One of the less well known benefits of location-based scheduling is that the quality of schedules is easy to evaluate. A well-planned schedule has certain characteristics that can be directly measured. During my year as a visiting scholar at Stanford University in 2005-2006 we developed together with master's students a list of attributes of a good schedule:

1) A good schedule should maximize productivity
~ continuous use of resources

~ durations based on accurate scope, quantities, resources, and productivity data

~ optimal crews

~ completely finishing locations before moving to the next location

~ aligned production rates

~ only one crew working in a location at the same time

~ good access and laydown space

~ right information and materials available

 

2) A good schedule should minimize risk
~ minimize the number of mobilizations

~ minimize the number of resources

~ buffers against uncertainty

 

3) A good schedule should be feasible
~ correct logic

~ resource constraints

~ accurate model of subcontractor behavior

~ coordinated with other subcontractors

 

Most of these issues can be directly observed from a flowline schedule. Continuous use of resources is the default planning option and can be seen as continuous lines in flowline. Durations require information about associated scope, quantities, and assumed productivity rates which can be checked against industry averages (such as RS Means in the US). Alignment of production rates can be checked by looking at empty spaces in the schedule - if a location is empty for a long period of time, the schedule is not well aligned. Overlapping lines in the schedule indicate multiple crews working in the same location at the same time. A linked 5D model can be used to evaluate access and laydown space. Location-based risk analysis tools show risky areas where more buffers are required. If the same risk assumptions are used for multiple schedules, the schedules can be compared quantitatively in terms of risk.

 

Vico has developed an internal scheduling assignment where all the participants solve the same scheduling problem with the same constraints and risk factors. The spread of results has been surprising. Using exactly the same assumptions for available crew sizes and exactly the same logic, the best submitted schedules have achieved three months faster schedules with lower risk than other submitted schedules which still follow the same constraints. This shows the power of schedule optimization but also that optimizing schedules is a skill which requires a different way of thinking compared to traditional scheduling where if the same durations and logic were used, everyone would arrive at the same result. If durations are allowed to vary as a function of crew size, and planners are free to optimize production rates based on resource availability and the overall sequence of locations within physical constraints, big differences in total durations emerge. It would be interesting to compare how well a CPM scheduler would manage using traditional CPM tools.

 

If someone wants to try this scheduling challenge, please contact me - I will send the files and instructions. I will also score the assignments and provide feedback as long as the total number of submitted assignments stays in reasonable limits.

 

If, like my students, you like to cram before a test, review these archived webinars:

BIM 401: Model-Based Scheduling

Scheduling Strategies in a Hard Bid

 

We also have a demonstration of Vico Control from Klorman Construction and a list for FAQs for Vico Control.  And here is a complete description of the Construction Scheduling Challenge.

 

Good luck with the challenge!

The goals of Lean Construction are to avoid waste, improve productivity and decrease variability.  (Review Part 1 of this post, Scheduling with Lean Principles.) 

 

The main tool to achieve this goal, the Last Planner System^TM has been developed by the Lean Construction Institute and has become almost synonymous with Lean Construction in the US. The Last Planner System is based on techniques of late planning and reliable promising.  Pre-planned master schedules are limited to milestones, and give constraints to phase schedules which are developed by the team, who will do the work.  Phase scheduling is pull scheduling - the schedule is developed by working backward from  a target completion date, which causes the tasks to be defined and sequenced so that their completion releases work.  The best known part of Last Planner is weekly or daily planning, where commitments are tracked and measured by calculating the percentage of plan completed (PPC) for each week.  For each failed weekly or daily assignment, a root cause analysis is done to prevent the problem from happening again.  This social learning process improves the reliability of promises.  Lean Construction tries to make all planning collaborative, and dismisses the traditional Command & Control planning approaches, where a centralized planning division creates plans and tries to make the field to follow them.

 

Location-based management systems try to achieve the same goals. Instead of focusing on the assignment level, the tools focus on planning continuous flow, and forecasting problems before they happen to enable timely reaction. The goal is to achieve an assembly line of continuous production on the construction site.  Location-based management systems have a long history, and have been dismissed by Lean Construction communities as Command & Control tools.  My PhD research shows that the Command & Control approach does not work, and the location-based management tools should be used in a collaborative way, incorporating the knowledge of the subcontractors into the planning process.

 

The Last Planner System lacks explicit tools for pre-planning, phase scheduling and look-ahead planning.  Location-based management system has tools to pre-plan the assembly line of production, adjust it during phase scheduling, forecast future problems to help in the look-ahead process.  However, it lacks the critically important tools of execution - transforming the plan into weekly or daily commitments and tracking those commitments.  My PhD research suggests that the PPC is strongly correlated to plan reliability on master schedule and phase schedule levels.  It seems that there is a big synergy between the two methods.

 

Location-based management gives a systematic process and tools for Last Planners to improve their PPC while maintaining the link to the master and phase schedules. During planning phase, continuous workflow can be planned to get a good understanding of required capacity.  This can be implemented in a phase scheduling meeting using a collaborative process.  An objective, mathematical procedure is used to forecast upcoming production problems based on progress so far, and the knowledge of available capacity in the upcoming weeks for each subcontractor.  This forecast is adjusted by the use of control actions to prevent problems.  (To see examples of this 4D coordination process, please review the archived webinar, Coordination Strategies for a Hard Bid.)  The adjusted forecast can be used to validate the weekly plan assignments, together with other constraint information.  Many things that earlier were learnt the hard way by first failing, and then learning from failures can be anticipated and prevented from happening.

 

Location-based management is a natural part of Lean Construction.  Currently we are working to integrate the Last Planner System and the Location-Based Management System into a new, more powerful tool and process, to better achieve the goals of Lean Construction. 

 

Learn more about the combination of Last Planner and Vico Control:

Vico Control homepage

BIM 401: Model-Based Scheduling

4D Scheduling Strategies in a Hard Bid

 

The paper, The Combination of Last Planner and Location-Based Management System, co-authored with Glenn Ballard, explores the best practices for scheduling and controlling commercial construction projects, and will be presented this Summer at the 18th Annual Conference for the International Group for Lean Construction in Haifa. 

 

After years of experience helping customer companies to be successful with model-based estimating and scheduling, I have come to the conclusion that running parallel processes is the number one enemy of successful process change. Our most successful projects have taken the brave approach and thrown away the old processes and just used model-based coordination, estimating and scheduling tools. Most of the companies we have worked with have not been ready to do this which results in a very difficult environment for change. The difficulties come from change resistance, and the difficulties of showing return on investment when the same process is carried out twice and the two estimates or schedules have to be reconciled.

 

If parallel processes are being used, there are, by definition, two estimates and two schedules. There is a traditional estimate, which has been created by the company estimator with years or decades of experience. And there is the BIM estimate which has been created by the Virtual Construction group of the company - typically young engineers who want to do things differently but don't have a whole lot of experience. Which estimate wins and becomes the official one? Typically, the traditional estimating team will incorporate some of the findings from the BIM estimate and validate their quantities but they are still using their own tools and processes. The Owner gets the official estimate from the estimating group.

 

The same happens with schedules. While the Virtual Construction group is building their schedule using quantities, productivity rates, and optimizing the schedule, the company scheduling department is working together with the superintendents to create the official baseline schedule. Superintendents are more familiar with the traditional process so they prefer to talk to the scheduling department who will put their duration estimates and logic into the schedule without complaint, instead of asking hundreds of questions based on the model (for example, "you just said that the pour will take 5 days in this area, based on the model you would have to pour an average of 500 cy / day to achieve that duration, is that feasible?"). Model-based scheduling process then typically degenerates into following the "official" schedule and trying to give feedback and make minor improvements in some areas. The whole process of optimization becomes difficult. The Owner gets the official baseline schedule from the scheduling group.

 

If parallel processes are being used, what is the ROI of BIM? The Owner is paying twice for the estimate and schedule. The benefits of model-based estimating and scheduling are largely lost because of inefficient processes of reconciliation and having to follow the traditional process. If the reconciliation does not happen, there are two sets of information. Which one is the superintendent going to use when making commitments to the Owner or subcontractors? Of course, the one he is more familiar with. Even if the schedule was optimized, it is not too useful if everyone is going to the traditional CPM schedule for information (unless the dates match exactly, such as by direct exports of data).

 

Caption: It is easy to think that parallel processes means nothing will be missed and any potential mistakes will be caught.  However, our research shows that running two schedules or two estimates just eats away at BIM ROI and keeps project team members at odds.

 

We have always achieved successful results when the estimators and schedulers learn the model-based approach and there is only one model-based estimate and schedule in the project. The Owner pays for the estimate and schedule once, the reconciliation costs and inefficiencies are not needed, cycle times are shorter and the information is more accurate and transparent. Estimators and schedulers are typically not the people who want to drive this change, so the decision needs to come from the executive management of the company. Most often there is an initial time of resistance and large lists of "shortcomings" of model-based approach. If the executive management decides to proceed even despite these difficulties, the project will have a high probability of success. We have experienced high benefits in all projects without parallel processes and limited benefit in most of the projects with parallel processes.

 

BIM should not be a separate line item but should be incorporated as a new tool and process for traditional coordination, estimating and scheduling line items.

 

For more information on the benefits of model-based scheduling and model-based estimating, please review their corresponding webinars: BIM 401 and BIM 501.  For an interesting case study comparing traditional construction methods used during the project and an after-the-fact analysis with BIM methods, please review the Torrey Pines Laboratory episode of Fridays with Vico.

 

For more information on change management in commericial construction firms, please read Mark Sawyer's blog on the topic.

For many people in construction industry, 4D means linking a model to a schedule to produce a 4D simulation of the model.  This may work for marketing purposes because it makes it easy to present the schedule to the Owner and helps the GCs to convince the Owner that they are proficient in the use of BIM.

 

However, this 4D simulation does not help to create better schedules (except for catching obvious issues, such as building 4^th floor before 3^rd floor) or to control the project better.  It is powerful as a visualization tool but if the plan is not optimized, it is of little value because it will not be followed anyway.  Gluing traditional CPM schedules to design models works for demonstrations, but is too time-consuming to be useful in real production - any changes will have to be manually updated, and everyone knows that traditional CPM schedules change monthly and design is continuously evolving.

 

Owners should not be content with pretty movies but should pay special attention to the intelligence behind the schedule - quantities, resources, means and methods.  This information content is a better predictor of whether or not a General Contractor can deliver rather than the ability of the General Contractor to create a snapshot simulation movie.  Owners should require that durations are justified with quantities and productivity rates.  Otherwise, there cannot be constructive arguments, because within reasonable constraints, duration is a choice and a function of quantity and resources.  Owners can be satisfied with the schedule only after all optimization opportunities have been exhausted.

 

For us at Vico, 4D means much more than just a simulation.  3D models are powerful tools to calculate quantities that are necessary to drive the schedules.   Construction is repetitive work in locations.  These locations can be visually identified in 3D models and the resulting location-based quantities can be used to calculate accurate durations for each location.  This automated duration calculation enables crew optimization to achieve continuous work-flow and reduced waste in the process.  This optimization can reduce durations by 10% without increasing resources.  20% reduction is achievable if strong location-based controlling methods are also adopted.  

 

Because everything is linked to the model, all the assumptions are easily verifiable and can be graphically demonstrated.  This 4D makes is possible for a layman to understand why the duration should be 10 days instead of 5 days and why the welding is the bottleneck trade which defines the pace of other activities. The model is used to generate the schedule instead of just linking a schedule to a model - this makes updates easy because any design update will update the quantities which will automatically update the schedule.  A 4D simulation movie is generated as a by-product, the intelligence is what really matters in a true 4D schedule.

 

 

For more information on model-based scheduling, please utilize the following resources:

Vico Control, the line of balance scheduling application

BIM 401 Webinar: model-based scheduling

Mastering the 4D Schedule in a Hard Bid Environment 

 

 

Perhaps the best example of an animation derived to show the schedule can be found in the Hensel Phelps Denver Justice Center webinar.  Scroll to the end of the presentation to watch the animated schedule as compared to the webcam images from the site.  This shows the power of optimizing the schedule AND controlling it on-site.

The book, Location Based Management for Construction - Planning, Scheduling and Control, is now available!  We have been working on the book with Prof. Russell Kenley for some years now.

 

The main challenge has been the amount of new information we learn by continuously implementating more projects. We did two major updates to the content during the writing process, but in January 2009 we decided to stop, freeze the content, and ship the product. Therefore the book contains most of our combined knowledge of location-based management as of December 2008.

 

The book starts by reviewing the history of both activity-based and location-based planning and control systems. It will be new information for many that location-based systems actually have a longer history than CPM. Activity-based systems were easier to code into computer algorithms so they started to win ground in the 1970s and 1980s.

 

The second section is a detailed account of how location-based planning works and how it should be used to optimize construction schedules. The third section is about location-based controlling. Typical books about CPM regard controlling as a natural consequence of planning and do not devote too many pages on control methodologies. This book tries to highlight the critical importance of controlling by having the controlling and planning sections of equal length. The controlling section includes information about how to monitor production, manage changes, how baselining works in location-based schedules, and how forecasts are calculated and used.

 

We also go on to discuss location-based Owner progress reports and running site meetings. In addition to just scheduling, location-based cash flow, cost control, payment systems, design scheduling, procurement, risk analysis and quality control are highlighted to give the reader a comprehensive view about the potential of location-based management to change the way your company does business.

 

The fourth section is intended for practitioners who want to implement the system and change their organization. It summarizes the required parts of location-based management system and discusses implementation challenges. It also gives some practical site tricks and techniques to make control on site easier. The final section of the book is devoted to case studies.

 

I will try to use this blog to discuss questions related to the book and to write about new things I have learned after penning the book. Please be sure to order your copy today!

 

As you wait for your book to arrive, take a look at the software solution that we highlight in the book, as well as the underlying methodology: Vico Control and flowline theory.  Vico Control is a model-based scheduling application.  And the BIM 401 webinar illustrates flowline theory and how it compares to CPM.

When an organization implements BIM, it is not simply adding a new BIM function to its existing business.  If implemented fully, the core processes will change and BIM will affect everyone and everything in the organization.  This article continues the theme of resistance of change to BIM-powered business, and focuses on Estimators.

 

Estimators

When we started to help General Contractors re-engineer their business processes to leverage the full benefits of BIM and integrated workflows, it came as a surprise that there is very little science or data behind today's estimates.  Creating a Standard Database for an organization proved to be difficult because there was no data.  Executives of multiple companies were sure that the data was there but after numerous starts and stops, we came to the conclusion that this data existed only for self-performed work.  For subcontracted work, the estimators relied on their subcontractors to provide estimates.  The skill of the estimator was to ask right questions from right people, and the successful estimators knew the subcontractors who would give reliable information for estimates.  This process did not require detailed understanding of quantities or unit prices.

 

Model-based estimating starts from a detailed quantity takeoff of the construction 3D model (at Vico, we call these construction-caliber quantities, and they serve as the linchpin for model-based scheduling and estimating).  The Estimator needs to define line items which will be quantified based on model data.  Because the downstream processes of model-based scheduling, procurement, and on-site production control will use the same data, it is important that the line items are determined as a team process (for more information on model-based scheduling, please review the recorded BIM 401 webinar from the BIM Master Class Series).  It is important to understand that this hand-off of information between the roles (estimator à scheduler à production control) was very limited in the traditional process.  Schedulers never used quantities and site people did not believe that they could get useful information from their estimators.  The Estimator then needs to apply unit rates for each quantity.  This is complicated because the data is not likely to exist today. To be useful downstream, the unit rates need to be separated to labor and material costs - also for subcontracted work.  The logical way to build this information is to ask for detailed breakdowns from trusted subcontractors (for more information on model-based estimating, please review the recorded BIM 501 webinar from the BIM Master Class Series).

 

All of this is new, so it is no wonder that estimators feel threatened by the new process.  Instead of just estimating an accurate GMP, estimators are now required to provide information which is used by the scheduler to build and optimize the schedule, and by the site guys to buy out the project and to control production (for more information on this preconstruction and production control workflow, please review the recorded BIM 502 webinar).  It would be easy, even motivating, if all of this was just better use of information which was already available.  However, because of lack of data, the information needs to be collected, and the estimators feel threatened by the huge new responsibilities they are suddenly getting.  Everyone will use the information and blame the estimator if it is wrong or not detailed enough!  The biggest threat is that the model makes the estimate and all the underlying assumptions totally transparent.  Double clicking an item in the estimate highlights the item in the model.  It is impossible to "hide" dollars any more.

 

The resistance of estimators typically manifests as downplaying the importance of detailed cost data, and emphasizing the risk of showing model quantities to subcontractors.  They also often highlight the difficulty of getting the information from subcontractors.  Often estimators continue to do their work in parallel with a Virtual Construction group building another model-based estimate with subcontractor data.  This has the same problem as building parallel schedules.  It is not only wasted effort to do things twice, but all downstream processes prefer to use a familiar source of information if it is available.

 

To overcome resistance, the estimators should be supported in the transition.  Instead of making them the only responsible party for the correctness of data, a team effort should be initiated to collect the initial set of data.  Ideally the team would include people from the site team, scheduling team, virtual construction team, and estimators (for more information on this workflow, please review the recorded webinar, Vico Office for Preconstruction and Estimating Teams).  Because the line items do not currently exist, they should be developed with the end-users in mind.  All quantities and information should be reviewed together with the team to ensure that they are correct.  If they are incorrect, the team takes joint responsibility. Estimators then use this jointly-created information to build the GMP, schedulers and site team use it for scheduling, buy-out and production control.

 

It is too much to expect one person to create and take responsibility for all this information.  By making it a team effort, the responsibility can be divided.  In the longer term, the estimator can start to manage it all which will give him/her a lot of power in the organization.

 

Implementing 5D BIM effectively requires a change of process.  Suddenly processes which used to be regarded more art than science are assigned to scientific formulae.  People are threatened by change.  There are some people who see the change as opportunity to get ahead and do their work better. However, most of the people are afraid of the change. The following discussion concentrates on schedulers and planners.  I will discuss estimators, superintendents, subcontractors and Owners in future blog posts.

 

Schedulers and Planners

Durations are no longer based on "experience" from previous projects but are calculations based on quantities, productivity rates, and crew sizes.  Schedulers are forced to reveal that they do not have these rates.  In fact, it turns out in many cases that durations never assumed any crew sizes.  Resources were considered to be "the subcontractor's problem."  Using the project estimate as the basis of the schedule is a totally alien concept to many.  Location-based schedule analysis reveals the weakness of any CPM schedule.  Schedulers often take this as a criticism of their scheduling skill, although they are just using an inferior scheduling methodology.

 

Transparency of location-based management is scary to schedulers who see the schedule mainly as a tool to show delays and as a tool for Owner reporting.  Schedulers often fear that being transparent about assumptions will backfire if the assumptions are wrong (and the author of the assumption is painfully clear!).  They also rely heavily on superintendents who often do not provide crew size assumptions, because it is not part of the existing process.  Schedules have spent years or decades becoming proficient with CPM, so it is understandable that a new way of working which challenges everything they have learned is very difficult to implement.

 

The schedulers' resistance typically manifests itself as running a parallel CPM schedule, persuading the superintendents into using the CPM schedule as the basis of buy-out and reporting to the Owner using the CPM schedule.  If there is a parallel CPM scheduling team, many benefits of location-based management are lost.  It is common that organizations which evaluate location-based management ask the scheduler for an expert view of the technical solution.  The resistance manifests right from the beginning because the planners will test whether the software works using the traditional process.  Most of the identified "shortcomings" will arise from methodological differences between CPM and location-based management, especially the need for planning continuous work.

 

The only way to overcome the resistance of schedulers is to convert them, or to replace them with people who are able to understand the new production principles.  Schedule analysis based on model-based quantities is a good way to start this.  First, quantities are estimated from the 3D model by location.  The location breakdown structure of the CPM schedule is replicated in the model.  The granularity of the quantity takeoff is selected so that each production task will have at least one quantity item.  The logic of the CPM schedule is exactly replicated in a location-based schedule.  The durations are adjusted to match by changing crew sizes (quantities come from the model and productivity rates from industry standard sources).  The end result is a schedule which is resource-loaded and an exact copy of the CPM schedule.  Next, an optimized schedule is created with the same data.  Comparison of the two will in almost all cases reveal that the same total duration can be achieved with less resources and more continuous production or duration can be compressed using the same resources.  This comparison is typically enough to convince most schedulers.

 

If they still resist, pressure from the management who see the same results is typically enough to force a change.  Sometimes the scheduler will go back to the CPM schedule and try to fix it, but it is easy to demonstrate that the problems just shift from one part of the schedule to another.  However, there is nothing wrong with the scheduler himself.  CPM as a scheduling method is failing to deliver results, not the scheduler.  This fact needs to be communicated clearly also to other team members - otherwise the scheduler will resist even harder.  Conversion rate with this method is around 80%.

 

For another opinion on the necessity of change in today's commercial construction industry, please see my colleague, Mark Sawyer's blog, on what you can do to lead the BIM charge in your organization.

 

To learn more about location-based management systems and flowline theory for scheduling, please continue reading through the Fit and Finnish Blog.  Your team can also view the BIM 401 webinar on model-based scheduling.  Vico Control is a unique location-based construction management system.  Vico Control's flowline scheduling system incorporate locations, model-based construction-caliber quantities, and productivity rates yielding clear, accurate and feasible yet significantly compressed schedules.

The most cited benefit of location-based management systems has been a reliable schedule compression of 10-20%.  However, in my opinion, even more important is the capability of location-based management to decrease the risk of the project.  Location-based management decreases the risk of the project in many ways:

-- explicit assumptions

-- continuous production

-- formula-driven forecasting; no "wishful thinking"

-- verification of the estimate through the schedule

 

Explicit Assumptions

CPM schedules record durations and dependencies. Only the person who planned the schedule is able to understand where logic and durations came from and what assumptions are included in the durations. LBMS makes everything explicit. Durations are functions of quantity, productivity, and the number of resources.  Anyone who challenges a duration needs to challenge one of these components.  In CPM, durations cannot be challenged because any duration is right!  You can choose any duration if you staff for it.  However, if the assumptions are not explicit, it is very difficult to actually manage to that duration.

 

Continuous Production

Planning continuous workflow is a realistic model of production.  Subcontractors need to work continuously to maximize their productivity.  If continuous work is not planned in project schedule, subcontractors may legitimately refuse to come to the site because "there is not enough work ahead of them."  If a subcontractor runs out of work and has to demobilize, they will generally not come back immediately when some work is available.  Superintendents realize that return delays are typically between one and two weeks.  If the activity is on a path with float less than 2 weeks, the return delay will, in fact, make the path critical!  This is the main reason why traditional CPM methods do not work and why critical path tends to fluctuate from path to path.  CPM schedules have no way of enforcing continuous workflow and thus are unrealistic models of production.  (For more information on the different between location-based flowline schedules and CPM, please view the archived BIM 401 webinar on model-based scheduling.)

 

In contrast, location-based scheduling forces the planner to make a decision whether a task will be performed continuously or discontinuously.  Continuous work does not increase project duration if production rates of predecessors and successors are aligned.  Additionally, subcontractors typically can perform work much faster and with more quality if they know that they will not be interrupted.  These factors decrease the risk of return delays and thus the risk level of the project.

 

Formula-Driven Forecasting

Forecasting with location-based schedules uses the actual production rate achieved to-date and projects that to upcoming locations and similar types of work performed by the same subcontractor.  The key difference to traditional CPM techniques is that the forecast duration of upcoming locations of similar work is adjusted.  CPM only adjusts the remaining duration of started locations but keeps the upcoming task durations constant.  The typical way of planning control actions to restore the schedule is to decrease the durations of upcoming activities on the critical path.  This is wishful thinking.  Original durations did not have any assumptions related to manpower, quantity, or productivity.  Therefore, changing the durations of upcoming activities will not require a change in any such assumption.  Effectively, CPM planners are cheating both themselves and the Owner.  Because durations are typically decreased in the end of the schedule, this approach leads to huge hurry, overtime work, and cost escalation in the last two months of the project when it is impossible to continue cheating.

 

Location-based schedules recognize that a small delay will morph into a large delay if something is not done now.  Location-based management systems promote early warnings and early reaction.  LBMS requires a change in underlying assumptions - something real must be done to adjust manpower or productivity.  In a large project, it is extremely risky to wait until the last two months to catch up.

 

Verification of the Estimate through Scheduling

In LBMS, the project estimate is used directly in scheduling.  The schedule is an estimate tied to time.  Typically this reveals severe errors in the estimate.  LBMS calculates durations by using the model-based quantity, the estimated productivity rate, and the optimized crew size.  Quantities are objective because they come from the model.  Productivity rates come from the estimate.  Crew size is a choice to synchronize and optimize production rates of preceding and succeeding tasks.  Superintendents typically know how large a crew a particular task needs.  If the required crew size is too large or too small, either the quantity is wrong, or the estimate is wrong, or the superintendent is wrong.  These discrepancies initiate fact-finding which typically reveals that someone on the project team is not on the same page.  The early resolution of these conflicts and the reconciliation of assumptions decrease the risk level of the project.

Owners want to achieve bottom-line ROI from BIM and production management systems.  Scheduling methods such as CPM often fall short because they cannot transparently show (1) whether all possible ways to compress schedule have been utilized in the contractor's schedule or (2) how risky the schedule is.  During production, the monitoring of critical path and float is insufficient because contractors routinely resort to manipulation such as artificially compressing durations on the critical path.  Because the durations were not originally based on a detailed understanding of the scope or resources needed, this manipulation does not really add more value to management and, in fact, can be argued to be pointless - it is like saying that we are delayed now but we will take some (undefined) action later.  Problems are being pushed to the end of the project and there is over-optimism about the ability to catch up schedules later.  Production progress must be made more transparent so that the situation looks more urgent earlier and gives early warning signals to Owners that something must be done to restore production.  All control actions should be explainable in detail by using quantities, resources, and production rates instead of just changing durations.  (For a more detailed explanation of this formula, please review the BIM 401 Webinar on model-based scheduling.)

 

LBMS before the Start of Construction

Because optimized production rates will be used in the buy-out of subcontractors, it is extremely difficult and costly to change the total duration after subcontractors have been bought out.  Therefore, schedule optimization and compression should happen before buy-out.  It is also extremely important that work is planned to be continuous because cascading delays during production happen because of discontinuous work - starts and stops of subcontractors.  Any discontinuity increases the probability that the schedule cannot be implemented as planned - subcontractors tend to balance their own work if the GC or CM cannot show that the plan is already balanced.

 

Owners should be interested in seeing the following deliverables for each major design release:

• Resource-loaded schedules with continuous flow
• Explanations for any discontinuous work patterns
• Clear identification of bottleneck trades by General Contractor or CM
• Detailed explanation why those bottlenecks cannot be accelerated
• Risk analysis of the schedule showing the risky areas and the probability of completing each major milestone
• Cash flow

 

LBMS during Construction

During construction, the schedule should function as an early warning system.  By tracking completed locations, the total quantity of work-in-place can be easily calculated.  Daily reporting of total manpower on site makes it possible to calculate actual productivity and actual production rates.  These rates can be used to forecast progress and identify problems much earlier than in CPM systems.  For a more complete explanation of production control exercises performed on-site, please refer to the BIM 401 Webinar.

 

Weekly or monthly Owner reports should include the following items:

• Comparison of completed tasks and locations to plan
• Explanations for not implementing continuous work as planned
• Schedule forecast before control actions (if everything continues on the same production rate)
• Schedule forecast after control actions and documentation of those control actions (for example, adding 5 carpenters to the drywall crew, working on Saturdays, 2-shift work for selected tasks etc)
• Problem tasks where production rate is < 80% of planned
• Percentage of production completed compared to the plan (for the whole project and each major construction phase, such as foundations)
• Resource forecast for the next month for each main subcontractor to reveal unrealistic resource assumptions
• Updated cash flow forecast based on schedule forecast

 

Parallel LBMS and CPM

Trying to run parallel schedules in a project does not work too well.  CPM schedules have traditionally worked only as Owner reporting tools and have had limited relevance in field.  The information content of a CPM schedule is extremely limited, typically lacking quantities and resources and showing only durations without revealing the manpower and quantity assumptions behind the duration estimates.  Critical path and float are faulty as concepts because they do not consider the requirement to have continuous, balanced flow of resources through the project. Only continuous workflow makes it possible to forecast production and give early warnings of problems.

 

Schedule optimization using location-based planning typically results in 10% compression without the need for additional resources or increase in risk levels. This compression can be achieved so that most of the tasks in the project have continuous flow. However, all optimization is theoretical unless the schedule is actually followed on-site.  Having parallel schedules undermines this goal unless the schedules match exactly.  Location-based scheduling typically results in better quality, optimized, resource-based schedules in 70% of the time required to build a CPM schedule.  Optimizing a schedule from scratch takes about 50% of the time of converting a CPM schedule to LBMS and using that as the starting point of optimization.  All of these facts speak to abandoning the practice of running parallel schedules.  Additionally, CPM is included in LBMS, but only about 20% of LBMS information content is included in CPM (specifically logic layer 5: random CPM links).  Why would the Owner want to pay for two schedules and often for two teams of parallel schedulers?

 

If you are an Owner interested in exploring these ideas, please contact us.  We are more than happy to help you!

In my previous post, I argued that Location-based management gets to the heart of BIM ROI. The goal of this article is to give examples of two projects where durations were successfully compressed by using BIM and analyze the factors contributing to the success of those projects.  I have been personally involved in both of these projects but most of them happened before LBMS was connected to 3D models.  In all of the projects it was a major challenge to get location-based quantities and a lot of manual work was invested to do quantity takeoff manually to get the starting data.  In addition to these example projects, the Finnish contractors have been able to achieve consistent duration compression in most of their projects.  Some projects and companies do not seek duration compression but instead try to increase the probability of achieving the originally planned duration.  This may be even bigger benefit than duration compression because I have never been involved in a project implementing LBMS which finished late!  I have been involved in five projects of this type from start to finish and will write about these case studies and schedule risk in another article.

 

Case study 1: Kamppi Center, Finland

Kamppi Center building complex is located in the center of Helsinki.  The total budget was €500M.  The center consists of a central bus station, a metro terminal, an internal parking area, a six-floor shopping center, and three combined office and residential towers. The total duration of the project was planned to be four years.

 

LBMS was adopted as the sole schedule planning and control system. The implementation started after earthworks phase and lasted until the hand-off of the building.  The goal of LBMS use was to optimize the schedule, to compress duration, and to control production so that the optimized duration could be met.  Two consultants from my team tracked production and analyzed information and generated reports and recommendations for production management.  We were also present in subcontractor meetings showing production status.  The LBMS room came to function as the nerve center of the project because the subcontractors, superintendents, and project managers knew that we had the best information.

 

Our goal was to compress the original four year schedule by six months. Through proactive production control, the GC was able to achieve this and saved millions in overhead costs.  The project was completed in June 2005.

 

Case study 2: Opus Business Park, Finland

Opus Business Park was a case study in my PhD research.  It was a simple office building of 14,528 m2 in Helsinki.  The contract duration was from May 2004 to February 2006.  However, the General Contractor wanted to hand the building over two months earlier than planned.

 

LBMS was again the only schedule planning and control system that was used in the project.  The implementation started from earthworks and continued until the hand-off.  The project engineer used the software, and I helped by having weekly meetings on site where we analyzed progress and planned control actions.  The project engineer communicated the required actions to the project manager and superintendents and to subcontractors in weekly meetings.  Initial schedule optimization and risk analysis revealed that it was possible to compress the duration by two months without adding too many resources.  Location-based milestones were given in RFPs to subcontractors and the project finished according to the compressed schedule.

 

Company -level implementations

In addition to individual case studies, five large GCs in Finland have implemented LBMS to be their only allowed planning and control system. All GCs have reported good results - better overall quality of schedules, consistent duration compression, and better project data for use in company -level systems, such as procurement systems.  An example implementation case study can be found in (Soini, Leskelä, Seppänen 2004).

 

Why are all of the successful case studies in Finland?

The Finnish construction industry has practiced location-based techniques since the 1980s, starting with manual techniques (pen and paper), evolving to simple flowline drawing software in the 1990s, and finally to Vico Control (previously DynaProject) in 2002.  Already in the 1980s, location-based management was shown to compress schedules and result in better control of production.  Gantt charts resurfaced in the 1990s because of better availability of automated software for CPM applications.  Because of this history, it was easy to convince people to throw everything else aside and do the project completely with LBMS from start to finish.

 

Although the earliest known implementation of LBMS in construction is in the US (the Empire State Building), US contractors do not have the same familiarity with LBMS.  The prevalent scheduling method is CPM, and production control is practiced informally in the field by experienced superintendents, who each have their own systems of tracking data.  In 2007 and 2008, the GCs have started to experiment with production control but most of the projects are only now entering construction.  Because typical large scale construction projects take years, we do not have case studies which have been carried out from beginning to end using LBMS.  Many projects are now getting close to production, but it will be years before the first case studies will be completed.

 

Similar results to those in Finland - 10-20% duration compression and productivity improvement - can be achieved in US projects.  The most innovative contractors are already doing this.  Definitive bottom-line end results will not be known until the first projects are finished, but all the contractors already appreciate the better quality of information for decision making.  It is best to start now - we have the experience required to make it work for you.

  

References:

Soini, M., Leskelä, I. & Seppänen, O. (2004) Implementation of Line-of-Balance based scheduling and project control system in a large construction company. Proceedings of the 12th Annual Conference of the International Group for Lean Construction, Helsingør, Denmark