AFE and cost against plan, from estimate to close-out

An AFE is the money a well is allowed to spend, agreed before the rig moves. Its value on the rig is as a yardstick: it tells the team each morning whether the well is heading for its budget, and early enough to do something about it. This article builds an AFE for a horizontal development well, tracks it through 44 daily cost reports to release, and shows how the estimate at completion moved, why, and when a supplement was the right call.

What an AFE is for

An authorisation for expenditure is the approved budget for one well: a list of cost lines, a contingency for the risks the plan knows about, and the assumptions behind both. Once it is approved, the well may spend up to its total. Anything beyond that needs a supplement, a second approval for new money with its own reason. The authorised total is the AFE plus its approved supplements.

Three numbers matter while drilling: what has been spent, from the daily cost reports; what was planned by today, from the AFE spread over the drilling programme; and what the well will cost when it is finished. The third, the estimate at completion, drives decisions, and it is the one most often worked out badly.

Building the estimate

Most of a well's cost follows one of two drivers. Time-dependent lines are charged for every day the rig is on the well: the rig's operating rate, mud engineering, mud logging, rentals, communications, fuel, water and camp. One-off lines are charged when something is bought or a job is done: casing, cementing, bits, wireline logs, the wellhead and the completion. Some services sit in between. Directional drilling, LWD and mud chemicals are charged daily, but only while their hole section is being drilled.

D

planned days, from the rig move to rig release, taken from the drilling programme

r_j

daily rate of each time-dependent line (USD a day)

L_k

each one-off line, including services priced as a lump sum per section

C

contingency: an amount for the named risks the plan carries, with its reason written down

Lines are grouped by class (pre-drilling and mobilisation, drilling, completion) and by cost code, so the daily cost reports can be booked against them. Giving each one-off line its hole section phases the budget: the 9⅝″ casing is planned for when the 9⅝″ string runs. That turns an AFE total into a planned cost curve.

The contingency covers risks the plan names and prices, such as losses in a known zone. New scope, such as a longer lateral or an extra casing string, is a change to the well and gets a supplement, approved by the people who decided the change.

The example well

The example is a horizontal oil producer in a Lower Cretaceous carbonate reservoir, set in typical onshore Abu Dhabi conditions: a development well drilled from an existing pad, with a 17½″ surface hole through near-surface carbonates prone to losses, a 12¼″ intermediate hole, an 8½″ build section landed at 90°, and a 6″ drain completed with swell packers and inflow-control screens. The plan is 40 days from the rig move to release.

The AFE engine also splits the lines two ways. Tangible equipment that stays in the well (casing, wellhead, completion, tree) is 1,820,000 and intangible spend 3,773,000. The dry-hole cost, without the completion, is 4,833,000.

A hand check of the build-up

1. The daily spread. The time-dependent lines add up to 32,000 + 1,500 + 1,200 + 2,500 + 300 + 6,000 = 43,500 USD a day.
2. Over the planned days. 43,500 × 40 = 1,740,000, which is 31 % of the lines. Each day the well runs late costs at least that day rate, before any section service.
3. The rest. 5,593,000 − 1,740,000 = 3,853,000 of one-off lines and section services.
4. Contingency. 280,000 is 5.0 % of the lines, sized for one losses event in the surface hole: about two rig days of spread, two cement plugs and the LCM.
5. After the supplement. On day 30 the drain was extended by 300 m, to 4,260 m. Supplement S1 added 241,000 (two rig days, two days of directional and LWD, drill-in fluid, screens and packers for the extra length, the daily services) and 2 days. Authorised: 5,873,000 + 241,000 = 6,114,000, over 42 days.

Tracking cost against the plan

The well had two NPT events. At 380 m in the surface hole it lost returns: 26 hours of LCM pills, two cement plugs and drilling blind with water and sweeps. In the drain the rotary steerable tool failed at 3,420 m: a 22-hour round trip. Both are logged on the daily drilling reports, which give the engine the depth and NPT for each day.

The actual line sits below the planned curve for most of the well, by as much as 600,000 on day 17, even though the well was a day behind from the surface hole on. That gap is a matter of timing. The planned curve spreads each section's one-off lines across the section's days; the cost reports book the 9⅝″ casing and its cement, 566,000 together, on the day they run. Compare the two at the end of each section, or compare the actual line with the authorised total.

By hole section the engine puts the surface hole at 6 days against 5 and 886,000 against 754,000 planned, with cost per metre up from 1,125 to 1,322 USD. The intermediate and build sections came in on their days and within 7 % of their money. The drain took 13 days against 9: two for the approved extension, one for the RSS failure and one because it drilled more slowly than planned.

The estimate at completion

There are two simple ways to forecast the final cost, and they fail in opposite directions, so it pays to run both. The remaining-plan method assumes every cost group still under budget will spend exactly what it has left; groups already over add nothing more. It reacts only when a group goes over. The day-rate method follows the pace: it takes the costs actually booked day after day and multiplies them by the days left, then adds the one-off lines still to come at their budget.

S

spent: the latest cost report's cumulative

r

daily rate: the average of the codes booked on the latest report and on at least 3 of the last 5

D_f, D_e

forecast days (planned days plus the days approved supplements add) and days elapsed

P_k, A_k

plan and actual of each one-off line not in the daily rate

On report 6, in the middle of the losses, the day-rate forecast jumped to 6,147,667, 274,667 over the AFE. The LCM had pushed mud spend to 26,667 a day, and the daily rate of 70,167 carried that over the 34 days left. As the LCM dropped out of the five-report window the forecast came back down. A jump is a prompt to look at the make-up before anyone reports the number.

Through the 12¼″ section the forecast swung between 91,500 under and 242,250 over the AFE, because directional and mud charges were in the daily rate while the section ran. Before S1 it crossed the AFE on 16 of the 27 reports from report 3. The remaining-plan forecast stayed at 5,593,000, the lines without contingency, until report 25, and never crossed. Against the final cost, from report 3 on, the day-rate forecast was 146,000 out on average, sometimes high and sometimes low; the remaining-plan forecast was 306,000 out and always low. Neither is enough alone, so both belong on the same screen. Here is the day-rate forecast checked by hand at report 30, the day S1 was approved:

1. Spent. 4,218,000; authorised 6,114,000.
2. Daily rate. The base 43,500, plus mud 13,400, directional 19,600 and LWD 12,000: 88,500 a day.
3. Days left. 40 planned + 2 from S1 − 30 elapsed = 12. Time cost: 88,500 × 12 = 1,062,000.
4. One-off lines left. The second 6″ bit 28,000, the lower completion 555,000 (with S1's screens), tubing and upper completion 170,000, the tree 110,000, and 4,000 the 9⅝″ cement job left unspent: 867,000.
5. Forecast. 4,218,000 + 1,062,000 + 867,000 = 6,147,000, 33,000 over authorised. The engine gives the same.

The well finished at 6,019,000, so that forecast was 128,000 high. Two errors pulled against each other. The daily rate carried the drain's mud, directional and LWD, 45,000 a day, for 12 days, but those services ran for at most 9 more: 211,000 too much. And the forecast counted 12 days left when the well ran 14, because the 48 hours of NPT were never added to the forecast days: 87,000 too little. The cement job's 4,000 that was never spent makes up the rest. A forecast is only as good as its day count and its split between daily and one-off costs.

Where the money went

Casing, completion, logging, bits and the wellhead ended on plan. The overruns were in time and in the losses. The rig ran 64,000 over (two days beyond the 42 authorised), cementing 58,000 over (the two plugs, 55,000, and a 13⅜″ top job, less 4,000 left on the 9⅝″ job), and the daily services 23,000 over for the same two days. Water haulage while drilling blind, 18,000, had no line in the AFE at all: the cost control check lists it as unbudgeted spend. Mud chemicals were 16,000 over and directional 6,000.

Those overruns add up to 185,000, the contingency's use: 66 % of it, with 95,000 left. Spent at release was 6,019,000, 98.4 % of the authorised 6,114,000, over 44 days and 4,230 m: 1,423 USD a metre. NPT priced at the rig rate, 48 hours at 32,000 a day, is 64,000, or 1.1 % of the spend. That leaves out the spread, the plugs and the LCM, so use it to compare wells and the cost codes to explain them.

Contingency or supplement

The losses were a risk the AFE named, so the contingency paid for them. The drain extension was a change of scope, so it got S1 before the extra metres were drilled. The RSS failure fell in between: the contingency still had room, so no second supplement was needed.

Raise a supplement when the forecast, read from both methods and checked by hand, sits above the authorised total by more than its own noise, for a cause you understand. After S1 the day-rate forecast was over only on reports 30 and 31, by 33,000 and 12,400, about 0.5 %, with the drain's services in its daily rate. The right call was to watch it. Had it stayed over once those services stopped, a supplement for the gap would have been due before the completion was run.

Close-out

At release, freeze the final figures with the AFE: spent, authorised, days, cost per metre, NPT hours and contingency use, so later imports cannot change them. Then explain every code that went over or was spent without a line, with its root cause and the lesson. Here: budget water haulage for the surface hole, and size the next well's losses contingency on this one's 185,000. Daily reports that book the right code each day make the loop work; automating and checking them is covered separately, as is how NPT is classified.

A checklist for the rig team

• Split the AFE into time-dependent lines and one-off lines, and give each one-off line its hole section.
• Book every daily cost to the AFE's codes and subcodes; give one-off charges spread over several days their own subcode.
• Read actual cost against the planned curve at section ends, and against the authorised total every day.
• Run the day-rate and remaining-plan forecasts side by side, and check the daily rate's make-up when they jump.
• When NPT adds days, add them to the forecast days; when a section ends, drop its services from the daily rate.
• Pay named risks from the contingency; raise a supplement for new scope before the work is done.
• Explain every overrun and unbudgeted code with a root cause, and close the AFE with its final figures frozen.

How DrillIQ handles it

In DrillIQ the AFE & Cost module builds the AFE from a template of standard cost codes, with classes, hole sections, contingency, assumptions and the tangible and dry-hole splits, and moves it through submission, approval and activation, with supplements approved one by one. Each day it reads the cost reports and DDRs from Daily Ops Reports and shows spend against the authorised total, the planned curve phased by the drilling programme's sections, both estimates at completion with their make-up, cost and cost per metre by hole section, NPT cost by category, unbudgeted and over-plan codes, and a variance register with root causes and lessons. At close it keeps the final figures with the AFE. Every number above came from the module's engine.

Book a demo → See all capabilities

References

1. Bourgoyne, A.T., Millheim, K.K., Chenevert, M.E. and Young, F.S. Applied Drilling Engineering. SPE Textbook Series Vol. 2, 1986. Chapter 1, drilling cost analysis.
2. AACE International. Recommended Practice 18R-97: Cost Estimate Classification System, as applied in engineering, procurement and construction for the process industries.
3. IADC Drilling Manual, 12th ed. International Association of Drilling Contractors, 2015.

About this example. The well in this article is a representative scenario set in typical onshore Abu Dhabi, UAE conditions and worked through with DrillIQ's engines. It is not a client's well, and its figures are not field results.