Allowing for Uncertainty in Turnaround Cost Estimates
Turnaround cost estimating is notoriously poor. Previous studies have revealed systemic under-estimation of costs and over-confidence regarding the tightness of the accuracy ranges assigned to estimates.
Part of the problem is the general lack of rigor in considering how to allow for uncertainties and risks in the estimate. There are essentially four areas of uncertainty that need to be considered in a turnaround estimate:
- Emergent Work Scope
- When compiling the cost estimate, the estimator prices the exiting “know” scope on the worklist for the turnaround event. But inevitably, there will be additional scope that “emerges” between the compilation and freezing of the worklist, and the start of the execution of the turnaround event in the field. This scope emerges because things tend to break unexpectedly and need to be added to the list of stuff to fix.
- Discovery Work Scope
- This refers to scope that was not on the worklist, but is “discovered” during execution of the turnaround. Examples include issues such as opening a column for routine inspection, but discovering that half the trays have collapsed; or opening a furnace and discovering that the amount of damaged refractory is far more extensive than expected.
- Known Estimating Uncertainty
- This refers to line items in the estimate that we know have some uncertainty around them. Examples might be the cost of materials we haven’t yet bought, or the exact number of labor hours for an activity.
- Known Unknown Estimating Uncertainty
- This is what in the project world is known as “contingency”. It’s money that historical experience has told you will probably be needed to cover cost growth on something, but you’re not exactly sure which line-items in the estimate it’ll be spent on.
Unlike in capital project estimating, escalation is less of an issue for turnaround estimates, since the timescales are much shorter, giving less time for escalation to impact prices.
Under capital project rules, the first three would typically be estimated “above the line” and an allowance added as necessary to each relevant line item in the cost estimate. Only the “pure” contingency would be listed “below the line” as a separate “bucket” of money in the estimate.
However, Turnaround estimating is currently less rigorous in its approach. The most common industry approach is to lump all four elements into a “below the line” bucket, labelled with the slightly inaccurate name of “contingency”.
In addition, many teams do not use any kind of calculation to decide the size of that contingency bucket. Instead, their own company rules often dictate that they are only allowed to include a flat 10% of the total “above the line” estimate as contingency. How that 10% was arrived at is unclear, but a likely explanation is that it arises from confusion about what it means to say that a cost estimate has an accuracy range of ±10%.
So, in order to improve the accuracy of Turnaround estimates we could consider doing the following for each of those four categories:
- Emergent Work Scope
- Review the historical data on how much scope has “emerged” every month in the past. (Or in cruder terms, think about how many things break every month, on average.).
- Use this, combined with the time remaining between now and the start of turnaround execution, as your guide to how much should be allowed in this category.
- Discovery Work Scope
- Ask your inspection team and your process technology team to make their best guess, based on current inspection data and plant performance data, of where and what damage there might be, hidden in the equipment, waiting to be “Discovered”.
- Use this as your guide to how much should be allowed in this category.
- Known Estimating Uncertainty
- For each item in the known scope, consider what level of uncertainty there is around quantities and prices. For example:
- Do we think the productivity might vary and hence our estimate of labor hours for each scope list item might be wrong?
- How firm are the prices for these materials? Do we have a good quote, or did we just phone the vendor and ask for a rough guess?
- Use this as your guide to how much should be allowed in this category.
- Known Unknown Estimating Uncertainty
- How inaccurate have our estimates been in the past?
- What does our risk register look like? Does it have some likely risks that have cost consequences?
- What is the level of detail and accuracy in the scope definition? (i.e. Consider the estimate “classification”; is this an early, top-down, rough estimate, based on first guesses; in which case, you need more contingency: or is it a detailed, bottom-up estimate, based on firm drawings and quotes; in which case you need less contingency.)
- If accuracy is extremely important to you, consider running an uncertainty and risk workshop, with monte-carlo modelling.
- Use this as your guide to how much should be allowed in this category.
By considering each of the four elements in turn, and by considering the detail of each element, you should firstly achieve a far more accurate estimate than the simple fixed percentage of 10%, and you should have a clear and defendable argument that you can use with management when you (inevitably) have to defend your estimate.
If you would like help with improving your turnaround cost estimating procedures and processes, contact Becht for support.