A geologist evaluating a new area rarely starts with a single well. She starts with a question like: which wells here are actually performing, and what do they have in common? That's a screening problem — filtering a county's entire well population down to the handful worth studying in detail. Done by hand in a spreadsheet, it's a day of exports, joins, and cleanup. Asked conversationally against the Wellsite data lake, it's one query.
Let's walk through a real screen the way a user would actually run it.
The question behind the question
Say you're looking at a Delaware Basin county with a few thousand horizontal wellbores on record. You don't want the all-time cumulative leaders — those are old, long-lateral wells that flatter the count with years of tail production. You want the best recent wells, because they tell you what current completion designs and current operators are delivering right now.
So the screen isn't "top wells by cumulative oil." It's something more specific:
Show me horizontal oil wells in this county with a first production date after January 2022, ranked by cumulative oil through their first 12 months, and only wells that have a full 12 months of history.
That's a screen with four filters stacked on top of a ranking: county, well type, first-production window, and a normalization rule (first 12 months) so a well that came online last month doesn't get unfairly buried under one that's been producing for a year.
Why normalization matters
Ranking raw cumulative production across wells with different ages is the most common way a screen goes wrong. A well that made 180 MBO over three years looks better than one that made 165 MBO in twelve months — until you realize the second well is on a much steeper, higher-value trajectory.
Normalizing to a fixed window — first 90 days, first 6 months, first 12 months — puts every well on the same footing. First-12-month cumulative oil is a clean proxy for well quality because it captures the flush period and the early decline without letting lateral length and vintage muddy the comparison too badly. If you want to go a step further, you can normalize by lateral length to get a per-foot figure, which strips out the "they just drilled a longer well" effect entirely.
The point is that the screen carries its assumptions with it. When you ask for a ranking, you also say ranked by what, over what window, and the shortlist reflects it.
Reading the shortlist
A good screen returns more than a list of API numbers. For each well you want the operator, the lease, first production date, the ranking metric, and enough context to judge whether the number is real or an artifact.
Suppose the top of the list comes back looking like this:
- Three of the top ten wells belong to a single operator, all on the same lease, all brought online within a four-month stretch.
- The metric spread is tight at the top — the #1 and #5 wells are within maybe 8% of each other on first-12-month oil — then falls off a cliff by #10.
- Two wells that should have ranked high are missing because they had a multi-week shut-in in month three, so their 12-month cumulative undersells the true rate.
That last point is why outlier and gap detection matter alongside the ranking. A well that got choked back or shut in for facility work isn't a bad well — it's a well with a hole in its record. Flagging those keeps you from throwing out a strong candidate or, worse, chasing a completion design that only looked mediocre because of downtime.
From shortlist to insight
The screen is the setup. The payoff is what you ask next, and this is where staying inside one conversation pays off:
- Who drilled them? If one operator owns half the top ten, that's your benchmark. Pull their book in the county and see whether the whole program performs or just a lucky lease.
- Where are they? Map the shortlist and the pattern usually isn't random — the best wells cluster in a fairway. That's a landing-zone or thickness story worth chasing.
- What's the decline telling you? A well can top the 12-month screen and still be fading fast. Layer in decline-rate analysis and you separate the wells with durable, shallow declines from the ones that front-loaded everything and are already rolling over.
- How do they compare to the county average? Benchmark each shortlist well against the county mean for its vintage. Being top-ten in absolute terms is one thing; being 2x the county average is the signal that something specific and repeatable is going on.
Turning a screen into a standing view
The first run answers today's question. The more useful move is to keep it live. New wells cross your first-production window every month, so a screen that was current in the spring is stale by fall. Setting an alert on new permits and new completions in the county means the shortlist refreshes itself — you find out when a well enters the ranking, not six months after the fact when someone else has already leased around it.
The takeaway
Screening isn't glamorous, but it's where most good field work starts. The value isn't the ranked list itself — it's how fast you get from a few thousand wellbores to the ten that deserve your attention, with the filters, normalization, and data-quality flags handled in the same breath. Ask the question the way you'd ask a colleague, and let the record do the sorting.