In an analysis of 1,200 IT projects completed over the past 5 years, on average, these projects turned out 8% under their approved budget and took 21% longer than their approved schedule. This means that too much money and resources are claimed and claimed resources become later available then planned. Too much money is left on the table that could have been allocated to other business opportunities.
Looking forward there are two trends that demand for faster delivery of IT projects:
Increasing pace of IT evolution
Shorter time windows for new business opportunities
To enable faster delivery of IT projects ExxonMobil is introducing P50 probabilistic project cost and schedule estimating for classified estimates. Classified estimates are estimates within the envelope of uncertainty when it is possible to make risk-based predictions of a possible range of the project outcome in both cost and schedule. For a single project the P50 estimate means a 50% probability of cost and schedule overrun. When these projects are accumulated into a portfolio the underruns and overruns even out to a level where money and resources are used as efficiently as possible.
Introducing P50 estimates is a culture change where overruns within the allowable range are accepted and results outside the acceptable range should be treated with equal scrutiny whether it’s an overrun or an underrun.
Probabilistic project cost and schedule estimating - Michel Sterling - NESMA 2013
Probabilistic Project Cost and Schedule Estimating
NESMA Najaarsconferentie in Baarn, NL - 21 Nov 2013
Michel Sterling – ExxonMobil IT
Case for action
Cost and Schedule Essentials
What is P50 estimating
Why P50 ?
Change Management Challenges
Case for Action
On average, IT projects completed over the past 5 years turn
out 8% under their approved budget and take 21% longer
than their approved schedule.
Opportunity to leave less “money on the table”
Trend : Increasing pace of Information Technology evolution
Trend : Narrower windows of business opportunities
Demand for faster delivery to capture full
benefits of business opportunities.
Before we jump in…
“Everything that can be counted does not necessarily count;
everything that counts cannot necessarily be counted.” –
“There are three types of lies : Normal Lies, Damned Lies and
Statistics” – Mark Twain
Cost and Schedule Essentials
• Basis definition improves as project progresses through
project stages, increasing confidence in the range of possible
• Unclassified Estimates : No historical statistical data available to
• Classified Estimates : Risk-based predictions of possible range of
Estimate Classification Concept
What is P50?
Actual / Estimate Cost Ratio
The P50 cost value is a probabilistic estimate of the project cost
based on a 50% probability that the cost will be exceeded.
Cost Estimate Performance
Projects of a portfolio are considered at the mean of a simulated
cost distribution, typically the P50 estimate.
Why P50 ?
For a project : an equal likelihood to complete under or over
the appropriation basis (cost and duration)
Maximize individual project’s predictability, avoid surprises
Not wasting, not shortcutting; Net zero impact
Avoid overly conservative estimates that encumber resources
Capital that can be used elsewhere
Resources that can be deployed elsewhere
Requires thorough comprehension of Risk events,
Opportunities and their impact on Cost and Schedule
Level of experience
Assessment of complexity
Within a Portfolio :
8 out of 10 projects to complete within a given confidence range of the
appropriation basis cost and duration.
Routine Project’s within +/- 10%
Known Risks, low uncertainty
One-off or first time
New Technology, Geography, Process
Number of Projects
Non-Routine Projects within +/- 20%
Managing the Change
A over-run within the allowable range (between 100-110% of
appropriation basis) is as good as an under-run (between 90-100%)
A result outside the confidence band should be treated with equal
scrutiny whether it's under or over
The confidence band should be broaden when there is more inherent
risk in the estimate : A result between 100-120% of appropriation basis
is as good as a result between 80-100%
Gate 2 – “Scoping” should represent a P50 too…
Process & Tools to implement P50
Norms and tools calibrated to P50
Cost and schedule controls earlier in the project life cycle
Risk management and contingency planning