This group brings together those who are interested in topics around oil and gas exploration, drilling, refining, and processing.


How accurate can lease-to-well production allocation be?

Enno Peters's picture
CEO, ShaleProfile

Background in AI, worked on developing Supply Chain Planning & Optimization solutions for Quintiq, setting up its business in China. Focus on company direction and the technical development...

  • Member since 2018
  • 255 items added with 546,264 views
  • Nov 1, 2022

A seemingly simple question: how much did a Texas well produce? A question fundamental to oil & gas analytics, yet difficult to answer due to the problem of lease-to-well allocation.

In contrast to most states in the US, operators in Texas report oil & gas production not on well level but on lease level. A lease is comprised of a group of wells in a similar geographic area, which can range from only a single well to over a thousand for some of the big ranches. For analytical problems like well forecasting, having production on well level is crucial to get acceptable accuracy. Armed with lease production, well-tests, and a few other inputs, the challenge is to allocate lease production back to individual well production for each hydrocarbon and water stream.

Again, sounds easy but in practice, the problem is very challenging due to the intricacies of the Rail Road Commission of Texas (RRC), the state agency where operators are required to file their reports. Changing leases, stacked laterals, pending productions, reclassified wells are few of the factors that need to be considered for an accurate allocation. No small task since in Texas, there are over four hundred thousand leases and six hundred thousand wells to allocate.

Novi approach

At Novi Labs we have a couple of key advantages at our disposal that help us with this puzzle. Our many years of working with operators and experience with forecasting have provided us an extensive knowledge base regarding well productivity. We leveraged this to design algorithms to create accurate production profiles on well level. This algorithm uses all the available inputs that are relevant for allocation. It can also handle all possible lease configurations, even those that change over time. The actual lease-to-well allocation then makes use of these production profiles to allocate production among all wells on the lease.

Without actual data, the construction of such an algorithm would be difficult. Which brings us to our next asset: the Novi Data Network (NDN). This is a set of about fifteen thousand wells (and growing) whose actual production data we can use to optimize our algorithms. It allows us to identify areas for improvement and test within minutes whether a given change has a positive impact on allocation accuracy. NDN has also been instrumental in flagging idiosyncrasies in the RRC filing system. Given the large variety of these, knowing which ones have the biggest impact on allocation allows us to narrow our focus and make big improvements in a short amount of time.

Knowing where the issues are is not sufficient. To solve them, we also made extensive use of the close contact we have with our operator customers. Since operators are the ones that actually file the reports to the RRC, they are in a unique position to inform us about the unique reporting requirements. Novi’s operator customers provided another important piece of solving the allocation puzzle. Combined with our fast iteration and instant feedback from NDN, this helped us reduce our allocation scores rapidly.

How good is the Novi allocation?

We measure a set of key metrics to evaluate our allocation accuracy, with the key performance indicator being Root Mean Squared Error (RMSE). Its units are the same as those of production itself, it is scale dependent and puts extra emphasis on outliers. We calculate this metric by taking all the production months in a test set of NDN wells and comparing them with our allocated production months. The graphs below show our most recent progress. It demonstrates how, in the course of only a few months, our allocation errors were reduced by 58% for oil, 46% for gas, and 79% for water. With these improvements our current lease-to-well allocation is now industry leading.

What’s next?

Because lease-to-well allocation is so critical for analytics purposes, Novi will continue to make further improvements. By growing NDN, talking to customers, and fine-tuning our algorithms, we will lower our error scores even further in the months ahead. Part of our mission at Novi is to have a world-class lease-allocation algorithm, and with the enormous progress made this year we believe we are already there and will work hard to stay there.

Want to compare Novi to your current solution?

With Novi Labs you can leverage the most accurate, time and comprehensive data available for lower 48 data. If you are an operator, mineral or royalty owner, analyst or investor, Novi provides an easy-to-use platform with powerful insights and real time updates.

You can request a free demo here:


No discussions yet. Start a discussion below.

Enno Peters's picture
Thank Enno for the Post!
Energy Central contributors share their experience and insights for the benefit of other Members (like you). Please show them your appreciation by leaving a comment, 'liking' this post, or following this Member.
More posts from this member

Get Published - Build a Following

The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.

If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.

                 Learn more about posting on Energy Central »