineer's judgment about what matters is exactly the underserved middle we build for.

The production engineer keeps the lights on, one barrel at a time. The right tooling just makes sure the barrel they're chasing is the one worth chasing.

Tools & terms named in this piece

SCADA / historians: CygNet, XSPOC, Weatherford ForeSite, Peloton, PI; nodal analysis (IPR/VLP) in PROSPER / GAP / MBAL (Petex IPM), PIPESIM (SLB), WellFlo; WellView (well/ops data); Spotfire / Power BI for surveillance; artificial lift (rod pump & dynamometer cards, ESP run-life, gas-lift allocation, plunger, PCP); well testing, flow assurance, LOE. Workflow details drawn from current production- engineering job specifications, vendor documentation, and practitioner accounts, 2025–2026.

One of Workollab's oil & gas engineering pieces.