The 4-Week Pump-by-Priority Pilot: What Actually Happens, Week by Week

A pump-by-priority pilot is not a software install. It is a structured experiment with a CFO-signed metric on Day Zero and a measurable result on Day 28. Most operators expect a multi-quarter build. The shape of the actual pilot is one week to integrate, two weeks to redesign the morning, and one week to measure. Here is what each week looks like, what gets touched, and what the operator owns at the end.

Why the Question Is Wrong Before the Pilot Starts

The first conversation usually opens with the wrong question. "How long is the implementation?" The honest answer is that a pump-by-priority deployment has two clocks. The integration clock is short. The behavior-change clock is the one that decides the outcome.

A small-to-mid operator can stand up the data plumbing in a week. SCADA, production accounting, GIS, EAM, and the historian are already producing the inputs the score function needs. Read-only connectors are a standard ETL exercise on a stack the operator already owns. The longer clock is whether the morning meeting changes shape, whether the pumper looks at a ranked plan in the truck cab instead of a fixed route, and whether the superintendent stops reconciling spreadsheets and starts adjudicating the top of the day's score. That clock is the one that decides whether the metric moves.

The four-week pilot is built around the second clock. It assumes the data is already there, the integration is plumbing, and the operating value comes from changing what the crew does on Tuesday morning. The pilot is the smallest legitimate experiment that proves the change is durable and the math is real.

Week 0: Sign the Metric, Not the Software

A pilot that starts without a signed metric is a demo, not an experiment. The first day on the operator side is the most important day of the engagement. The operations leader, the asset manager, and the CFO or controller pick exactly one metric and write it down. The score function the WorkSync Work Engine is going to optimize against is not chosen by the vendor. It is chosen by the operator and tied to a specific line on the operating budget.

Common Week-Zero metrics:

• Deferred production recovered per crew shift, in barrels of oil equivalent
• Mean time from anomaly to first field response
• Value-added time per lease operator shift, measured by SCADA-stamped intervention windows
• LOE per BOE on a defined route, baseline against the trailing 90 days
• Run-time delta on artificial lift on a defined set of wells

The metric is named. The baseline is measured. The threshold for "moved" is written into a one-page document the controller signs. There is no kill fee. There is no license fee in the pilot. If the metric moves past the threshold by Day 28, the operator signs the annual. If it does not, the operator walks away.

This is the Impact Guarantee. It is the single most underweighted decision a small-to-mid operator can make before a software pilot. The metric anchors every conversation for the next four weeks. The score function inside the Work Engine optimizes for what the operator wrote down, not what the vendor wishes were measured.

Week 1: Read-Only Integration on the Stack You Already Own

Integration is plumbing. Most operators imagine an enterprise software project. The reality on a pump-by-priority pilot is a small set of read-only connectors onto systems that are already producing the necessary signal.

What gets connected in Week 1:

• SCADA or historian for production rates, run statuses, alarms, and tag-level deviations. The Work Engine reads from AVEVA PI, Ignition, Cygnet, eLynx, and similar without changing the historian configuration.
• Production accounting for daily allocated production, well-level economics, and netback by stream. Enertia, Quorum, Peloton, and similar are typical sources.
• GIS for well locations, route polygons, and pad relationships. ESRI or an existing internal GIS export is sufficient.
• EAM or CMMS for equipment history, work order status, and maintenance backlog. Maximo, SAP PM, or a mid-market alternative all work.
• Field data capture currently used by pumpers. The pilot does not replace the form the pumper uses in Week 1. It reads from it.

There is no rip-and-replace. There is no new sensor. There is no field-side hardware install. Operators on 15-year-old SCADA installations have completed integration in five working days. The bigger lever is whether the data is reaching the score function, not how new the historian is.

The output of Week 1 is a working ingestion pipeline, a baseline of the chosen metric measured against the prior 90 days, and a one-page integration map that the operator's IT lead signs. The score function runs offline against historical data for the rest of the week so the operations team can see what a ranked plan would have looked like over the trailing quarter.

Week 2: The Score Runs Live, the Plan Hits the Truck Cab

Week 2 is when the operating model changes. The Work Engine begins running the scoring loop nightly against live data. Every well, every artificial lift unit, every open work order is scored on cash-flow impact, risk, and crew availability. The output is a ranked daily plan published to the truck cab and the operations center by 6 AM.

This is the visible change. The pumper opens the route on a tablet or phone. The route is no longer a fixed sequence of wells in calendar order. The route is the top of the day's score, ordered for the crew, the basin, and the truck. A 60-bopd well drifting 8% off forecast outranks the routine check on a separator that pinged a nuisance alarm at 3 AM. The pumper does not need to know why. The crew runs the top of the list.

Two things happen alongside the ranked plan in Week 2. First, the operations center starts adjudicating the score function. The superintendent reviews the top 20 items each morning, flags any item that the score over-ranked or under-ranked, and feeds the correction back into the Work Engine. The score function gets better in a week. Second, voice-first field capture becomes available where the operator wants it. The pumper can speak the observation while standing at the well. The observation flows back into the score before the next morning's plan. The radio chatter, the handwritten ticket, the gut call about a well that felt wrong, all become scoreable inputs.

Week 2 is also when the cultural friction usually shows up. Pumpers who built routes around personal knowledge of the field will sometimes resist a ranked plan that puts a well they "know is fine" at the top. The pattern that resolves this is letting the score show the math. When the 60-bopd well that drifted 8% off forecast actually had a backside leak, and the visit recovered three days of deferred production, the pumper becomes the most vocal advocate for the score. This is consistent across deployments.

Week 3: The Loop Closes

By the start of Week 3 the ranked plan is the default morning artifact. The integration is stable. The score function has absorbed one full week of operator corrections and pumper observations. Week 3 is the closed-loop week.

Three things happen in parallel:

• The scoring function retrains nightly against the field outcomes of the prior 24 hours. Items the crew visited and confirmed get weighted. Items the crew skipped that later became failures get weighted up next time.
• The dispatch loop tightens. Crew assignments are routed by the Work Engine on cash-flow impact and proximity, not on personal preference or last week's route. Pumpers cover more high-value wells per shift. The miles-per-visit ratio drops.
• The morning meeting changes shape. The superintendent stops rebuilding the day in a spreadsheet. The asset manager reviews the ranked plan, adjudicates the top items, and signs off. Total meeting time drops by 30 to 50% on most deployments, with the time reallocated to field response and crew coordination.

By Day 21 the operator is running a different operation. The metric chosen in Week 0 is measurable against the baseline. Most deployments see the metric move inside Week 3, before any optimization passes are run on the score function. The reason is simple. Eighty percent of the lift comes from doing the right wells first, not from a clever model. The remaining 20% is where the model earns its keep over the rest of the year.

Week 4: Measure, Decide, Sign

The fourth week is the decision week. The chosen metric is measured against the Week-0 baseline. The controller reviews the math. The asset manager reviews the operating change. The decision is one of two outcomes.

If the metric moved past the threshold:

• The operator signs the annual on the same Impact Guarantee terms
• The pilot route expands to the full asset on a defined schedule, typically 60 to 90 days
• The score function adds the operator's preferred secondary metrics into the ranking
• The closed-loop retraining runs continuously

If the metric did not move:

• The operator walks away. No license fee, no kill fee, no awkward sales call
• The integration is decommissioned. The operator keeps the baseline data and the integration documentation
• The vendor publishes a one-page post-mortem with the operator's permission

This is the structural difference between a pump-by-priority pilot and a typical software pilot. The decision criterion is named on Day Zero. The exit is documented and clean. The operator is buying an operating change, not a license. The pilot is the smallest legitimate experiment that proves the change works on their assets, their basin, and their crew.

What the Operator Owns at Day 28

By the end of the four weeks the operator owns:

• A live integration onto the existing stack with all read-only connectors documented
• A baseline of the chosen metric and a measured delta over 28 days
• A ranked daily plan published to every truck cab by 6 AM
• Voice-first field capture flowing back into the score
• A scoring function tuned to the operator's basin, well mix, and economic assumptions
• An adjudication workflow in the operations center
• A one-page operating change document the controller signed

Nothing in the field changed at the hardware level. No new sensors. No new SCADA. No new headcount. The bottleneck was never collection. It was scoring. The pilot moved the scoring layer onto the data the operator already had.

This is the shape of the deployed reference at a top 25 private producer running over 5,000 wells across the Western Anadarko, Permian, and Wyoming. The metric the controller signed in Week 0 was free cash flow uplift on the same well count. The metric moved. The annual signed.

What the Pilot Is Not

A pump-by-priority pilot is not a SCADA refresh. It is not a sensor install. It is not a multi-quarter integration. It is not a license bet. It is not a vendor demo dressed up as a deployment. The Impact Guarantee structure makes each of those non-pilots commercially unviable.

It is also not a model deployment in the sense the average AI vendor pitch implies. The lift in the first 30 days comes from doing the right wells first, not from a model. The Work Engine is the scoring and ranking infrastructure that makes "doing the right wells first" the default morning artifact. The model layer compounds the lift over the rest of the year. The pilot only needs to prove the first 30 days.

Where to Start

The shortest path to a four-week pump-by-priority pilot is one conversation. Send your well count, your basin, and your current dispatch or scheduling system. WorkSync sends back a per-asset adoption math estimate, a draft Week-Zero metric, and a one-page pilot scope inside 30 minutes.

For operators who want the numbers, the metric draft, and the pilot scope on paper before requesting the pilot itself, the Adoption Math intake returns the per-asset estimate without committing to the deployment. The same score function applies to specific operating questions, including cutting tank-gauging visits in half on a 30-day pilot using the SCADA and meter data the operator already owns.

Request Your Free Trial. 4-week adoption pilot on the stack you already own. The metric is signed on Day Zero. The exit is clean if the metric does not move. The operator owns the integration either way.

FAQ

How is a 4-week pilot different from a typical software proof-of-concept? A proof-of-concept usually has no signed exit metric, no operating change, and no documented decision on Day 28. The pump-by-priority pilot inverts each of those. The metric is signed by the controller on Day Zero. The operating change is the deliverable, not a demo. The decision criterion is named and the exit is clean. The pilot is structurally an experiment, not a sales motion.

What if our SCADA stack is older than the average vendor pilot expects? It almost never matters. The Work Engine reads from AVEVA PI, Ignition, Cygnet, eLynx, and 15-year-old historian installs without changing the historian configuration. The bottleneck is whether the data is reaching the score function, not whether the historian is current.

Does the pilot require hiring new headcount on our side? No. The pilot uses the existing operations center, the existing pumpers, and the existing superintendents. The operating change is what those people do in the morning. The integration is run by the WorkSync deployment team in coordination with the operator's IT lead.

Can a pilot run on a single route rather than the full asset? Yes, and that is the default for most operators. A pilot route is typically 80 to 150 wells across one or two pumpers. The score function trains on the full asset but the operating change is contained to the pilot route until Day 28. Expansion to the full asset is a 60-to-90-day rollout after the annual is signed.

Why does the pilot have no license fee? Because the experiment only works if the operator and the vendor share the risk. A license fee in the pilot biases the operator toward continuing even if the metric did not move. The Impact Guarantee removes that bias. The pilot is the cleanest test of the approach on the operator's assets, basin, and crew. If the metric does not move, no one pretends otherwise.

What metric should we sign on Day Zero if we have not run a pilot before? The most common starting metric is deferred production recovered per crew shift, measured in BOE, baselined against the trailing 90 days. It is hard to fake, easy to instrument off existing SCADA and production accounting, and directly tied to free cash flow. Controllers consistently sign it.