1

SMSHAR Tutorial

© ir JP Tollenboom 2014

2

Table of ContentIntroduction

Section 1: SchedulingSection 2: MonitoringSection 3: Portfolio

ConventionsSymbolsVideos

Section 1: SchedulingBefore you startProject Tree

SummaryConstructing the project treeClean and Simple TreeRecap

Task InventorySummaryHow to startHow to structureMore structureGranularityTasks mapTask namesObservabilityRecap

DurationsThis video shows how the working days of the week can be set.Weight factors

Weight factors explainedThe showing

LogicsA few definitionsTypes of dependenciesLeads and lagsExamplesGood practiceThe showing

Advanced topicsMilestone blocks

ProblemSolutionDriving MilestonesBuoysUsageThe showing

Packages

3

ProblemSolutionMethod 1Method 2Preferred methodMore on the absence of a sequenceExamplesRememberFinal remarkThe showing

AuditSecion 2: Monitoring Progress

A few disturbing words.What is monitoring about

The pastThe presentPercentage completeAggregated progressTwo defining parameters

How do we perform the monitoringMetrics

Binary metricFixed stepsCalculated progress HintSpecial case: tracking documents

Driving the reporting processS-­curves: the kernel of the progress reports

The basicsInspect the Daisy treeInspect the S-­curves

Identify the area in which the last point laysCheck the trend lineDiverging trackConverging trackParallel track

Action!Section 3: Portfolio

BasicsMethod

ReferencesTutorial videos

4

Introduction

This is a tutorial for the optimal usage of the integrated project scheduling and monitoringsystem, SMHAR.This system is based on Smartsheet for the setup of the project schedule.The project progress reports are generated by the DPC engine in a fully automatic mode.

It is assumed that the reader has set up an account for the usage of Smartsheet and theDPC reporting engine.

This tutorial has three major sections: Scheduling Monitoring Portfolio Monitoring

Section 1: SchedulingIt is assumed that the reader has no prior knowledge about project scheduling.This section will introduce the reader to a reduced set of functions enabling him to build aworkable schedule in a minimum of time.These topics will be addressed:

Project tree How to look at a project through a hiërarchycal structure

Task inventory How to set up a complete task inventory

Durations How to define task durations

Weight factors How to set weight factors and what they mean.

Logics How to define the workflow logics by linking tasks

Advanced topics A few advanced techniques in scheduling.

Audit How to self-­check the schedule

Section 2: MonitoringMonitoring is also often called “tracking”. To monitor a project or to track a project is the samething.In this section we explain why monitoring and scheduling always have to come in pairs:

If you don’t plan to track, don’t bother to plan.

We will recognize that a project is build from processes which are very similar to productionprocesses. We will understand that we can monitor the output of these processes usingsimple means. This output, in project management speak, is called the physical progress

5

and can be visualized with the so-­called S-­curves.An S-­curve is the timeline of the process’s output, as per schedule and as observed by themonitoring process.An S-­curve looks like this:

We will learn that only two pieces of information are sufficient to be able to asses the healthof the process being monitored:

its current position: the black dot its current average progress rate (progress speed): the slope of the trend line.

Section 3: PortfolioThis section will show how easy it is to monitor a whole bunch of projects once every singleproject is monitored with the system.

6

Conventions

Symbols

This symbol marks the beginning of a procedure: A procedure is displayed as a bulleted list:

step one step two …

A procedure is displayed as a numbered list when referencing to certain steps is necessary:1. step one2. step two3. if condition go to step 54. …5. …

A key item, an important note, is preceded by this symbol:

A warning, a “don’t”, is preceded by this symbol: ⇏

ExampleThis is the procedure to do something.

do this then do that this is a very important step

⇏ Don’t do this

VideosSome parts in the tutorial are illustrated with videos.Those videos are set up in HD.They can best be viewed in full screen mode and HD. Remember to set your video player to full screen and HD.

7

Section 1: Scheduling

The schedule defines the reference frame against which the project progress will be traced.It is therefore important to understand that a schedule should not be modified frequently.

Before you startYou will need the template Gantt. This template is called “Gantt SMSHAR (T)”.It is provided to the users as a shared sheet.

In Smartsheet, open the template and do “Use Template”. Create a new sheet by giving it a name of your choosing.

These are the columns of the Gantt template. You should get familiar with the definitions given in the next table.

1 Link to attachment Standard Smartsheet column

2 Link to discussion Standard Smartsheet column

3 Row action indicators Standard Smartsheet column

4 At Risk flag To mark a task “at risk”

5 Flag when set to “-­p”: drop this task when computing theprogress

6 Select Use this field to organise partitions.

Enter some value (x, y, z,…) then use the filter to filter tasks with same value

7 Task Name The name of the tasks.

use short precise names

8 Responsible The responsible of the task

9 WF The tasks weight factor.

8

Use integer values. When left blank, the task duration is the task’s weightfactor

10 Predecessors The task’s predecessors.

See subsection “Logics” in this section

11 Duration The task duration.The number of working days needed to complete thetask

12 Start date Computed start date of the tasksWhen set manually: the imposed start date

13 End date Computed end date of the tasksWhen set manually: the imposed end date

14 % Complete The degree of physical completeness of the taskexpressed in %.How much % of the task

has been produced, has been accomplished has been realized is visible can be observed can be measured

See subsection “Metrics” for clarification.

More columns: These columns are located to the right of column 14

15 Comment First row (the root) can contain a link to the ProgressReport page, providing direct access.

16 DLS Days left to start expressed in working days

9

computed from today negative value: date is passed due calculated field: need to fill the column down

when new rows are created error message will disappear as soon as a valid

date will appear in column 12

17 DLF Days left to finish expressed in working days computed from today negative value: date is passed due calculated field: need to fill the column down

when new rows are created error message will disappear as soon as a valid

date will appear in column 13

Project Tree

Summary

The project tree is the hiërarchical structure of the project. It is constructed by using theindenting and outdenting (demote -­ promote) of individual tasks. Its is easy to do, but it is alsoeasy to construct a complicated and confusing tree. So a set of simple rules must beapplied.

Constructing the project tree

In this video you see how easy it is to construct the project tree.Simply use the indent (outdent) button to move items to the right (indent) or to the left(outdent).It is in a way like using the tab key when editing text.

The next little list is an exact replica of what has been shown in the video. You recognise astructured list.

1. My Project1.1. Summary 1

1.1.1. a1.1.2. b1.1.3. c1.1.4. d

1.2. Summary 21.2.1. x1.2.2. y1.2.3. z

10

The codes preceding the items, e.g. 1.2.1, are called the WBS codes.WBS stands for “Work Breakdown Structure”. It is a fairly common term used in scheduling.

This is another way of representing the project tree: the hiërarchical structure.

Clean and Simple Tree

We prefer clean and simple trees to complicated and “unclean” trees.

Apply these rules when constructing the project tree

There can only be one item at the top level.It is the item with WBS = 1

This item is called the root of the tree. There can thus be only one root. The summaries -­ in the circles above -­ are called the nodes of the tree The tasks -­ the rectangles above -­ are also called

the atomic tasks the leafs of the tree

These are the tasks that have to be executed.Summaries are collection of tasks or collections of lower summaries(sub-­summaries)

Do not create too many levels if you run a small project, try to limit the number of levels to 3. if you run a larger project try to limit the number of levels to 5 all the tasks need not to be on the same level

Avoid hybrid structures: keep the tree clean.This is simple to understand, and still this one rule is cause of almost all structuredefects in schedules. So we will spend some time explaining and showing.

The explaining Summaries should be collections of

11

only atomic tasks, or only sub-­summaries, but not ⇏ collections of atomic tasks and sub-­summaries

In this case we have a hybrid structure. Remember this: hybrids will not explode the system, but will blur thereporting.

The showingThis video shows

How a hybrid structure looks like How to correct this

Take the time to clean up hybrids. It will pay back in easy of handling of theschedule and in providing un-­blurred progress reports.

Recap

The project tree is constructed by using the outlining facility: indent / outdent items. There should be only one root. Summaries can contain only atomic tasks or sub-­summaries. Hybrids should be avoided and eventually corrected.

Task Inventory

Summary

The setup of the tasks inventory is extremely important. Great care must be taken during thatprocess, so that a clean and manageable structure is constructed. Never start with opening afresh Gantt sheet: start with doing some thinking.If this part is not done in the right way, the penalty will be problems and difficulties over thewhole duration of the project.

How to start

The setup of the tasks inventory is not a trivial thing to do. One would expect that to be theeasiest part: not true.

Best practice is to start with the construction of the project tree by sketching it.Then to use this sketch as guideline for the setup of the schedule.In order words: never start by opening a new empty Gantt.Start by doing some thinking.

How to structure

It is clear from the description of the project tree, that at a certain point in time we have togroup atomic tasks into “summaries”.Sometimes we also group summaries into a summary of higher level.

The question now is: how to define such summaries?

12

The answer is: there are many ways.The structure must however serve a purpose:

it must facilitate the handling of the schedule it must create a clear and easily understandable decomposition of the project it must contribute to creating meaningful monitoring results it must satisfy the project manager’s need to control the project in a top-­down

manner. This in turn must avoid being drawn by the details it must conform to the internal culture of the project group. This means that the

project decomposition must produce small lumps of activities that comply with theway the project team is used to look at similar projects.

The last item in the bullet list is helpful to guide us: if it is the habit to think in phases, then we can use project phases to define the first

level of the project tree if the habit is to think in “units” -­ production lines, buildings, groups, machines -­ then

we can use such “units” to define the first level items

An example with phases tasks denotes any series of atomic tasks. Every tab denote one level in the project tree.

My projectPhase 1: Literature study

tasksPhase 2:Designing an experiment

tasksPhase 3: Purchase of components

tasksPhase 4: Building the prototype

tasksPhase 5: Performing the experiments and tests

tasksPhase 6: Reporting on the results

tasksPhase 7: Presenting the results

tasks

An example with “units”For the sake of the example, units are here taken as “”buildings”.Could be “Lines”, “Machine”, “Group of people”, …

My projectBuilding 1

tasksBuilding 2

tasksBuilding 3

13

tasksBuilding 4

tasksBuilding 5

tasksIn both examples, we have the atomic tasks at level 3.

More structure

When it is useful and meaningful, we can chose to create more complicated tree, i.e. withmore levels.⇏ Do not however create levels for the sake of it. Too much levels will contribute to moreconfusion, not less and contribute to complicated and often difficult to interpret, progressreports.

This is an example

My projectPhase 1: Design

Item 1: Buildingtasks

Item 2: Clima systemtasks

Item 3: Liftstasks

Item 4: Power distributiontasks

Item 5: Interior designtasks

Phase 2: ProcurementItem 1: Building

tasksItem 2: Clima system

tasksItem 3: Lifts

tasksItem 4: Power distribution

tasksItem 5: Interior design

tasksPhase 3: Construction

Item 1: Buildingtasks

Item 2: Clima systemtasks

Item 3: Liftstasks

14

Item 4: Power distributiontasks

Item 5: Interior designtasks

Phase 4: ReceptionItem 1: Building

tasksItem 2: Clima system

tasksItem 3: Lifts

tasksItem 4: Power distribution

tasksItem 5: Interior design

tasks

This example shows a project decomposition in phases, with every phase beingdecomposed in sections.The sections are here defined according to the dominant trade or specialty of the futureactors:

Building → Civil construction engineers, ArchitectsClima system → HVAC engineersLifts → Electro-­mechanical engineersPower distribution → Electrical engineersInterior design → Interior architects

This style of structure is very common.

Granularity

The next question is about the level of detail.How deep should we go into the detailing of tasks?In other words what is the smallest “grain” of activity we should plan for?Hence the term granularity.

Should we plan every bolt, every nut, every washer?Or should we plan on a factory unit level?

There is no fixed rule, but there are some good practice advises: if you go too deep in the details, you will eventually lose control by information

overload if you stay at a too high level, not detailed enough, you will eventually lose control by

lack of visibility

The choice lays in the hands of the project manager. There are some guidelines however: When choosing to define summaries and sub-­summaries think of processes.

Immagine that the summary, sub-­summary, that you want to define, is like a

15

production process, and that you want to keep the tasks together that are of the samekind. Doing, you will automatically define homogeneous processes. This is importantin relation to the monitoring. Why this is, is explained further in this tutorial.

A summary should always be the answer to one of the following questions do I need to be able to report on the status of this (part of the project)? do I want to track, monitor the progress of this (part of the project)?

Tasks map

The set of tasks must map the whole project.In other words, the schedule must tell the whole story. Down to the chosen granularity asstated above.⇏ It is a bad habit to construct a schedule that handles only part of the project, while thebalance is handled “on the side”.This is a guarantee for problems, and often ultimate failure.

Task names

The tasks names should be short and precise.

⇏ It is bad practice to define tasks that tell a story on their own. These are called narrativetasks. The story they tell can be so complicated, that ultimately no one will be capable ofmaking any statement about the % completeness.

Never hesitate to cut a narrative task into many small bits. The scheduler will cope.

Observability

This will sound trivial: the progress of every task of the schedule must be observable.We need this to be, because we will monitor the progress of every task by observing itsphysical progress and express it as a % complete. This is further explained in details in thistutorial.

Sometimes however, we need to add a few tasks to our schedule, for which tasks theprogress is not observable.E.g.: interventions by utility companies -­ gas, water, power distribution -­ that are not part ofour project. We will not be able to observe their progress. Still we need those tasks in ourschedule for general coordination purposes.The trick consists of flagging such task to “-­p”. This will cancel these tasks out in thecomputing of the overall progress values.

Flag “-­p” is mentioned in section “Scheduling / before you start / Column 5 “

Recap

Think before you schedule. Define the structure that fits your needs in terms of reporting and monitoring. Chose the granularity level that optimizes need to control and quantity of information

to handle.

16

The schedule is the whole project, no stuff “on the side”. The progress of every task must be observable;; exceptions are flagged “-­p”. Task names are short and precise. Narrative tasks to be split up in many smaller atoms.

DurationsEvery tasks must have a duration.The duration field (11) cannot be left blank.The duration is expressed in working days.This is not the same as calendar days.

A milestone is a task with its duration set to 0 (not blank).To define a task as milestone, set its duration to 0.

ExampleSay we set the duration of a given task to 10 (working days).

Assume that we declare that a week only has 2 working days.The task would then take 5 weeks of calendar time.

Assuming we declare that alls weekdays are working days (7/7), then the task wouldtake one week and 3 more days of calendar time.

This video shows how the working days of the week can be set.

When setting the task duration one can use one of the following methods: copy from a previous similar project copy from a similar project and multiply by a scaling factor (1+α), with |α| < 1, to take

local conditions into account. estimate using personal experience use figure provided by contractor estimate using this formula: 6

m + 4e + M

m: smallest reasonably possible valueM: largest reasonably possible valuee: expected value

impose a value: the contractor will have to adjust his resources in order to cope.

⇏ When the project is live, do not modify the durations all the time. Modify the durations if and only if it has been decided to reschedule the project.

Weight factors

Weight factors explained

The weight factors may be set in column 9.Only use integer values.The weight factor of a task expresses its weight, its importance, in the overall progress.The higher the value, the more this task will contribute to the overall progress.

17

By default, when the value is left blank, the task duration expressed in hours is taken as itsweight.

Weight factors are used to reshape the progress curves.Sometimes, the task duration create a distorted image of the progress curves.This is particularly the case when tasks have short duration (low default weights) but doinvolve large amount of resources (manhours).

Possible methods are: estimated manhours content abstract values, expressing the relative weights (task at 200 weighs double the task at

100) estimated head count estimated $ value of the task. In this case a kind of earned value method is used.

Estimated values are good enough.⇏ One should not spend lots of time looking for high-­precision values: it would be a loss oftime.

The showing

This video shows how to set the weight factors.

LogicsIn this section we will address the basics only. These basics however, solve at least 85% ofall schedules. The more advanced techniques, such as, backbone planning, backtracking,fish grate planning, etc., are left out.In section “Scheduling/before you start/Column 10” we mention the task predecessor.It is here that the logics are defined.

A few definitions

Logics: the imposed dependencies between tasks. The logics state that task i precedes task j: task j starts as soon as task i is

finished. Or that task k and l start at the same time. Or that task u and v have to finish on the same date. Or that task y starts 5 days before the end of task x. Etc.

Predecessor of task t: any task that has to be finished before t can start.Put otherwise: t can start as soon as all its predecessors are finished.

Task reference: when declaring the predecessors of any given task, we referencethese predecessors by their row number.The row number is found in the leftmost column of the sheet.Once used in the predecessor field, the row numbers will modify automatically should

18

tasks be dragged to other places.

Types of dependencies

FS: finish to startThe task can start as soon as the predecessor is finished

FF: finish to finishThe task and the predecessor finish together

SS: start to startThe task and the predecessor start together.

Leads and lags

For every type of dependency, we can add a lead or a lag expressed in working days. lead of n days: the dependency is shifted n days ahead. lag of m days: the dependency is shifted m days later.

Examples

We consider that we are looking at task in row number 20.In the column “Predecessors”, we find this:

5: task in row number 5 is predecessor of 20 in FS (FS is default) 5 FF: task in row 5 is predecessor of 20 in FF 5SS: task in row 5 is predecessor of 20 in SS 5FS+2 days: task 5 is predecessor of 20 in FS, but 20 starts 2 days after 5 is

finished (lag) 5FS-­ 2 days: task 5 is predecessor of 20 in FS, but 20 starts 2 days before 5

is finished (lead) 5;;6: both tasks 5 and 6 are predecessors of 20. 5FS+2 days;; 6FF-­2 days: this is a bit complicated.

20 will start 2 days after the end of 5 or finish 2 days before the end of 6,whichever leads to the latest finish date for 20.

Predecessors are separated by a “;;”

Good practice

Follow these simple rules Use the FS type as much as possible Avoid spaghetti logics;; avoid linking to many tasks to one given task Use simple sequences as much as possible: a sequence is a series of tasks linked in

FS. Avoid setting fixed dates: this will block your schedule and cause tremendous

problems later. You can link summaries

19

The showing

This video shows some common examples for defining the logics.

Advanced topics

Milestone blocks

Problem

Planners often use many milestones in their schedules. Most of the time these milestonesare spread all over the schedule. This makes it difficult to find a specific milestone quickly,especially when the schedule is large and more especially in the midst of a stressed projectmeeting.There is a simple mechanism to create a tidy and efficient milestone overview in theschedule: create a milestones block.

Solution

The milestone block is an easy planning trick, which when used adequately, will prove to be avery powerful tool. The concept is that the physical planning is constructed with a block ofmilestones located above the planning body and just under the planning title row. Themilestone block is organized in two sub-­blocks: driving milestones and buoys. The planningbody is tree structured and contains any number of tasks.

20

Milestones Block

Driving Milestones

Driving milestones are externally imposed constraints, emphasis on externally.Typically:

obliged starts, mandatory events.

These driving milestones are grouped together;; they form the only group of tasks where dateconstraints are allowed.The driving milestones are SS linked to one and only one task (or summary) of the planningbody.As the planning body is constructed in pure logic, managing changes becomes very easy:simply relocate the driving milestone and the depending parts of the planning body will moveaccordingly.

Buoys

Buoys are passive milestones that "pop up" from the planning body (they appear at thesurface, hence "buoys"). They mark important starts and stops within the planning body.They are passive because they are only linked in FF or SS to one predecessor.

Usage

Taken together, driving milestones and buoys, are a powerful tool for:

21

Quick visualization of a synthetic status on the project: all major events are neatlydisplayed in a tidy survey;; no need to dive deep under the surface of the planning bodyfor a desperate search for a specific task.

Immediate visualization of the consequences of moving a driving milestone: when adriving milestone is moved, we will immediately see that the affected buoys movealso.

If these buoys have been equipped with base lines and deadlines as shownhereunder, then all consequence are made visible in a snap.

TipWhen using driving milestones and buoys, keep it simple.

The showing

This video shows how to create and handle a milestone block

Packages

Problem

It often happens that we need to insert a list of tasks in a planning, and: they must be grouped in one summary, we don't know the sequence of execution, or there is no preferred sequence, and we want to track the individual tasks

As we go along, we hesitate to insert this -­ eventually long -­ list of tasks, because we fear thatit will overload the planning and it's updating.There is very a simple solution to this problem.

Solution

Arrange a list of tasks in a package.

22

A package is a summary grouping the tasks. The package is given a window of time for execution. All tasks must be completed at the end of this time window. We don’t care about the sequence. We track the package on the task level. We link the package to it’s predecessor(s) and successor(s)

Method 1

Tasks have equal length and weight factors.In this case we create a summary containing a series of tasks of equal length (duration). Thispackage in it's whole is linked to predecessors and successors, by linking the summary task.We can assign weight factors to the individual tasks (Weight Factors are explained in anotherarticle), so that their respective weighted contribution to the overall progress will be taken intoaccount.The tasks will be progressed individually. We will monitor the progress of the package.We don't care about the sequence into which the tasks are executed.

Method 2

Tasks have unequal lengthIn this case we create a summary containing a series of tasks of unequal length (duration).We add a start milestone and a finish milestone separated by a distance equal to the durationallocated to the package. All tasks are linked in SS to the start milestone. This package in it'swhole is linked to predecessors and successors, by linking the summary task.We need not assign weight factors to these tasks as their individual durations will be taken asweight factor.The tasks will be progressed individually. We will monitor the progress of the package.We don't care about the sequence into which the tasks are executed.

23

Preferred method

In our experience method 1 is the preferred method, because Easiest Fastest Produces no perturbed S-­curve images

More on the absence of a sequence

One can wonder, “why is there no sequence ?”.The answer is: “Because it’s not relevant”. (in the case of a package)Consider this:

Contractor (external or internal) has to execute a set of related tasks that togetherform a package.

He must respect a given window of time. Say the set of task are numbered 1,2,3…,N He may start with task 6, then jump to task 3, then to 9, then come back to 6, then

jump again to j, to k, to whatever. There is no pre-­defined or even pre-­definable sequence. It is his productivity that is at stake, he knows best how to proceed, we don’t But still, we track the progress of the package by tracking the progress of the

individual tasks.

Examples

Here are a some cases where the use of packages can be considered: Issuing a document package: related set of drawings, related set of diagrams, bills of

24

material, etc. The installation of a set similar and related items: a series of pumps, a set of iso's (in

piping), a set of office equipments, etc. Everytime we consider a set of related, similar tasks, the sequence of which is not

relevant for the planning, but whereby the monitoring of the progress of the whole setis relevant.

Remember

Package → set of related, similar tasks Sequence not relevant Tracking the progress of the individual task is requested. We link packages “in globo” to predecessors and successors

Final remark

I noticed that packages can be used more often than one would originally think.This techniques simplifies the planning construction and speeds it up considerably.Hence: More (tasks in the planning), still less work and easier done.

The showing

This video shows how to create and handle packages

Audit

Before a project schedule goes live, we want to audit it.These are the items to be checked:

Is the project tree free of hybrids?The automatically generated diagnostics helps us with the listing of the wbs codes ofdetected hybrids.

Have we reduced the usage of fixed dates to minimum? Have we avoided spaghetti logics, can we simplify the logics? Are we confident with the tasks durations? Are we sure that every task’s progress can be observed?

25

Have we flagged tasks to “-­p” when we cannot monitor their progress? Is the task inventory complete? Have we assigned a responsible to every task?

Once all checks have been performed and problems corrected, the schedule can be clearedfor going live.

26

Secion 2: Monitoring Progress

A few disturbing words.There is a lot to say about the mechanics of project monitoring.In this tutorial, because of its practical nature, we will not deal with the theoretical foundations,but focus on the practical aspects.We do however provide ample references that provide such theoretical foundations.

But before we continue, it is necessary to clarify a few points.

1. There is no need to set up a schedule if it is not for monitoring the progress of theproject.

2. Monitoring a project’s progress and reporting it is a delicate matter. It interacts withsome of the most profound and basic psychological behaviour of people: the ability toface reality, the acceptance of being confronted with the truth about a situation.

The first point is less trivial than one would think on first sight. Monitoring a project’s progressin a rational, analytical manner, is not a widely practiced activity.It is a strange situation, because setting up a schedule, and then avoiding to monitorprogress, is like planning a trip, and, while en route, refusing to look at the road map, one’sposition, one’s fuel consumption, etc., and to hope to arrive on time at the planned location.No one does that, except when managing a project.So, our attitude is: we plan to able to monitor. And otherwise, no need to plan, go onadventure.

The second point addresses one of the major causes of many disasters. In all kinds ofhuman activities.Countless are the politicians, ceo’s, team leaders, and … project managers, that have driventheir organisations into the wall. Because of one common reason: avoiding, refusing to lookat reality, while choosing to believe fairy tales, rosy dreams, the good sayers, shamans,gurus, etc.Project monitoring as we propose to practice it, namely by looking at the project dynamics, isnot for the “negationist”, those refusing to look at the real picture.

So, at this point, if you don’t feel comfortable with the idea of having to look reality into theeyes, maybe it is better to discontinue this reading. And maybe also to quit projectmanagement altogether: it will avoid lots of frustrations, and who knows, even accidents.

To make things even more clear: there is nothing wrong with experiencing anxiety when truthis about to come out. But the difference between the fit and not so fit, lays in the capacity toovercome this anxiety.

27

What is monitoring about

The past

Monitoring the progress of a project is often tedious and complicated. Lots of informationmust be collected and processed, and in the end a comprehensive report must be issued.Too often, reporting time is experienced as a recurrent nightmare.No anymore for those who have chosen to use our system.

The present

Today, all you will have to do is to enter the % complete values per tasks in the Gantt sheet.This is done like shown in this video.The data capture, the reporting, the archiving of the progress histories, the report distribution,everything is done automatically, without the need of the slightest intervention.

Percentage complete

Consider a project to be a production process, a special case of it to be precise. Instead ofproducing “things” like cars, pieces of furniture, appliances, PC’s, whatever, in a repetitiveway, a project produces outputs of many kinds, and most of the time as single, one-­timeitems. These outputs can be anything:

concrete structures steel structures lines of code articles scripts movies equipment roads …

Any output of a project must have a physical existence, it must be observable as a physicalobject.In monitoring the progress of a project, we monitor the outputs produced by the project. Wehave taken the habit to express the progress in term of percentage of physical completion (%complete). This is a value that we allocate to any tasks at any moment.These values range from 0 (nothing done) to 100 (task is completed).

Aggregated progress

We do enter % complete values on a task level.But what we are really interested in, is the progress of sets of tasks, collections of tasks.Sets of tasks define what are called processes, these are parts of the complete project, thaton their own need to be monitored.Indeed, we do not limit the monitoring to the complete project only, but extend the monitoringto every process (defined by a set of tasks) of interest.By monitoring parts of the project on their own, we will be able to detect parts that start to goastray at an early stage of the project. This will allow us to correct the course of smaller parts

28

of the project in time, so avoiding the complete project to capsize.The project processes are most often defined by the “summaries” of the project tree. So, yesindeed, we monitor the progress of every summary of the project Gantt on its own.Another way of defining processes is by allocating “attributes” to the tasks. Attributes valuescan be:

a contractor area unit phase etc.

How aggregated progresses are computed from the tasks progresses is explained in detail inthis document. We advise to read it later, once you will be familiar with the basics, and havegathered some experience with practical project monitoring.

Aggregated progresses are also expressed as % values. They also range from 0 (nothingdone) to 100 (completed)

Remember: you do not have to be concerned by this: it is all done automatically.

Two defining parameters

At any given time, when we will be assessing the health of a project, or any part (anyprocess) of it, we will be looking at only two parameters:

aggregated progress aggregated progress rate

We look at aggregated values because we are not interested in tasks, but rather in sets oftasks, the processes.We remember that we are dealing with a kind of production process, so we are interested inthe instantaneous value of the achieved output of any process: this is the aggregatedprogress value of that process.We are also interested in the instantaneous, eventually the averaged, value of the productionrate, the progress rate. That is the speed of the progress expressed in % per day.

This whole thing is so simple, it almost is trivial. Still I will illustrate the point with a fewexamples.

Assume we are looking at process P, i.e. a given and meaningful collection of tasks.1. Assume the achieved progress in 25% (complete). The balance to complete is thus

75%.2. Assume the average progress rate is 0.5% per day.3. This means that we need another 150 days (75/0.5) to complete, assuming that we

keep the progress rate steady at 0.5% per day.4. We can now easily assess our position

a. say we have 155 working days left till scheduled end: we are ok.

29

b. say we have only 125 days left till scheduled end: we are in trouble (for thatprocess)

That is all there is to it.Notice that we are looking at one process (P) and that the same procedure goes for allprocesses, including of course the complete project.

We have been combining statics (aggregated progress) and dynamics (averaged progressrate) to create a complete picture of the situation of all processes.We can thus easily detect those processes that may be in trouble and correct their course ina timely way.Again, generating this information is automatic, the analysis and decision making, is ofcourse left to the project manager.

How do we perform the monitoring

Metrics

As said before, the only thing a project manager has to do, is to enter % values on the tasksin the Gantt sheet (video).The question is now, how do we proceed to set realistic % values. This is about the metrics.

The most commonly used metrics for expressing the actual progress are the following:

Binary metric

The task progress only takes two values: 0% or 100%.Binary metric will be used for short tasks.A task is short when it's duration is small compared to the progress measurement intervals.If this interval is f.i. 1 week, then a task of a few days is short.

Fixed steps

A series of fixed steps are defined and progress is registered accordingly.Example: 0-­10-­25-­50-­75-­95-­100%10% stands for: execution of task has just begun.95% stands for: execution is finished but, a small detail is still pending.Setting a task at 95% will keep it under attention.Setting it at 100% will cause it's disappearance from the radar screen.A variation can be: 0-­1-­25-­50-­75-­85-­99-­100.Fixed step metrics are widely used.See more below.

Calculated progress

Some activities allow for an easy calculation of progress.Eg. handling enumerable items:

30

pct=100 (Installed items)/(Total to install) pct=100 (Issued documents)/(Total to issue) pct =100 (Constructed length)/(Total length to construct)

The piping and cabling contractors traditionally have sophisticated progress monitoringsystems. The suit our needs to capture progress perfectly.

Bear in mind that in such case, there must be congruence between elements in our planningand the corresponding elements of the contractors planning.

Hint

A strange twist in peoples brains makes that when questioned about the percentagecomplete, most of the people can estimate this value without difficulty till 50% is reached.After that, most seem to have difficulties in estimating the fraction of work done.

Think of asking the complementary question: how much work is left?You will see that once over the 50% barrier, this question is easily answered.

Special case: tracking documents

There are some special cases, e.g. the tracking of document production.In such case, we must define fixed and agreed meanings for every % step.Example:

just started 5%draft ready 25%draft reviewed and commented 35%comments processed 50%issued for approval 75%comments processed 85%issued final approval 95%approved and filed 100%

Or any similar system.

One might be a bit overwhelmed by these metrics systems. Bear in mind, that most of thetime, we use the same, or similar and familiar systems. So, once the we have had our firsthands-­on experience, we will see that the next cases will be extremely simple to implement.

More details about the fine mechanics of the metrics can be found in this document.

Driving the reporting process

We still need to tell the system what we want about the reporting: we need to tell whichproject must be tracked and at what time intervals.This is done through a special sheet, the “configuration sheet”, aka “job list”.This sheet is shared and looks like this:

31

The columns are:

Proj: The project name Freq: The reporting frequency TheHour: The hour of issuing the report (local time) E-­maill to: Reference Active: Project is fit for reporting when checked Link to Report:The link to the report's web page

The values for “Freq” are: D: daily M1: every last day of the month M2: every 15th and every last day of the month Monday: every Monday of the week etc

The DPC engine fetches the data from the configuration sheet (using the Smartsheet API)and using these, produces the progress reports for every project with the set frequency.The reports are deployed as web pages that can be accessed using the link as set in theconfiguration sheet.The more advanced reports are deployed as CDF documents when requested.The whole process is transparent for the user: there is no action needed whatsoever.

S-‐curves: the kernel of the progress reports

The basics

Remember: we monitor (aka track) processes defined as sets of tasks.The most commonly used sets are those defined by the project tree, i.e. by the summarytasks of the Gantt sheet.So expect one S-­curve per summary task.

A basic progress report looks like this:

32

The left part is called the Daisy tree. It is the display of the project tree as a daisy chart.

The Daisy tree acts as a navigator for the project's S-­curves: on the web page, clicking onan item will display its S-­curve. You can already have a “hands-­on” experience here. The Daisy tree can be zoomed and panned .1

Every segment of the Daisy tree corresponds to a specific part of the project. The central disk corresponds to the complete project. The items of the first ring correspond to the first level items of the project tree, the

second ring items to the second level of the project tree, etc.

1 Best with these browsers: Chrome -­ Safari -­ FireFox -­ IE10 supporting SVG formats

33

The color of any item corresponds to the "color" of the present project status: seebelow.

Important remark.This style of reporting not only displays Information on the global project, but also, and moreimportantly, on the all sub-­sections of the projects.This allows for early detection of parts of the project that are developing problems (generallythey are then "red").This in turn, acts as an early warning allowing for timely correction, so that problempropagation is averted.

This is an S-­curve. It has been displayed in the right part of the report by clicking on an itemin the Daisy tree.

1 Horizontal axis days elapsed

2 Vertical axis % complete value (0-­100)

3 Blue line the progress line as scheduled

4 Gray area the activity profile as scheduled

5 Black dotted line the observed progress line

6 Fat black dot the latest status point: where we stand now

7 Green dot a marker at 50% progress and 40% duration

8 Red dot a marker at 50% progress and 60% duration

9 Black dot midpoint, 50%, 50% duration

34

The color areas are related to different degrees of safety.Safety in terms of "safe position" regarding the likelihood of finishing in time.

Blue area: ahead of schedule Green area: safest area Yellow area: less safe area Orange area: lesser safe area Red area: problems must be taken care of

By simply looking at the area where the latest status point (6) lays we already know howsafe we are.

The color of the item in the Daisy tree is the same as that of the area of the last statuspoint.

So the color of the Daisy tree item already tells us where the problems are: the "Red"ones and to a lesser extend the "Orange" ones.

Inspect the Daisy tree

Spot the red items: these are the ones having troubles. By clicking on the item, display the corresponding S-­curve.

In this one simple step, we already know where the problems are

Inspect the S-‐curves

Identify the area in which the last point lays

In Blue: ahead of schedule In Green: safe In Yellow: less safe

35

In Orange: should get worried In Red: problems to take care of

Check the trend line

The trend line is the average progress line of the observed track (the black dotted line).Its slope is the actual average progress speed (in % complete per day)Compare this slope to the average slope of the schedule progress line (the blue line) for thesame period.

Rules1. V1 is the progress rate of the scheduled S-­curve2. V2 is the average progress rate of the track3. If V2<V1, then delay will increase4. If V2=V1, then delay will remain stable5. If V2>V1, then delay will decrease

Diverging track

This picture displays a diverging track. We are in the condition V2 < V1 (see above).Delays will continue to increase.If you see this: the process is out of control.

36

Converging track

This picture displays a converging track. We are in the condition V2 > V1 (see above).Delays will decrease.If you see this: control over the process has been regained.

Parallel track

When the process track runs parallel to the scheduled S-­curve, then one can conclude thatthe process is well under control. We are in the condition V2 = V1 (see above).The delay canbe recovered during the end phase (see in diagram);; eventually a limited acceleration might

be needed.

37

Action!

S-­curves are not for contemplating but for defining actions.

As soon as we detect problems -­ remember that this goes for any part of the project, anysub-­process as we call it -­ we should consider to act upon the process in order to correct thecourse.

The domains upon which one can act (more or less) are a.o.:resources quantity -­ resources quality -­ information flows -­ logistics -­ worked hours -­ scope

When and only when no room for improvement is left, then rescheduling can beconsidered.

⇏ Never contemplate re-­scheduling every time a delay is incurred: you will only fool yourselfand the whole project organisation.

This "strength of action" matrix can give some guidance

Trend / Color Blue Green Yellow Orange Red

Divergent Monitor Monitor Monitorclosely

Prepare foraction

Act strongly

Parallel Considerslowingdown a little

Do nothing Do nothing Consideracceleratinga little

Act

Convergent Do nothing Do nothing Monitor Monitor Monitorclosely

38

Section 3: Portfolio

BasicsIn the portfolio report we look at a whole bunch of projects at once.This is an example report in style 1

This is an example report in style 2

In style 1, the project trail is displayed, in style 2 we see the last position and the trend as avector (direction and magnitude).

In style 1, every dotted line is the observed trail of a project.The color of the dots corresponds to the color of the area where it lays (or laid in the past)In style 2, only the last point is shown. Its color is the color of the area where it lays.

All projects, regardless of their duration, fit into the same rectangular area, a dimensionless

39

space.

That is because the horizontal axis corresponds to the "fractional elapsed time": theratio (in %) between elapsed time and project duration.

The vertical axis is the score of the project. The score is a measure for the distancebetween the status point ( (6) above) and the scheduled blue curve. It is alsoexpressed as a fraction.

The more a project point is close to the zero line (score 0), the better. A negative value indicates a deficit and thus a delay. And vice versa. The blue area is where we find the project that are ahead of schedule. The red area is where we find the projects when the finish date is past due. The green area is where every project should converge to.

The portfolio report is issued as a web page. When clicking on a "last point", the fat ones, thecorresponding S-­curves are displayed.

An important feature is the trend line as shown below.

When we see a negative trend line, the situation is degrading and vice versa.The complete portfolio report contains separate views for:

All projects Projects in good shape The red ones Projects that are improving: positive trend line Projects that are deteriorating: negative trend line

Method

40

The portfolio report displays the observed trails of all projects (style 1) or the last point andtrend as a vector (style 2).

When looking at the section "Deteriorating", attention should focus on all red projects.This example shows how easy this is done.

Remember that this report is issued as a web page.Hovering over any dot will display the project's name.Clicking on the last dot in the trail will display the S-­curves of the project.

Both are needed: portfolio and the individual S-­curves.

Only the S-­curve can tell how much of the project is completed, and at what progress rate itis advancing.The portfolio report helps detecting projects that are in trouble, and see how they evolveglobally, without having to dig into the details of every project.

In this case, based on the “Deteriorating set”, one would advise to Check on all the red projects Check also the one marked with (*): potentially dangerous evolutio

41

References

These references are provided for further reading.Some of those reference clarify a number of concepts, some provide the theoreticalfoundations.

S-­curves in detail: construction and usage PDF

Tasks versus Processes PDF

Metrics PDF

Milestone Blocks PDF

Packages PDF

Tree PDF

42

Tutorial videos

When viewing the videos, set you video player to HD and fullscreen for best results.

1 The project tree

2 Hybrid structures

3 Setting the working days

4 Setting the weight factors

5 Setting the logics

6 Milestone blocks

7 Packages

8 Entering % values

9 SMSHAR at work: a recap

10 Daisy Tree

11 S-­curves

12 Analyzing S-­curves