Over the past dozen years, my involvement in large capital projects has provided abundant opportunities for observing how things went wrong. So, let me offer a few thoughts on what can undermine project execution and how to avoid these traps. First, the most obvious culprit is project delay. Extensively stretching out a capital project or working only fitfully on it practically guarantees the people who start the project won't be around to finish it. This seems to be one of the major reasons for missteps.Second, confusion breeds more confusion. A constructor really doesn't care about the careers of two refinery unit supervisors fighting over a flare stack. Nor do they want to manage the stream of heat tracing change-orders resulting from incomplete knowledge of steam supply and electric loading.[pullquote]Andrew Jergens' expansion of its Cincinnati plant was a success because the company spent two years identifying what had to be done and how it would do the work. This methodical pace allowed sufficient time to weigh operations' pet projects on their merits. It's called front-end engineering — if you don't do it or muddle through it, prepare for disaster.Do you really want a half dozen welders sitting around waiting on a work order? Training everyone on the work order system helps but it isn't enough. You need adequate enforcement. You must police the system. Recently, I talked to an electrician on a subcontract crew. He sat around all morning waiting on a production operator, who then went to lunch. You must discipline people like that operator. Study how long it takes to complete a work order from the time a project package is approved until it's signed off by production. Inventory management is part of this process. People in maintenance departments sometimes have sticky fingers; so, establish a secure site for project inventories. In a well-run turnaround, the approval shacks are staffed 24/7; delays and thievery don't happen. That brings me to my next point — you need an inspector general. This person's role is to root out and smooth over interactions between the overall site contractor, subcontractors, and your maintenance and production staffs. The inspector general must have the authority to over-ride your people (like that operator), if they become the bottleneck. One useful approach is an anonymous system for complaints. This gives the person a means to improve efficiency. It helps if the inspector general potentially could get promoted based on performance on this assignment. Staff then may worry that any difficulties they cause the person during the project could come back to haunt them later. Here's something else to consider: As projects come to a close, re-works and add-ons accumulate. Assign a team of contractors and your project engineers to list and prioritize these items. This group should be separate from the construction team.Quality control is another concern but one that large projects generally do well. Leave some money in budgets for critical re-works. Don't expect the constructor to pay for everything as if it's solely responsible. If an operating company insists on a bad design, despite the constructor's warnings, the company should shoulder the cost of any necessary modifications.There's more to doing a job well than efficiency. Too many projects fly through hazard and operability (HAZOP) reviews, but once built simply won't work. A few years ago, I saw a design with a flame arrestor installed on an atmospheric tank vent and the overflow. This is an unworkable design. Remember, operability is part of the HAZOP review process. Squash stupid ideas during the design HAZOP review process.Now, let's consider some changes you can make during the front-end design. Every project should begin with a basic engineering design (BED) report that details things such as elevation, drawing standards, maximum ambient temperature and barometric pressure. It also should define software tools for models. I wish I had a thousand dollars for every time I've seen an engineer trying to use an incompressible fluid package to design for vent flow. In addition, the BED report should specify what goes on each drawing. For example, why would you crowd your P&IDs with programming? Why aren't your instrument details on the isometrics? A good BED report should include workflow diagrams for changing drawings and monitoring quality, i.e., how equipment purchases will be processed from selection through handover. Frequently during large expansions, drawings and specifications lack uniformity. Too often, a different index is used in distillation than in gas oil hydrotreating. If you build hurdles, people are sure to fall over them!