APM – ACostE

Estimating Guide
For Project, Programme and Portfolio Managers
 Contents

Table of Figures v
Foreword vi
Acknowledgements vii
Purpose and reason for this guide viii
Estimating framework x
1 Developing the estimating plan 1
1.1 Stakeholder engagement and mobilisation 1
1.1.1 Stakeholder commitment 1
1.1.2 Responsibility assignment matrix (RAM) 2
1.1.3 Interpretation of stakeholder objectives and targets 2
1.2 Understanding the estimate scope 2
1.2.1 Scoping the estimate 2
1.2.2 Understanding the estimate maturity requirements 5
1.3 Manage the information requirements 6
1.3.1 Identifying the information requirements 6
1.3.2 Storing information 7
1.3.3 Reviewing the maturity of the information available 7
1.3.4 Revising the ADORE in line with information
maturity risk 8
1.3.5 Data normalisation 8
1.4 Prepare the estimating plan 9
1.4.1 Agreeing the estimating approach(es) to be used 9
1.4.2 Agreeing the competency requirements of the team 9
1.4.3 Agreeing the schedule of estimating activities and
project dependencies 9
1.4.4 Updating the estimating RAM 10
1.4.5 Defining the toolsets and information communication
media 10
2 Creating the base estimate 11
2.1 Estimate management 11

iii
Contents

2.1.1 Iterative process 11

2.1.2 Configuration control 11
2.2 Estimating approach 12
2.2.1 Top-down approach 12
2.2.2 Bottom-up approach 14
2.2.3 Relying on the work of others – ‘ethereal’ approach 16
2.3 Estimating methods 17
2.3.1 Analogy 19
2.3.2 Parametric 20
2.3.3 Trusted source 20
2.4 Documenting the basis of estimate (BoE) 21
3 Risk, opportunity and uncertainty assessment 23
3.1 Assessing uncertainty in the baseline activities 23
3.2 Link to the risk management process 24
3.3 Link to schedule risk analysis 26
3.4 Holistic review of risk, opportunity and uncertainty 27
3.4.1 Top-down approach 27
3.4.2 Bottom-up approach – Monte Carlo simulation 29
3.4.3 Dealing with inherent optimism bias in risk and
opportunity 32
3.4.4 Taking a balanced view of risk, opportunity and
uncertainty 33
4 Overall estimate validation, challenge and approval 35
4.1 Estimate maturity review 36
4.1.1 Maturity definition 36
4.1.2 Maturity assessment 38
4.1.3 Why is maturity assessment important? 41
4.1.4 Estimate review 41
4.1.5 Estimate approval 43
5 Estimates to completion 44
5.1 EAC top-down approach 45
5.2 Bottom-up approach 46
5.3 Risk and opportunity management in EAC 46
6 Ethics in estimation 47
Glossary 49
Acronyms and abbreviations 55
References 56

iv
 Table of Figures

0.1 Estimating framework x

2.1 Top-down example 13
2.2 Project life cycle 14
2.3 Bottom-up example 15
2.4 Potential suitable use of ethereal approach 17
2.5 Estimating methods 18
2.6 Example two-factor analogy 19
2.7 Example parametric with two cost drivers 21
3.1 Different types of distribution 24
3.2 Schedule risk analysis 27
3.3 Time phased cost profile 28
3.4 Realistic distribution 29
3.5 Monte Carlo 30
3.6 Risk, opportunity and uncertainty evaluation 31
3.7 Monte Carlo optimistically biased 31
3.8 Comparison of bottom-up and top-down approaches to
risk, opportunity and uncertainty 33
3.9 Rolling wave estimating 34
4.1 Example of maturity definition 37
4.2 Example of estimate maturity assessment 38
4.3 Example of maturity assessment 39
4.4 Example of a maturity scale 40
5.1 Earned value S-curve 44

v
 Foreword

Estimation and its diligent application within any project is one of the cornerstones
of successful project delivery. This guide has been created to assist the current
and next generation of project stakeholders to understand the core values,
information sets and underpinning knowledge that, if applied diligently, will
improve the clarity and robustness of an estimate, informing the organisation
with transparency and clarity and supporting better decision making.
The Association for Project Management (APM) and the Association of Cost
Engineers (ACostE) have collaborated to bring this guide to you. All the
collaborators have an excellent understanding of the challenges faced when
generating an estimate that is fit for purpose, having had the experience of real-life
situations where good estimates made the difference in delivering the project, and
a burning desire to share their combined knowledge for the benefit of all.

Professor Andy Langridge

Director of business development, ARES Corporation

vi
 Purpose and reason for
this guide

The purpose of this guide is to provide a fundamental understanding of the

methods of cost estimating and it explains a number of standard approaches
available to promote good practice.
Estimates are critical to the project manager as they are needed to make
informed decisions about projects across the different stages of the whole project
life cycle, hence the cooperation of the Association of Cost Engineers (ACostE)
and the Association for Project Management (APM) in publishing this guide. In
project management, effective monitoring of a project’s performance depends
on having an appropriate, high-quality estimate against which progress can be
measured.
For instance:

n Investment committees require estimates in order to predict return on

investment and hence determine whether to support project proposals and
the level of finance to invest in them.
n Cost estimates form part of option appraisals.
n Mature organisations often review projects at key stages and require them to
meet internal governance guidelines on cost-benefit-risk.
n Organisations that manage a portfolio of projects require credible cost
estimates in order to manage the spending profile of the portfolio.
n Good management practice needs to understand the impact of a project
change, including risks and opportunities.

There is long-standing evidence that underestimation of project costs is a key

reason for project failure. For instance, the National Audit Office’s 2001 report
Modernising Construction (Comptroller and Auditor General, 2001) found that
limited understanding of the true cost was one of the main reasons that 70 per
cent of public sector construction projects were delivered late or over budget.
Industry surveys (KPMG, 2015) indicate that this is still the case. Academics
have collected evidence for overruns costing billions of pounds in major

viii
 Purpose and reason for this guide

infrastructure projects worldwide (Flyvbjerg, 2003). More recent studies

(Comptroller and Auditor General, January 2017) have shown that these errors
were often due to not taking estimating seriously enough, hampered by poor
quality data and unrealistic assumptions (National Audit Office, December 2013).
By their very nature estimates are speculative; the word estimating is
synonymous with approximation and guessing, yet estimates are vital for sound
decision-making, planning and financial management.
Different techniques may be appropriate at different stages in a project’s
development. This guide will focus on cost estimating method and approaches.
However, the advice in this guide is not limited to initial cost estimates; it is equally
applicable to forecasting and to other forms of estimating, for example, time,
schedule or performance.
The guide is not limited to public sector construction, infrastructure or defence
projects; it is equally applicable to the private sector and a wider range of projects,
for example, if you are recruiting a new member of staff, or building the world’s
fastest car, launching a new service, or licensing a new drug – every project needs
a robust estimate.

ix
 Estimating framework

Figure 0.1 Estimating framework

Estimating consists of a number of activities, which provide a framework for

generating and continuously improving an estimate. The diagram above shows a
typical estimating framework, which includes the activities covered in the guide.
A definition of the terms used in the diagram above can be found in the
glossary and are explained throughout this guide.

x
 2

Creating the base

estimate

2.1 Estimate management

2.1.1 Iterative process
Before starting any cost estimate, it is important to understand that it will be an
iterative process. It is extremely unlikely that all information will be available at
the same level of maturity at the start of the project. This means that the estimator
must manage the estimating process as an iterative one. The estimator will
progress through the various stages of the estimating process (see Figure 0.1:
Estimating framework), and at each stage the individual’s knowledge will develop
as more up-to-date information becomes available. If the estimator plans for an
iterative process and operates configuration control, it will facilitate these changes
in a controlled manner.

2.1.2 Configuration control

A key principle of cost estimating is the ability to trace and document all aspects of
the analysis, from raw data to final outputs. To achieve this, robust configuration
control is required. Considering that estimating is an iterative process, configuration
control will provide a method to ensure that the most current information is used.
It is important to tailor the extent of configuration management required. For a
simple project with a small number of data points, a simple naming convention
could be sufficient to control the configuration (e.g. yyyy.mm.dd [name] v1.n).
Whereas for more complex projects, full baseline management and change control
may be required to ensure all information is configured appropriately. APM
provides comprehensive guidance on configuration management which can be
tailored to the needs of the estimating process; see APM Body of Knowledge 7th
edition (APM, 2019).

11
Estimating Guide

2.2 Estimating approach

An estimating approach is the direction, or means of arriving at an estimate, and
to some degree implies the level of detail at which the estimate is created. With
complex projects, it is often considered to be good practice to create an estimate
using more than one approach as a means of providing a greater level of
confidence in the output advised, thereby testing the robustness and
interpretation of the data, the assumptions and the methodologies employed.

2.2.1 Top-down approach

In a top-down approach to estimating, the estimator reviews the overall scope of
a project in order to identify the major elements of work and characteristics
(drivers) that could be estimated separately from other elements. Typically, the
estimator might consider a natural flow down through the work breakdown
structure (WBS), product breakdown structure (PBS) or service breakdown
structure (SBS).
The estimate scope may be considered as a whole, or broken down to different
high levels of WBS as required (see Figure 2.1). The overall project scope must
be covered by the range of non-overlapping work packages selected, although
not all work packages need to be estimated at the same level of WBS. This allows
the estimator to use the maturity and/or uncertainty in the key information
available to produce the most appropriate level of estimate. The overall project
base estimate would be created by summing these high level estimates. This
should not be confused with the bottom-up approach where all the lower levels
would be aggregated.
Over the life of the estimate these higher-level work packages and the
associated higher-level estimates may be broken down into more detailed or
more refined elements, which ultimately will facilitate a bottom-up approach (see
Figure 2.2).
A top-down approach is frequently used for creating rough order of magnitude
(ROM) estimates, otherwise known as ball-park estimates, where the level of
detail available is limited. As a general rule, a top-down estimate requires less
time and effort to produce than one produced using a bottom-up approach.
Top-down estimates are appropriate at the beginning of the life cycle when large
numbers of alternative options need to be estimated and considered. As the
solution matures and more information becomes available, there is an increased
opportunity to produce bottom-up figures. However, a top-down approach

12
 Estimating framework

Figure 2.1 Top-down example

Source: © Alan R Jones 2019

can still be useful throughout the life cycle of a project, e.g. for validation
purposes.
The main benefit of working at a higher level is that there is a tendency to
use more holistic data from previous projects or products, including unmitigated
and unforeseen risks, and scope creep. This can reduce the risk of emerging
work activities or costs being overlooked. As a result, top-down estimates are
typically greater than those created by a bottom-up approach.
Base estimates created by a top-down approach should exclude consideration
of additional risks and opportunities. These should be considered separately by
either a top-down or bottom-up approach as part of the formulation of the project
baseline estimate. See section 3.
It is considered good practice to express an uncertainty range around a
top-down estimate, based on the maturity of the information available, and the
estimating methodologies employed. According to the NAO survival guide to
challenging costs in major projects, (National Audit Office, 2018) “Early cost
estimates should be presented as a range and never a point estimate”. Note that
APM and ACostE believe this should apply to all cost estimates.

13
 3

Risk, opportunity and

uncertainty assessment

3.1 Assessing uncertainty in

the baseline activities
There will always be uncertainty (or a lack of exactness) in the cost estimate of
projects. This happens for various reasons, including uncertainty about the
detailed scope of work and uncertainty about the levels of productivity that will
be achieved. In addition to these factors, there will be risks and opportunities;
these are things that may or may not happen, but if they do they will impact the
estimate. Risks are discrete events that will have a degree of uncertainty over the
exact value. Baseline tasks will occur, but the actual value will have uncertainty.
For example, while driving home (baseline task), the journey time is uncertain
due to variable traffic conditions (i.e. traffic lights). However, in addition there
could be a risk of a delay of uncertain duration due to an accident.
Uncertainty is the inherent and potentially uncontrollable variability in
estimating the actual cost and schedule. It can be considered as a tolerance band
on the understanding of the scope. Uncertainties arise because the organisation
does not have a complete understanding of the proposed task or the solution. An
uncertainty is an expression of something that will happen; the actual project
value will not be known but is expected to lie within a defined range. Some
uncertainties will express natural variation; for example, my journey home each
day varies by maybe five minutes less or 10 minutes more.
Most baseline tasks in a work breakdown structure will have uncertainty,
which can be expressed as three values: the minimum (unlikely to be less than),
most likely value and maximum (unlikely to exceed). The 3-point estimate should
express the range of uncertainty – excluding risks or opportunities.
Where there is uncertainty in the scope definition of the baseline activities, the
3-point estimate can be used to express the extremes of a minimum (or simplistic
scope requirement) and a maximum (or complex scope requirement).
Alternatively, any potential but improbable extreme in the scope requirements

23
Estimating Guide

might be expressed as a risk or opportunity. Care must be taken not to duplicate

or overlap any extremes expressed in both ways.

3.2 Link to the risk management process

The evaluation of the net effect of risks and opportunities cannot be performed
in isolation from the baseline activities, nor can they be evaluated in isolation
from the project’s risk management process (i.e. do not re-invent the wheel).
Most of the data requirements to manage risks and opportunities are also needed
to evaluate their net impact; for example, 3-point estimate of the individual
risk or opportunity cost impact, probability of occurrence, risk retirement date,
mitigation plan, etc. In addition to these parameters, in order to evaluate their net
impact, there is a need to express the 3-point estimate range of potential values
as a probability distribution. This then allows probabilistic modelling of the risks
and opportunities to be performed in Monte Carlo simulation in conjunction with
baseline activities.
Risks and opportunities are discrete events whose occurrences are expressed
by a probability of occurrence, modelled by a Bernoulli distribution.
The most common forms of probability distributions to model uncertainty
include normal, lognormal, exponential, triangular and uniform.

Figure 3.1 Different types of distribution

24
 Estimating framework

When to use these different types of distribution:

Normal distribution:

This symmetrical distribution often represents the spread and frequency of

values in naturally occurring observations in nature, such as the height of
adult males or females of a given ethnicity. It can also represent the distribution
of values from man-made systems such as the accuracy and/or precision of
machining operations. The distribution can often be used to represent the
range of values for system or sub-system level costs, even where the
constituent elements of those systems or sub-systems are not normally
distributed. The scatter or deviation of values around a linear cost estimating
relationship is also expected to be normally distributed.

Lognormal distribution:

This distribution can be used in reliability analysis to model the repair time of
items, in particular in relation to the fatigue-stress characteristics of
mechanical systems. It is often an empirical distribution observed in natural
growth or human behaviour systems. When a variable is deemed to be
lognormally distributed in linear space, its values will be normally distributed
in logarithmic space. As a consequence, the scatter or deviation of values
around a power or exponential cost estimating relationship is expected to be
lognormally distributed.

Triangular distribution:

The triangular distribution is frequently used as a default distribution where

there is some knowledge or perception of a most likely value, and also an
appreciation of the likely minimum or maximum values of a variable. In cost
and schedule scenarios, data is more likely to be positively skewed; i.e.
where the difference from the most likely to the maximum is greater than the
difference between the most likely and the minimum. Aggregated system
level variables are more likely to be symmetrically distributed.

Uniform distribution:

The uniform distribution is frequently used as a default distribution where

there is some knowledge or perception of the likely minimum or maximum

25
 6

Ethics in estimation

The estimating codes of conduct lay down the behaviours we would like to see
from both the estimator and the customer of those estimates. The customer is
defined in this guide as the person who asks for the estimate: project manager,
chief engineer, etc.

1. No individual employee, team, organisation, project (or programme), or

vendor shall be required to develop, submit or certify any estimate for which
they do not have appropriate confidence.
2. There shall be means to address without retribution any concerns about the
integrity or ethics in the development of any estimate and those means shall
be communicated clearly.
3. In order to protect the integrity, security, image and reputation of the company,
senior leadership will confirm the compliance of their respective organisations
to estimation policy and standards, be held accountable for the same, and
shall delegate as appropriate levels of assurance and compliance to the
estimation policy and standards.
4. Any known impacts to estimates, including those for remaining costs of
projects in progress, shall be documented and reported as quickly as possible,
and no later than in accordance with documented policy.
5. Estimate values, changes and associated impacts shall be communicated
honestly, ethically and on a timely basis, to all customers, both external and
internal.
6. At all times, the estimator and customer shall create an environment of mutual
trust and respect. They shall provide open feedback and views without
criticism. At no time shall bullying, intimidation or disrespectful behaviour be
tolerated.

All professional bodies have a code of conduct which their members are expected
to follow. These include, but are not limited to:

ECUK Spec – Engineering Council www.engc.org.uk/ukspec

APM – Association of Project Management www.apm.org.uk

47
 Glossary

The terms in this glossary represent the views of both APMBok (APM, 2019),
ACostE and the authors of the guide.

3-point estimate/ [APM] An estimate in which optimistic best case, pessimistic worst
three-point case and most likely values are given.
estimate [ACostE] A three-point estimate represents three cases produced by
estimating. Some organisations (and this guide) refer to these as the
minimum, the most likely and the maximum.
The three-point estimating technique is used for the construction of
an approximate distribution representing the uncertainty of future
events; this will ensure the estimate is credible.
Accuracy The correctness of an estimate. This can be measured as the
percentage error between the estimate and actual. In the case of
3-point estimates, an estimate is considered accurate if the actual
cost/schedule lies inside the estimate uncertainty range.
ADORE Assumptions, dependencies, opportunities, risks, exclusions
(Shermon D, 2017)

Assumptions A statement that is taken as being true for the purposes of estimating,
but which could change later. An assumption is made where some
data is not available or facts are not yet known.
Baseline The reference levels against which a project, programme or portfolio
is monitored and controlled.
Bottom-up [APM] An estimating technique that uses detailed specifications to
estimating estimate time and cost for each product or activity. Also known as
analytical estimating. This should not be confused with the
‘Estimating Method of Estimating by Analogy’ (Section 2.3.1).
[ACostE] An approach to estimating all individual work packages or
activities with appropriate level of detail, which are then rolled up to
higher-level estimates. The accuracy of bottom-up estimating is
improved when individual work packages or activities are defined in
more detail. See section 2.2.2.

49