An offset becomes a useful benchmark only after the question and comparison population are defined. Proximity alone does not control geology, trajectory, drilling mode, BHA, bit, measurement basis, or operational limits. A defensible benchmark makes those differences visible and yields to current-well evidence as it accumulates.
Define the benchmark question
“Best offset” is not a complete question. Best at what? A well can lead one metric and perform poorly in another. For example:
- Fastest ROP: may identify the most aggressive drilling interval.
- Highest drilling efficiency: may identify the mechanically cleanest operating condition.
- Lowest tortuosity: may identify the best directional outcome.
- Lowest section time: includes both drilling and flat-time performance.
- Best bit run: may emphasize footage, ROP, dull condition, or avoidance of trips.
- Best connection performance: has little to do with bit or formation performance.
The benchmark population should therefore be constructed around the question. If the question is:
What WOB and RPM worked well in this formation?
then the relevant observations should ideally be:
- actual drilling,
- in the comparable formation,
- with comparable BHA/bit conditions,
- under similar hole geometry,
- and without major dysfunction.
If the question is:
Which rig drilled this section fastest?
the population and normalization requirements are different. The benchmark follows the engineering question, not the other way around.
Align the correct coordinate
Most offset comparison begins with depth. This is sensible because drilling decisions are heavily depth-dependent. SPE/IADC-184743 describes a real-time offset-comparison system that plotted current and historical well data together by depth, allowed adjustment for TVD shifts, and marked formation tops.[1] If Formation A begins at 9,800 ft MD in the offset and 10,050 ft MD in the current well, a raw measured-depth overlay shifts the formation response by 250 ft. The engineer may conclude:
The current well is drilling differently at 9,900 ft.
But the two traces are not actually in the same rock. The comparison failed before the drilling parameters were even considered.
Measured-depth alignment can misplace comparable drilling behavior when geological boundaries shift between wells.
Depth normalization is not one universal operation:
| Basis | Useful for | Limitation |
|---|---|---|
| MD | Trip distance, torque and drag, bit footage, and section length | Does not align shifted geology. |
| TVD | Vertically separated geological position | Does not guarantee equivalent stratigraphic position. |
| Formation-relative depth | ROP, MSE, lithology response, and parameter effectiveness | Depends on reliable formation picks. |
For example, one might compare:
$$Depth_{relative} = Depth - Formation\ Top$$
so that zero represents the same geological reference in each well. Formation-relative depth is not mandatory; the depth basis should match the engineering phenomenon being compared.
Build a comparable operating cohort
The rig-state article established a critical rule: Valid sensor data generated during the wrong operation is still inappropriate for the analysis. If Offset A contains:
- rotary drilling,
- reaming,
- connections,
- circulation.
Current Well B is filtered to:
- rotary drilling only.
If all Offset A data is averaged by depth, the comparison is contaminated. ROP may be lower, torque behavior may differ, and WOB may include off-bottom observations because the populations were defined differently. A fair comparison matches rotary with rotary or slide with slide, not merely the same depth. SPE-186166 filtered its six-well comparison to rotary-drilling observations for this reason.[2]
An offset at 240 ft/hr versus a current well at 180 ft/hr indicates underperformance only if the formations are comparable. A stronger or more abrasive interval may legitimately require:
- higher mechanical energy,
- lower ROP,
- different WOB/RPM combination.
SPE-186166 emphasizes formation strength as relevant context for interpreting drilling efficiency and also notes that formation changes can contribute to apparent dysfunction or MSE changes.[2] This suggests a better benchmark:
How does current performance compare with offsets while drilling comparable rock?
rather than:
Which ROP trace is farther to the right?
Historical drilling data becomes much more useful when the performance trace is attached to the configuration that generated it. For each interval, useful metadata might include:
- bit manufacturer/model,
- bit size,
- dull history,
- motor bend,
- motor power section,
- RSS versus conventional assembly,
- stabilizer placement,
- BHA run number,
- accumulated bit footage,
- on-bottom hours.
Without that context, a high-performing interval can become misleading. The offset may have entered with a fresh bit while the current bit already has 5,000 ft on it; the same ROP should not necessarily be expected. Likewise, a BHA optimized for directional responsiveness may operate differently from one optimized for lateral stability.
Compare distributions
Three offsets can have nearly identical average lateral rotary ROP and still represent different performance:
| Well | Average ROP | Distribution |
|---|---|---|
| A | 205 ft/hr | Most footage between 190 and 220 ft/hr |
| B | 198 ft/hr | Half near 130 ft/hr and half near 270 ft/hr |
| C | 200 ft/hr | Most near 200 ft/hr with a few extreme spikes |
Those wells do not represent the same drilling behavior. Useful statistics might include:
- median,
- P25,
- P75,
- P90,
- interquartile range,
- footage-weighted distributions.
This is especially useful when constructing a benchmark from several offsets. Instead of choosing one best well, build an envelope. At a particular formation-relative depth, that might be P25 = 175 ft/hr, median = 205 ft/hr, and P75 = 230 ft/hr.
Now the active well can be evaluated against a population rather than one historical path.
A multi-well performance envelope can provide a more robust benchmark than a single selected offset.
Let current-well evidence take over
SPE-186166 found that efficient operating regions could vary substantially even among wells on the same pad.[2] Historical parameter selections are therefore starting points; real-time data should guide adjustment.
Before entering the interval, use offsets to establish:
- plausible WOB range,
- plausible RPM range,
- expected ROP,
- known dysfunction areas,
- known mechanical limitations.
Once drilling begins, use the active well to determine:
- whether performance matches the expectation,
- whether dysfunction appears,
- whether formation response differs,
- whether parameter adjustment is justified.
The historical data establishes the prior. The current well supplies new evidence. Offsets are most valuable before the active interval has been drilled; after several stable stands, the current well may provide a much better description of:
- its own formation response,
- its bit condition,
- its BHA behavior,
- its hole-cleaning state.
At that point the analytical emphasis should gradually shift from what offsets predict to what this well is demonstrating. This is where real-time benchmarking becomes more useful than a static pre-job roadmap.
Depth-aligned offsets expose the comparison cohort and operating ranges, allowing current-well behavior to be judged against visible historical evidence.
Practical offset-benchmarking workflow
A defensible workflow is:
| Stage | Required choice or check |
|---|---|
| Define the question | Specify whether the benchmark concerns ROP, efficiency, WOB/RPM, connection time, BHA performance, or total section time. |
| Select offsets | Match formation, location, hole section, trajectory, bit/BHA, and operating environment as far as the data allows. |
| Normalize depth | Choose MD, TVD, formation-relative depth, or another basis suited to the phenomenon. |
| Filter state | Compare rotary with rotary, slide with slide, and trip with trip. |
| Attach context | Include BHA, bit, mud, motor/RSS, hole size, and bit age. |
| Compare distributions | Use medians, percentiles, and variation rather than averages and maxima alone. |
| Check quality and dysfunction | Interpret ROP alongside torque, MSE, dynamics, and wellbore quality where available. |
| Build an envelope | Prefer demonstrated ranges and multi-well distributions over one exact parameter prescription. |
| Update live | Let current-well evidence modify the historical expectation as the interval develops. |
Related technical resources
- The Average Trap: Comparing Drilling Performance With Distributions
- When to Trust a Statistical Drilling Benchmark and When Not To
- Formation-Aware Drilling Optimization: Interpreting Performance Across Changing Rock
References
-
Behounek, M., Thetford, T., Yang, L., Hofer, E., White, M., Ashok, P., Ambrus, A., and Ramos, D. Human Factors Engineering in the Design and Deployment of a Novel Data Aggregation and Distribution System for Drilling Operations. SPE/IADC-184743-MS, SPE/IADC Drilling Conference and Exhibition, The Hague, Netherlands, 2017.
-
Ambrus, A., Ashok, P., Chintapalli, A., Ramos, D., Behounek, M., Thetford, T. S., and Nelson, B. A Novel Probabilistic Rig Based Drilling Optimization Index to Improve Drilling Performance. SPE-186166-MS, SPE Offshore Europe Conference & Exhibition, Aberdeen, United Kingdom, 2017.