There are two basic approaches to project planning: charts and network diagrams.
Charts means using bar charts often called Gantt charts. These are very useful for visual display. Vertical links can be used to show dependencies between tasks. These can be, however, difficult to read for more complex projects. Microsoft Project is a good computer program to use for Gantt charts. The logic is based on Critical Path analysis. This is a network analysis method recommended when the times comes to begin work on all but smaller projects. They depend however, on knowing the great deal of detail about the project tasks, so generally, these are produced later in the project life cycle.
Very early in the project life cycle the actual date when the project will start cannot be predicted. When the project does receive the approval to start it is highly probably that the project will be replanned bottom-up with much greater detail by a professional planner using critical path network analysis.
Project Breakdown Structure
Most projects contain hundreds of tasks that need to be planned before they are done. Therefore, we need to organise those tasks in some logical way. A typical project has many dimensions: work dimensions (tasks and its practical sequence), organisational dimension (people distribution), cost dimension ( estimated and real costs of all the work) time dimension, and resource dimensions ( detailed view of personal skills and material resources). These dimensions are interrelated together. They can overlay each other and become extremely complex. These seem impossible task to schedule for an average human mind. These are very difficult to prescribe visually and in pictures. Therefore we need to manage its proportions and prepare work Breakdown structure for it to be scheduled later on.
The WBS is a logical and hierarchical tree of all the tasks needed to complete the project. The top of the tree is the project itself. The next layer consists of work packages. Levels below that get more and more detail until the bottom level is reached to show all the smallest day-to-day tasks of project components.
The figure below presents a simple WBS of automobile project.
It can often be used by engineers and designers, when organising their drawings, bills of materials and parts lists into logical patterns.
It is possible to visualise the WPS as a jigsaw puzzle with every piece put into a right place and no piece missing. It is therefore important when designing the WBS to every piece of the puzzle to be introduced within piece missing. It is sometimes difficult to achieve. Checklists can be used for the risks of omissions. Each piece might be given a reference number that is used within logical coding system.
People might find it difficult to produce a Work Breakdown Structures. We will, therefore look at some examples.
Example 1
A company decided to organise national fund-raising week and draw WBS for the project. The WBS consists of four main elements and it is easily seen that each can be allocated to different managers responsible for it. One manager will organise public relations activities (newspaper advertising, television commercials, posters, billboards, commercials).
Second manager will be put in charge of local events (coffee mornings, student rugs, village fetes etc).
Flag days are also organised for street collections by the volunteers. Someone must organise production of cans, labels, find volunteers.
As hundreds of pounds might be collected there is also a scope for fraud, therefore there is a need to appoint someone to look after the money, banking and security.
Example 2
Another example, might be a copper mining project.
The figure above shows some of the work that might be needed. Everything must be provided as part of an enormous project. The level 1 presents of all the Infrastructure needed. Level 2 Expansions and Level 3 presents all the processes broken down.
The WBS must be continued until the lowest level os reached in which the tiniest components of the plan and new town are listed.
Example 3
A wedding project can also be detailed in a Work Breakdown Structure. Organising a wedding might be a trivial venture. In large weddings, there might be large quantities or guests arriving, displays or flowers, music and entertainment, meals and food, drinks etc. The WBS for such projects is not very straight forward.
We can organise the projects by locations or by functions as presented in future below.
In the upper WBS it is assumed that different manager will be responsible for different specialist function. The lower WBS is determined by the location. The recommendation is however, to use upper WBS. This is probably more cost effective.
Coding System
Every task will need to be given its name and description. These should be augmented by a specific code. A code might be a sequence of an alphabetical character and some a mix of these two. A code could be applied to anything within a project, a component, drawing, job, operation and so on. Each item has cost that must be estimated, budgeted, spent, measured, reported, assessed and recovered.
Example 4
For example, a following code elements are used in WBS for telecommunications project:
Project identifier 110-00 in a register of all projects.
Item identifier - for example - Transformer 110-221
Relationship with the project - for example, project 110-000 has transformer 110-221, and it starts with 110.
Operation identifier, for example the task of winding the transformer 110-221 will be coded as 110-221C where the C represent the winding operation.
Identifies for departments, discipline or trade - for example the discipline of designing the transformer 110-221 could be 110-221-153.
Figure below presents a simple guide for determining the coding for WBS.
Further the equipment and machinery can also be coded accordingly.
There are many benefits of the coding system. These benefits increase in time and allow the project to be used consistently. It allows for data efficacy, easy retrieval, easy reach of design information, assembling of components, cost, budgets, time etc.
Example 5
Some codes are used during the design process of a machine as presented in figures below.
Coding allows to prepare estimates tables, man-hours, material costs, allow planning and resource scheduling, forecasting and so on.
Once the coding system is determined it is difficult to reverse, therefore it needs to be designed with care. It is also tempting to be ambitious and try to make numbers include too much information. This might result in codes to contain many characters. We have to remember, however, that people will need to work with these numbers. When codes are too complicated people may not waiting to use them.
Often Project Work Breakdown Structures (WBS) and Organisational Breakdown Structures (OBS) relate to each other. Sometimes project need consideration of business structure and its representatives to carry out the work. In practice, the project manager needs to produce the WBS and OBS separately and do not try to built one diagram. This can become very complex. The OBS codes, however, might be used to determine the cost budgets for the tasks scheduled across departments and functions. Staffing codes might be available for labour costs.
Critical Path Networks
Gantt Charts are not always able to show complex dependencies between tasks. They can become very cluttered and difficult to follow. Critical path methods can be introduced to overcome these difficulties. They simply show all the project tasks in a logical sequence. Generally, it is better to convert network into Gantt chart. The planner can ensure by using them that some tasks will not be started before its needed foundation. Such logical mistakes are relatively easy to make with Gantt Charts. Another great benefit of networks is that they allow critical tasks to be prioritised. The tasks can be ranked according to their criticality. Networks cannot, however be used for resource scheduling and that's when the Gantt Charts can become useful.
There are two main methods identified: "Activity-on-arrow" and "Activity-on-node". Arrow methods are often better for initial planning because they are faster and easier to sketch. These can be particularly useful for brainstorming sessions when logic is being scoped into shape. Then, it is much easier to make them into precedence notation (activity-on-node).
The serious professional planner should be aware of both methods.
The Arrow Logic Diagram
In arrow diagrams, every network is constructed with careful thought to show accurately as possible the logical relationship and interdependence of each task. These are always drawn from left to right. It is important to estimate, however, how long each task will take. The activities and their time are generally written above the arrows and once the diagram is drawn up it is easier to determine the project time.
In the example above there is more than one path through the arrows from project start to finish. There are three possible routes to the final event (6). Numbers have been written above the activity arrows to show their estimated durations. These durations should be always consistent, for example minutes or days. Days or weeks are used most in project plans generally. The earliest possible time for each event and the earliest possible time for project completion at event 6 has been calculated by adding activity duration estimates along the arrows from left to right.
This is always the first step of full-time analysis and known as "forward pass". The earliest time indicated for each event will depend on which path is followed but only the longest path will give the correct result. The earliest possible completion time for event 3 might be 1+2 = 3 if the path through event 1,2 and 3 is taken. Event 3 cannot really be achieved, however, until the end of week 5 and the earliest possible start time for activity 3-6 is the end of week 5. Thus the earliest possible time for any event is found by adding the estimated durations of all preceding activities along the path that produces the greatest time.
The earliest possible time of event 5 is the end of week 6, 3 weeks before the earliest possible time for finishing the project at event 6. Activity 5-6 could be delayed for up to 1 week. This can be indicated on the arrow diagram by writing the latest completion time underneath the event circle. The result is found this time, not by adding from left to right along the arrows but in the opposite way by subtracting the estimated durations of activities from right to left (9-2 = 7 for event 5). This can be repeated throughout the network, writing the latest time underneath the circles. Where more than one path exists, the longest most always be chosen.
The term 'slack' indicates the amount of leeway available for achieving an event (the difference between the earliest possible time to finishing the project). "Float" is the correct word for the leeway for activities (as opposed to slack for events). However, these are used synonymously.
When all the earliest and latest possible times have been added to the diagram, there will be at least one chain of events from the start of the network to the end, where the earliest and latest events are the same indicating zero float. These events are critical to the successful achievement of the whole project. These are called the 'critical path' and are critical to the project.
There are many different ways of showing on arrow networks.
The times written on arrow networks usually refer to the event rather than activities. Project managers need to know the time when each activity will start and finish. Figure A above shows arrow network according to BS4335:1987. These method might not be suited for sketching because large circles are generally required and each event must be written very carefully.
Version B allows rapid sketching. However both A and B do not allow the time analysis easy from the network. Version C can be used to avoid this issue.
Activity on node
Precedence diagrams (activity-on-node) are easier to understand. They allow clearer illustration of activities. They can also show activities that can overlap. As with activity-on-arrow, the activity-on-node networks cannot be used to schedule resources, therefore further conversion to Gantt chart is required.
Above figure shows conversion of an activity in precedence notation. The flow of work is always from left to right. the length of the boxes or arrows have no significance. Every activity is given an unique identification number (ID Code). A single code for each activity is all what is required. These can range from small serial numbers to complex codes. The activities are linked by lines which, unlike in arrow diagrams always have zero duration. Activities can be shown to overlap rather than follow each other in a sequence.
The earliest estimated start and finish times for these activities are found by adding up the estimated durations from the left to right are written in the activity boxes. We can use minutes, days or weeks.
The earliest overall project duration possible can be calculated by adding activity duration estimates from left to right in node network diagram.
In the figure above, there are few possible paths through the network and the path with the longest duration will determine the earliest possible finish time for this project. The earliest possible time for activity 4 might seem 0+1+2 = 3, the end of week 3, however the activity 4 cannot start at the end of week 5 because it considers the longest path through activities 1-5.
Let's present it on an example.
Example 6
A company A is conducting a development project to design and built chair and desk. The project has been re-planned from Gantt Chart to network analysis.
The company was able to produce the complete list of project tasks using brainstorming sessions.
These activities are presented in figure below.
The activity on arrow diagram was as below.
And the Activity-On-Node diagram is presented below.
Time analysis has been completed on both diagrams and the data is identical. Below table tabulate the data.
All times are given here as the days numbers. This is typical for projects when the start dates are not known. As soon as authority is given for the project to start, the start time of every activity can be converted to the appropriate calendar date and then all subsequent start and finish times can be given as calendar dates. Only working days would be valid.
There are also more complex network notations. For example, there might be overlapping activities in arrow networks. In belie figure three activities are shown: design engineering, drawing and procurement of components/materials.
Drawing (a) shows the network fragment as it was originally drawn, with the three activities following each other. That it tells us that each activity cannot begin until its immediate predecessor has been completed. The combined duration is therefore 28 weeks. But there is a flaw in the logic. The design have not have to be finished before drawings can start. These activities can be allowed to overlap. Similarly, some of the long-lead purchase items can be ordered in advance as soon as the designers can specify them. In figure (b) an attempt has been made to indicate the permissible overlap of activities by re-drawing the network. The start and finish have now been relaxed by inserting dummies, each has been given a duration value. Drawings can start 2 weeks after the start of engineerings but cannot be finished after 3 weeks after the completion of engineering. Purchasing can begin 3 weeks after the drawing. Some parts cannot be ordered, however, until the part list is issued along with the drawing and delivery of late purchase items is not expected until four weeks until the completion of drawing.
This means that timescale for this small part has been reduced from 28 days to 17 weeks. The product can start to come off the production 11 weeks earlier. The drawing (b) is called a "ladder network convention". The drawing could start 2 weeks after week 0 even if no engineering had been carried out and procurement could start at week 5, whatever the state of engineering or drawing. In figure (c) the same sequence of activities has been depicted but, by splitting engineering into two phases and doing something similar with drawing and the true relationships and constraints are more clearly defined. But there is a mistake with the start restriction imposed on purchasing. It is not dependent on completion of engineering but only upon Phase 1 of the drawing. The real picture is presented in graph (d) with all the dummies correctly placed.
Activity links
Overlapping activities and other complex issues are best planned in precedence notation. Different types of links can be used. Below figure show four different types of precedence links. Figure (a) below shows normal start-finish relationship with time value on the link. The time would force a delay between finish of the activity and the start of the next. This can be useful for example when waiting for cure or dry of equipment.
Whenever significant number of activities have to be linked independently to several following activities the use of dummies can present more clear diagram. This makes the precedence system a disadvantage. In the figure below (a) and (b) it can be visible.
Dummies can be added artificially to the precedence network (figure (c) below). The two dummies have improved the clarity without changing the actual logic.
Let's put it into practice
Let's look at some methods and steps to implement project network successfully.
Step 1. Develop the Logic
Network diagrams should always be fully utilised. Brainstorming is probably the best approach as it allows for different perspectives. The number of people who will need to be involved will depend. It is always a good idea to have one person who can speak to each department involved. The participants should have relevant authorities to ensure commitment to the plan.
The network should be drawn rapidly, as large as possible and in full view to those who contribute. Everyone should see the whole logic pattern. Sometimes a large piece of paper rolled out in the middle of a table is still the best approach.
Arrow notations are best suited for this brainstorming sessions, because it can be sketched very quickly. It is always easy to convert to precedence later, even by using Post-it notes for each activity. The initial sketch can be drawn more tidy with a ruler and template for drawing event circles. If it is produced with a little care it can become unreadable and many corrections might be missed.
The person who sketch the plan needs to be competent to do so. The person need to be skilled in network analysis. This takes time to practice and it cannot be learned in few hours. Getting the logic right can be challenging. The person should be also allowed to control the brainstorming session. The logic needs to develop around the right lines. Questions to check the network already built will need to be asked time to time in order to prove logic and avoid errors.
These questions can be, for example:
- We have shown work on site immediately, but should we deliver some machinery first, deliver materials, get licence, approval etc?
- Should we check the design or drawing before it can be released or issued?
- Is customer approval needed before work can start on this activity?
- Would the tool need some protective work before erection?
- This method failed us on previous project, can we come up with something better?
- Does the start of this activity really depends on all of these activities?
The diagram that emerges from the initial meeting may have may changes that need a careful thought. The network analysis must never be too lazy. The planner must always ask whether every activity really depends on all the activities entering it. It is possible to make logical errors unwittingly. Insertion of dummies can correct these mistakes in arrow diagrams.
Step 2. Consider the level of detail
The size and duration of the project will determine the level of detail necessary.
Short durations
The duration units chosen as well as the amount of knowledge at the time of building the network.
One tip, would be to avoid showing jobs as separate if they are only a fraction of an expected overall timescale, especially if they do not require resources. These can be added to the network as part of other activities. For example of a project needs many drawings to be completed, let's say 15. Each drawing will need to be checked. Instead of writing each drawing and check as separate activity (30 in total) we can put two: complete drawings and check drawings.
For projects which complete in several weeks it is a good idea to plan the project in unit of days or half days. For project lasting few years, we might use weeks as planning units with very few activities lasting less than a week duration.
There are exceptions, of course, and some activities with very short durations might be important. These could be obtaining authorisation or approval, for example.
Responsibility
A network path should also be interrupted when tasks or activities move from one department to another. Often, that's where the immediate responsibility changes.
In arrow diagrams generally a new event is being created whenever responsibility progress from one manager, department or organisation to another. In precedence networks this can be done by saying that if single activity covers consecutive tasks by two different managers or departments, it is split into two separate tasks or activities. One task, therefore has one manager responsible for it. No network activity should be so large that it cannot be assigned for principal control of one department or manager. Additionally, an activity that is long compared to the project timescale, should in reality be broken into more activities. It allows frequent progress checkpoints.
A network that is detailed appropriately will enable project tasks to be easily identified, planned and progressed. For example, authorisations, approvals, financial authorisations, authority consent or applications, a start of finish of design for any manufacturing of assembly (if longer than 2-3 weeks we might want to break into shorter activities), release of drawings to production, start of purchasing activity for each raw material, any long delivery times materials should have information released in advance, issues of invitation to tenders, receipt and analysis supplier bids and quotations, the issue of purchase order to supplier, delivery of materials to the point of usage, transit time might be shown as separate consecutive activity, the start and completion of manufacturing stages, the start and finish of subcontracts, handover events of completed work packages, maintenance, manuals release, etc.
Cost
A network should also consider the costs. Generally it is possible to assign a cost to activity when we buy materials. Then, purchase order value can be assigned to activity. This makes possible the preparation of reports. Cost schedules can be derived when the invoice payment is due. This will indicate the cash flow requirements rather than purchase commitments.
Smaller networks
Networks with too many activities are prone to errors and difficult to proceed. Therefore, we need to choose the level of detail carefully. For high level management controls some companies use 100 to 150 tasks. Sometimes networks are backed up by more detailed networks produced by relevant department for the project.
Step 3: Interfaces
At times, activities in one network can have a logical link with another network. This happens when generally projects have been divided into smaller sub-projects and sub-networks. This can be a result of Work Breakdown Structure or Organisational Breakdown Structure. Sometimes interfaces appear with more than one network and have to be merged.
Each interface in the precedence logic diagram should be shown by a box with double-ruled line as per picture below.
In arrow diagrams we can use double circle.
Step 4. Milestones
We need to produce a schedule from which progress can be managed. Events of activities with significant importance are called "milestones". It is achieved when relevant milestone activity is finished. Some modern computer programs allow to filter only milestones that helps the project assessment and reporting to higher management.
Step 5. Estimating
As soon as network is drawn we need to estimate the duration of activities. From the start of each task to its finish.
We need to decide what units should be used. Days are often used in computer software packages with weekends as non-working days. For larger projects it might be more convenient to use weeks as the units. However, even with projects lasting several years, using days as units should not be a problem. Half-days are also very useful as units.
Once we determined the units, the same units should be used consistently. This will allow to convert later into calendar days.
We should also draw a procedure for estimating as soon as possible that will allow all managers or department to work to the same rules. Another solution would be to sketch the network from one department to another with each manager or department adding estimates for their responsible activities or tasks. We should also allow as much impartiality as possible, so estimates are not influenced by other demands rather than by the true nature of the job and the proper time required to complete it.
Any assumption that overtime will be used should be avoided. We should never add overtime into the network view. These should be treated as reserve resource when shortening of activity duration is required.
We must also look at the resource availabilities and constraints. These can involve shortages of people, materials or money when estimating activity durations. For example, when single individual needs to consider several network activities. One person cannot perform two activities at the same time. Sometimes we may need to add dummies or links to the network to indicate this resource constraint and prevent those being planned as two simultaneous tasks. It can be rather difficult to determine this is a complex networks with sub-networks and interfaces. Therefore time analysis needs to be completed. Assumptions should never be made. There is a simple solution to resource constraints, however. We need to ignore them. This is because the purpose of drawing the network is to establish the logic. We than follow with the time analysis to establish the amount of float available. Subsequent resource scheduling would be defined in a separate procedure. We, of course, must use common sense here. The planners should ask managers to determine the number of people or hours needed for a task and write the durations in the network.
Step 6. Review total timescale
The new network diagram should always be checked to ensure it presents the most efficient and practicable way. Generally the first arrow diagrams will predict timescales which are way too late. Often, to satisfy delivery time the planner is put under pressure to come up with a shorter plan.
Here, we should never consider to cut estimates. Unless a good reason is presented that tasks can be completed earlier.
When projects are approved and project managers appointed the first task for the project manager would be to familiarise themselves with the project design, the business plan and the intended project strategy. Sometimes study of the organisation structure needs to be performed to define and know the key persons and resource availability. The project manager would generally need to developed more detailed plan to control the project. Planning meeting would need to be organised with relevant senior staff and, using brainstorming methods, a draft network to be produced. Tasks would need to be established for each person responsible, including subcontractors in greater detail. Tasks will need to be reviewed and confirmed by everyone, including the durations. If the total timescale is not reaching a target deadline, something needs to be done about it. This is very common in project planning. Several methods have been developed to deal with this and reduce overall project durations. We might:
Shorten the estimates until the plan is reduced to accepted timescale. But this can lead to a disaster and should be exploited no further. It relies on wishful thinking and unfolded optimism.
Shorten the estimates for some of all of the tasks on the critical path by injecting more resources, or by changing the proposed working methods. This is called "crashing" in project management or "cost/time optimisation".
Re-examine the network logic and ask whether every finish-start constraint is strictly necessary. Perhaps there are some tasks that could be started earlier. It might be possible to overlap tasks or run them concurrently. This is called "fast-tracking".
Combine "crashing" and "fast-tracking".
Crashing
For example, we might apply some crash actions but attention must be focused on critical activities, because there is no point in spending extra effort and money trying to speed up tasks that has plenty of float and thus, these are not critical. We might choose to reduce time needed for some tasks by employment temporary staff, but there might be additional cost accumulated as a result. For some tasks, work might need to start earlier and finish later (overtime), or night shifts might be considered. This might also bring additional costs. Highly skilled tasks might be replaced by subcontracting, but additional hight costs might be paid. Working weekends, would be also an option.
If we attempt to crash complex networks with a great number of critical tasks, we should also consider cost per day saved. This can vary also depend on the cost of the tasks. We might not need to crash all the activities to achieve required earlier completion. Therefore, it would be wise to choose first those tasks that cost least money per day's time saved.
When one or more critical tasks are crushed, it could be that the original critical path is no longer the longest path through the network and other tasks become critical.
The optimum cost-to-project time saved is achieved by always choosing first the activities that are cheapest. The process will produce the network in which the project can be completed for the shortest possible time and least additional costs. We must remember, however, that every task duration in a network diagram is only an estimate based on the best judgement that people can make. When work begins to start the costs begin to change and problems occur.
Fast-tracking
Even using crashing methods, it might still not be possible to meet the project deadline. Network logic should always be questioned to ensure that constraints shows are necessary. It is usually possible to find tasks in the network that can be allowed to overlap other tasks or even run concurrently with them. We might be able to shave some more days or even weeks off.
Most networks use simple finish-start relationships but a good planner knows the availability of complex precedence constraints and use them in those cases when the project might benefit from the potential time saving. One benefit of fast-tracking by overlapping tasks or allowing jobs to run concurrently is that it adds no cost to a project. It can, however, add risk. When one task starts before another is completely finished, there is always a risk that some work will have to be scrapped and done again.
Step 7. From Back-to-Left
An experienced network planner can control the brainstorming meeting to extract a plan even from a seemingly hopeless situation. It might be better, at times, to draw the network from right to left.
For example a project might be progressing already for several months. Planning might of been neglected or there might be a poor coordination between departments in an organisation. The timescales and durations might drift out of control. Sometimes no one can say what the current state of a project is. We might therefore, re-plan by looking at the end event. We draw this event at the right side of the network. Than, another activity is draw up just before the final event. Questioning of the previous events continues going into a greater detail. This is, until the diagram fully terminates at the left hand side. Sometimes in a number of staggering start events. These would be further evaluated for the current state. These might also be activities not previously thought of.
We now have a detailed plan from which we could progress the remaining work. Order can be achieved from chaos.
Summary
Project networks can contain thousands of activities, however, the elements of a network notation are simple. We must practice and gain competence in using the notations. Critical path techniques can be applied very effectively.
We must remember, however, that planning is not a precise technique and the networks should be kept as simple as possible. The benefits of drawing networks are generally worthwhile, even if no duration estimates are made, no time analysis take place and the network is not used to control the processes. The networking encounter logical thinking and planning.
Brainstorming meetings can be rather beneficial. Network planning should not be done in isolation by project manager. He or she should cooperate with each manager and department. This applies not only to planning, but also control and monitoring of the project. Encouragement must come from the top - company senior management. The battle might more likely be won if all departments accepted the plan and schedule.
Comments