Maximizing Well Economics Using Tamr

Every industry is facing stark new challenges because of COVID-19. However, the oil and gas industry, which prior to the virus was facing stiff headwinds, is now in a state of extreme disruption. Due to a combination of poorly performing projects, high debt from borrowing against overly aggressive production forecasts, and the current global price war stemming from oversupply, many once lauded oil companies are looking at going bankrupt – with many assets potentially coming to market in the near future. Last week, Whiting Petroleum, filed for bankruptcy. And, unfortunately, this feels like a scenario that will be common over the next year or two.

For companies looking to squeeze more value out of their current assets to stay afloat, have properties going to market, or are those that are looking to buy assets at record low prices, it is critical to understand all of the metrics around well economics.

In this article, I am going to show you the benefits of using Tamr to master your parts and services providers  to improve well economics. It is amazing how much of a value increase happens with even a small change over the life of the well. We will cover the basic process of creating a “type well”, list the assumptions an engineer would make when thinking about future wells for end of year reserves or acquisitions, show the price sensitivity runs out of PHDWin, then sum up some of the metrics. This article should help non-engineers see where technical people get a lot of the information for this type of work and how Tamr is essential for all E&P companies going forward in the digital transformation. A note about well economics in the article, the data was pulled mid-January 2020.

What is a “Type Well”?

When a company is estimating their total asset value, they primarily look at producing developed wells (PDP’s – currently flowing wells), producing non-developed wells (PDNP’s – wells being drilled/completed at the time), and producing undeveloped wells (PUD’s – future locations).  There are also probable wells (PRB’s) and possible wells (POSS’s). To determine the classification of a future well location, a company follows rules set out by the SEC. That is a subject for a whole blog series, but if you are having trouble sleeping, you can read the rules that determine the classification here.  There is also another set of guidelines people use in conjunction with SEC rules on this site.  For this article, we will be considering future well locations as PUD’s – the classification where a company is very certain of a well location being economically productive and within a distance of 1 mile, or 1 section, of an already producing well in that same target formation.

We will pick an area in Eddy county, southeast New Mexico that is a popular part of the Delaware Basin.  We will look at this area and create a “type well”, an expected average well for the area. To make that well as realistic as possible for comparison to other producing wells, companies choose analog producing wells whose metrics are as close as possible to what they plan on booking for their PUD locations.  So, we will pick wells that have, generally, the same characteristics.  The area you are working in may not have statistically enough wells to choose from, but sometimes you just have to go with what you have.  In this case, we are picking local wells that:

1. Are completed in the same formation (Bonespring)
2. Are of similar lateral lengths (~9,800 feet)
3. Are the same vintage – frac jobs typically change over years (2017-2019)
4. Were drilled by the same company (not necessary, but great if available)

The figure below outlines the area of interest:

Zooming in on that portion shows 9 wells in the area that fit the criteria.  Of course, the more wells, the better, but this should be good for the purposes of evaluation in this article.  There are mixed theories on how to best look at all of the production to get the best version of the average, but I like to “high-align” wells – have all the highest points on early production line up.  These historic production curves will generate the average oil (single green line graph) and natural gas (single red line graph) profiles. There is an average water graph, too, but I don’t show it here.  The reason you calculate water is because it is an expense stream over the life of the well that has to be taken into consideration. If a well produces a lot of water, even if you are making a decent amount of oil and gas, it can kill the economics, making a company plug and abandon it earlier than expected.

Next, take those average production graphs, or “type curves”, and put them into your favorite reserves calculation software.  The two main ones used in E&P are Aries (Haliburton’s version) and what I use here, PHDWin. They both have their pros and cons, but PHDWin is very affordable and does a solid job for small to medium operators and consultants.  At this point, you would put a decline curve on the production. This is another very involved topic, but just assume I know what I am doing.  You can read more on this topic at Petrowiki.  The software will calculate the future production, economics, and associated metrics based on your skill in estimating what the decline will look like based on the production data available and the assumptions you put into the model.

You Know What Happens When You Assume…

You get the best answer you can based on what you know/can find out?  Here are all of the economic assumptions for southeastern NM I have made and where you can find this information in the future for your own evaluations.

Assumptions

1. Investment cost: $1,042/lateral foot – total cost for 9,800’: $10.22 million

1. Pull this from investor presentations on company websites or from 10-K SEC reports.  It is best to find publicly traded operators that only work in that basin as their numbers will be more “pure”, hence those operators being called “Pure Players”.  Also, some operators are awesome enough that they provide numbers by basin.
2. I mostly used EOG, Concho, and Cimarex for the Drilling and Completion costs (D&C)

2. Expenses:

1. $6.00/BOE (barrel of oil equivalent) on produced volumes
2. $0.95/BOE for gathering, processing, and transportation
3. $0.75/BOE for water produced and gathered
4. $2,500/month/well fixed cost (G&A and similar costs)
5. $60K for plugging and abandoning the well at the end of its life.
6. I found all of these values in the same places as the investment cost

3. Taxes:

1. Used the tax rates for NM and Eddy County.  PHDWin automatically puts them in if you specify state and county.  State taxes for oil and gas respectively are 7.09% and 7.94%, with local taxes being 5.00%

4. Well interests:  Working Interest:  87.5% Net Revenue Interest:  70%

1. These are guesses on my part, but you can find/infer this information in the same places as the above assumptions.  I went with two higher numbers where the NRI was 80% of the WI.
2. In the Delaware basin WI’s of 72-87.5% are common.

5. Well life is 35 years – Depending on who is doing the evaluation, 35 and 50 years are common well life limits.  After 35 years, no matter if a well is economic or not, the evaluation stops. Going the full 50 years does not add a great deal of value, and 50 years is a little long for a horizontal.  Also keep in mind, before the 50 years, the well may just end economically (expenses greater than cashflow)

6. For those reservoir engineers out there, the type curve metrics are:

Finally, let’s talk about product pricing.  There are a lot of places to get commodity pricing, but I usually use CME for oil and gas.  Two things about natural gas pricing in this model:

1. Natural gas in this part of the world suffers from a huge negative price differential because there aren’t enough pipelines for the amount being produced.  So, it is relatively common for the realized price to go from, say, $2.25/Mcf to $1.25/Mcf – sometimes less than 0.  I use a $0.81/Mcf differential, so the gas price for that year minus this number.
2. Finding area price differentials for natural gas can be difficult, but two good sites I use together are Natural Gas Intel and the EIA website.  You could also use CME again.

The pricing deck I used is more what you would use for a property acquisition.  It is a five year strip price (each month of that year averaged together), holding the last year constant in perpetuity.  If this was an SEC reserve report (official year end report to the government), you would use the average of the first day of every month for the past year and hold that price constant forever.

The other assumption in this run is that there are 2 rigs drilling and completing 1 well each per month for 1 year, starting Jan. 1, 2020, for a total of 24 wells.

Tamr is Going to Save you Some Serious Cash

Now that we have stated our assumptions and put all of the data into PHDWin,  we can run our cash flow model and determine the value of adding 24 more wells to the area of interest.  Just running these numbers, with no benefit from Tamr we get:

Not bad.  The project has an IRR of 44.47%, a PV 10 of $114.0 million, and pays out in 2.3 years.  Those are solid numbers, but we can do better.  You can use Tamr to master all of the data related to service vendors and suppliers.  Many companies don’t realize they are buying the same parts and services from multiple providers.  Like having clean, unified data for wells, having financial and operations costs in clean, working order can do amazing things for an organization’s budgeting and flexibility on projects.

Consider a scenario where you employ Tamr to master that data where you can save on both lease operating expenses and well capital investment.  In terms of LOE, a company can easily nail down who fairly prices water disposal, product transport, and pump maintenance.  To go along with that, on the drilling and completion side, you find a slightly cheaper sand provider and identify a better water vendor.  That leads to saving 5% on both LOE and D&C capital investment. Edit the PHDWin model to reflect the savings and that yields a new IRR of 56%, and PV10 of $131.6 million.  With Tamr software you were able to boost the IRR by 26% and the PV10 by 15.5%, a value increase of $17.7 million.

What could a company in this situation do with that extra value?  Have room in the budget for 2 more wells? Not question running more expensive logs in some test wells? In another exercise the company can run SEC reserve pricing, take whatever that value is, and borrow from a reserve based lender against it.

Keeping the LOE savings the same (5%), and building off of the D&C cost savings, in this next scenario let’s conservatively add in that the company was able to get another 5% savings on D&C.   That savings across a mix of better identified services, parts, and equipment vendors and improved completions from mastering all of their well data, along with public data, to understand the target formation better – totaling a 10% cost savings on the D&C cost.  That price sensitivity run looks like:

Between the run without Tamr and this run, the company is gaining ~$28 million in value…in one year of drilling…and a 66.7% rate of return.  That is a 50% increase on IRR and a 25% increase on PV10 value.

So, to summarize the 3 scenarios:

No-Tamr = Just our normal assumptions on economics

Tamr-5 = 5% savings on both LOE and D&C costs

Tamr-10 = 5% savings on LOE, 10% on D&C costs

Final Word

In this post, we covered what “type curves” are and how to create them.  You also saw a majority of the information that goes into building a financial model for oil and gas development projects.  Most importantly, we went over how Tamr can add significant value to a company’s reserves at the cost of an incredibly small fraction of the value a company will gain.

‍