Oil wells: reconcile drilling, completions and first production
A fall in drilled-but-uncompleted wells can reflect completions, new drilling and classification changes at the same time. It does not directly count wells that began producing that month. Research should preserve the well-stage ledger and obtain first-production dates before converting a DUC change into an output claim.
The deliverable is a well-ID queue from drilling completion through completion activity, connection and first production. All counts and volumes below are fictional. The worksheet concerns operating evidence; it does not recommend drilling, stimulation or well-operation procedures.
Locate the historical estimation method
EIA’s June 28, 2021 article on DUC estimates explains the difference between drilled and completed well records and the assumptions used for its regional estimates. Locate the paragraph beginning We estimate DUCs and the discussion of timing between drilling, completion and production.
This is a historical description of the then-Drilling Productivity Report for selected basins. It is not a current 2026 nationwide well-by-well inventory, a company asset register or a guarantee of how quickly a particular well will produce. Do not reuse the article’s old count as a current market fact.
For an issuer, obtain stable well identifiers, the meaning of each reported stage, the covered geography and the data vintage. Drilled, completed, connected and first-produced can have different reporting conventions. A field name that sounds similar across providers does not establish identical coverage.
Rebuild the DUC main case
Assume opening DUC stock of one hundred wells. Forty newly drilled wells enter the defined DUC population, fifty-five leave it through completion and five leave through an explicitly documented reclassification. Closing DUCs are 100 + 40 − 55 − 5 = eighty.
| Movement | Wells | Meaning |
|---|---|---|
| Opening DUC inventory | 100 | Defined initial well population |
| Newly drilled entrants | +40 | Enter the uncompleted state |
| Completed exits | −55 | Leave that state through the defined completion |
| Reclassification | −5 | Evidence-based change, not assumed output |
| Closing DUC inventory | 80 | Net decline of twenty |
The twenty-well net decline is neither the fifty-five completion count nor the number of first-producing wells. New drilling replenished inventory, and reclassification changed it separately. Calling twenty wells newly online would ignore both mechanisms.
The five reclassified records require an identified reason and source in a real ledger. They could not simply be a residual used to force the equation to close. An observed revision is also not automatically abandonment or a completed well; retain its actual classification evidence.
Follow the completed cohort to production
Assume the fifty-five completed wells have distinct IDs. Thirty-five have documented connections and dated first production in the current month. The other twenty have no production by the cutoff and a plan for connection in the next month. The plan remains forward-looking until actual records establish its outcome.
Connection alone would not have proved the thirty-five produced. This teaching case explicitly supplies both records. If only connection evidence were available, first-producing count would remain unverified. Keep stage dates and evidence fields separate rather than compress them into one online flag.
The thirty-five belong to this completion cohort. Other wells completed earlier could also start producing this month, so this cohort does not necessarily equal the company’s whole first-production total. Use the well IDs to link the relevant stages and identify older-cohort additions separately.
Do not count a well once when completion occurs and again as another new asset when it connects. It is the same physical well moving through states. If an identifier changes, preserve an evidence-based alias rather than duplicate the underlying asset.
Rebuild producing days before barrels
For the current month’s thirty-five wells, assume a thirty-day period and an invented constant oil rate of five hundred barrels per producing day per well. Twenty have verified first production on Day 1, and fifteen on Day 21, with no interruption after starting. These rates are teaching inputs, not estimates of actual well performance.
The first group contributes 20 × 30 × 500 = 300,000 barrels. The later group contributes 15 × 10 × 500 = 75,000. Total cohort production is 375,000 barrels under the stated assumptions. Counting all thirty-five for thirty days would yield 525,000, overstating the case by 150,000.
This arithmetic does not forecast future decline or establish the next month’s rate. Real production needs observed volumes, producing days, shut-ins and the applicable oil/gas unit basis. A well count cannot supply those quantities. Longer wells or different completions also cannot be treated as equal output units merely because each counts as one well.
Test a different boundary: data revision
Suppose a later release restates opening DUCs from one hundred to ninety-five while the listed physical stage movements remain unchanged. Closing inventory would then be seventy-five under that restated series. The extra five-well reduction from the earlier closing eighty is a baseline revision, not five additional first-producing wells.
Preserve original and restated series with their publication dates. Compare periods on a consistent vintage where possible, while retaining the forecast history. Do not select an old opening stock and a revised closing stock and call the entire gap current operational progress.
Geography also matters. A basin estimate and a company’s selected assets are different populations. Dividing company completions by a regional DUC total does not establish the company’s completion conversion unless the corresponding inventory population is known.
Distinguish stage movement from commercial output
First production is not necessarily first sale or cash collection. Storage, processing, transport and contractual measurement can separate those dates. If the research concerns realised sales, obtain the actual relevant volume and recognition records rather than multiply first-flow counts by a price.
Likewise, an operationally connected well can be temporarily shut in or await another constraint. This worksheet does not infer a permanent productivity problem from a missing current record. It identifies the unsupported stage and the evidence needed to establish it.
Use the well-stage verification sheet
| Field | Obtain | Check |
|---|---|---|
| Identity | Well register and geographic scope | Stable asset IDs and aliases |
| Drilling/DUC | Defined entry date and data vintage | Regional estimate distinct from company records |
| Completion | Dated completion evidence | Same well, actual event rather than plan |
| Connection | Relevant infrastructure record | Not substituted for first production |
| Production | First date, producing days and measured volumes | No full-month assumption for late starts |
| Revisions | Reclassification and restatement bridge | Statistical change not relabelled physical output |
When sources disagree, compare stage definition, reporting lag, geography, ID matching and preliminary versus revised data. Keep the unresolved record visible. Do not force it into completion, abandonment or production merely to close a queue.
If first-production records are missing, report the supported drilling and completion counts and leave output unknown. A planned connection date is not a substitute. Update after actual connections, production records, shut-in information, reclassifications or revised data. The conclusion should state which wells changed stage, when they actually produced and which volume assumptions remain unverified.
Keep cohort and revision evidence together
A practical register needs one row per stable well identifier and one dated observation for each stage. Save the source that establishes the stage, not only the latest label. If a well is renamed, retain an evidenced alias instead of creating another physical well. If a row disappears between data releases, do not automatically call it completed; first investigate a coverage or classification change.
In the fictional production calculation, the first group contributes six hundred producing well-days and the later group 150, for 750 in total. Multiplying that matched exposure by five hundred barrels per producing day gives the stated 375,000 barrels. This independent unit check is useful because the count of thirty-five wells alone does not expose the delayed start.
For a portfolio summary, keep the period bridge and the cohort follow-through side by side. The first explains why DUC inventory changed; the second explains what happened to the fifty-five completed wells. They answer different questions and should not be forced into the same denominator. A later production record for one of the twenty waiting wells updates that cohort’s stage and producing days, rather than retrospectively treating all twenty as producing throughout the earlier month.
FAQ
Does a twenty-well DUC decline mean twenty new producing wells?
No. The main case includes fifty-five completions, forty new entrants and five reclassifications.
Can every first-producing well be assigned a full month of output?
No. Its actual start date and producing days determine the applicable period.