In this first installment of our project estimation deep dive series, we show you the most common reasons midstream companies end up with inaccurate estimates. By the end of the series, we'll provide you with specific tips, tools, and strategies for addressing every one of those challenges.
Wondering why project estimates are so important? Click here to learn why accurate estimates are the cornerstone of any robust investment strategy.
PowerAdvocate has worked with many of the world's leading midstream companies, and when it comes to project estimates, we've seen it all. So we know that there's nothing more challenging than developing a defensible project estimate when you lack the data to predict the future. Here's four of the most common ways that midstream companies end up with inaccurate estimates (and some tips on how to solve them):
1. The Experience vs. Data Dilemma
Go with your gut or go with the data? It's the perennial question that any project cost estimator must face.
And it's a nearly impossible one to resolve, as midstream companies face the choice between:
- Relying on personal knowledge fraught with biases, lapses, and errors of memory
- Relying on project data that's inaccurate, inaccessible, and incomplete
Given the lack of options, we've seen many midstream companies take bits and pieces from each: a spreadsheet that sums up the costs of a few categories, a few phone calls to Project Managers to gauge past project costs, a review of the most memorable past overruns, or recourse to rules of thumb like the old "materials are 35% of costs" rule.
But it's precisely because midstream companies are left to these insufficient devices that we've seen project misses of +/-40% in the industry in just the last year. Experience can yield best guesses, but best guesses just aren't good enough for estimating projects worth 10's or 100's of millions of dollars. At scales like those, even the smallest margins of error can have impacts worth millions.
The data is often no better: it's out of date, inaccurate, miscategorized, and riddled with errors. In fact, we've worked with customers who've had up to 50% of their projects categorized in the wrong accounting or WBS codes. It's no wonder that estimators end up defaulting to past experience when they can't trust the numbers they're working with.
Wondering how to improve historical project data? Click here to read our how-to guide.
2. The One-Size-Fits-All Obstacle
Not all projects are the same. So how can midstream companies estimate future projects that don't exactly replicate projects from the past?
In order to generate accurate estimates, estimators need to be able to calculate the relative impact of a series of project-specific cost drivers, from the rockiness of the soil to the affluence of the surrounding areas to the season in which work will occur.
The magnitude of the problem is overwhelming: our research shows that more than 10 cost drivers, from diameter to percent grade to population density, can affect project costs by a total of more than 600%. With statistics like those, insight into historical project data quickly becomes erased.
But it's nearly impossible for any single company to generate a large enough number of data points to conduct precise statistical analyses and quantify the impact of each cost driver. And without normalizing projects using hard math, estimators either end up working off inaccurate baseline assumptions or defaulting back to experience. The result is that even a wealth of project data becomes useless when it can't be brought to bear on new project scenarios.
Wondering how to begin normalizing projects? Click here to read our guide to the 6 most overlooked project cost drivers.
3. The Process Hurdle
There's a reason why the Government Accountability Office calls centralized estimating processes a "best practice" for estimation teams. We've too often seen project estimates generated Project Manager by Project Manager, each of whom has to replicate the entire process from start to finish.
Not only does that kind of decentralization multiply the effort of creating estimates, but it's also the quickest way to generate a wealth of inaccurate estimates all biased in their own ways. That's because no one person has insight into the full scope of project costs and because different estimators operate under different assumptions, use different data, and have different analytical biases.
It's not just team decentralization that causes project estimate inaccuracy. Decentralized data collection processes too often mean that estimators rely on calling contacts from engineering firms to provide cost information or on contacting competitors for project insights.
The result is that executives end up with inconsistent estimates based on data from unreliable sources. With no agreed-upon, centralized way to estimate projects, entire investment portfolios are being driven by outdated data, inexact rules of thumb, and idiosyncratic insights.
Interested in centralizing your estimation process? Click here to read our tips for starting.
4. The Predictability Problem
Otherwise accurate estimates quickly become inaccurate when they don't account for future changes in market conditions. But how can project estimators predict the future when they already have limited visibility into the market conditions of the present?
We've seen many midstream companies resolve the forecasting problem by tacking on a static measure of inflation to the end of an estimate, but stop-gap solutions like those ignore the fact that different projects will be exposed to differing commodity changes, based on variables like project type, region, and mileage. For instance, PowerAdvocate predicts that welding labor costs will spike in the Gulf Coast in 2016 and beyond. When building projects in the region, midstream companies should be accounting for those region-specific wage spikes.
Without actively predicting specific market conditions months and years into the future, midstream estimates risk being based in an outdated and irrelevant past unrelated to the costs of the future.