Creating Proper Truss Labor Estimation instead of Board Foot or Other Flawed Methods

Recently I unintentionally invoked an intense argument about the common practice of using board footage (BF) as a means to monitor truss shop labor efficiencies. My argument was that BF is not an accurate means to gauge truss manufacturing efficiencies and others said that it is so widely used that I must be mistaken. First, let us look at what we are trying to achieve.

You cannot use total times of completed built trusses to develop time standards but you can use them to confirm your time standards. This is a very important distinction about developing time standards. You cannot use history of production to develop accurate time standards because of the following important factors:

  • Time standards define what should be, not what has been. Just because it took a certain length of time to build the truss types in the past does not mean that past results are good indicators of what future performance should be. Past performance may have been poor.
  • Historical data does not allow you to understand which work orders need to be ignored. Averaging the good and poor performances in previous work orders’ time summations only skews your time estimates towards lower expected rates than they should be.
  • The time standards can only be derived from individual tasks. You cannot break down the individual tasks with start and end times of all the different tasks. How do you know the assembly time of the truss versus the setting up of the jigging with a combined time?

When you create your time standards (labor estimations), why not use properly defined methods developed by industrial engineers? What are proper time standards defined by industrial engineers? The time standards must be developed using the following criteria:

  • They are based on specific tasks performed by experienced crew members working at a steady pace.
  • You do not group different tasks together. An example would be measuring how long it takes to move a 2x6x14′ onto the build table’s jig. You do not combine the other task of moving the 2x6x14′ by including other smaller or larger lumber onto the jig. The industrial engineer wants to measure just the statistical average of movement of lumber types (specific task) onto the build table’s jig. Another example of creating time standards the wrong way would be combining the assembly of the truss lumber parts with plating of joints and press rolling of the truss. This would be a huge mistake because each task needs to be measured individually.
  • You do not want to define your time standards with unskilled individuals.
  • 100% of the time standard is a normal experienced crew member working at a steady pace. This is not like a machine that if it exceeds 100% it will become broken. Workers can achieve 120 to 140% if properly motivated.
  • It can measure consistently in clearly defined units.
  • If you measure the task repeatedly, the same task should show very small deviation statistically. An example of this feeding a 2x8x10′ into the component saw. If you measured the task today or a year from now, the time measurement result should fall within a certain range to be valid.
  • The unit of measurement should be consistent with time units of man-minute, RE, or SU. Reasonable expected units (R. E.) and scheduled units (S. U.) are each defined in decimal minutes. This means 1.5 R. E. is simply 1:30 minutes. You can give the time unit other names such as “truss nuggets” but it has to be consistent with a measurement of time.
  • The time measurements are divided into three categories: run/assembly, setup, and other.
  • Run/Assembly time – Two examples of this are the assembly of trusses or running the component saw. This type of activity is the hardest to increase in speed without incentive programs.
  • Setup time – Two examples of this are the build table jig setup or setting up the saw. This is an easy area to improve by spending large amounts of money on automation.
  • Other/support time – Examples of this are lumber picking, plate picking, truss stacking, and any other support activity. This cannot be measured consistently and is not included in the measurement units. An example is picking lumber from the lumber units. If you pick a quantity of 20 pieces of lumber from a unit of lumber versus taking the whole unit of lumber to the saw, how would you break down the time units per lumber piece? The answer is you cannot do it consistently and therefore it cannot be used.

Units for Time Standards

  • It is easy to understand why most truss shops use the common types of units to estimate labor time. All of them are easily derived from the truss design program. However, none of them can be directly linked to actual labor man-minutes required for any given order. They are: 1) board footage, 2) linear footage,  3) material cost,  4) sales dollars, and 5) number of lumber pieces (latest seriously flawed method).

Board Footage – Do you think your employees would rather build pole barn trusses or a hip roof system based on board footage? Every production supervisor will tell you that board footage is a very poor method of understanding how long an order will take to produce. Board footage does not have any consideration for setup time for saws or tables. So anytime you have numerous setups in a build order, your BF will suffer tremendously at the build tables. And to further make my point, using average rates of BF by combining many different work orders over a given time period will not provide accurate levels of rate efficiency for the per workstation’s task! Here is an example: one build table might have a run of very simple trusses to build with very few setups while another table has many setups with complex shapes. How can they be both measured the same using “average” BF rates? The answer is they cannot.

Linear Footage – Are common trusses as easy to build as customized profiles with multiple pitch breaks? Again, setup time is completely overlooked and other factors such as number of varying time requirements of piece types is completely ignored.

Material Cost – When the material costs fluctuate because of market conditions, does the ease of building the trusses coincide?

Sales Dollars –Stating that you need a certain sales dollars per day from your production translates into needing a certain amount of margin dollars to cover costs and make a profit. This has nothing to do with rate of efficiencies for production. When an order has been discounted, does it mean they can build the order faster?

Number of Lumber Pieces – The number of lumber pieces measurement does not properly handle the setup. Often there is no adjustment for lumber length and size. Some companies will create an “average” setup time and add it to the individual piece rate. The “average” setup time divided over numerous pieces skews the labor estimations. For orders with very few setups, the time estimation is overestimated and complex orders with complex setups are extremely underestimated.

Estimating time by the number of pieces example

Order #1

  • Automated saw has a setup time of approximately 30 seconds per unique setup
  • 1 Sawyer and 1 lumber catcher for 30 seconds equals 1 man-minute for piece setup
  • A quantity of 6 of 2x4x8’ with an average run time of 5 seconds each
  • 5 seconds multiplied by 2 crewmen equals .17 man-minutes per piece
  • Total time = (.17 runtime per piece * Qty of 6) + 1 saw setup = 2.02 man-minutes with proper setup.
  • Average time per piece is (2.02/6 = .337)

Order #2

  • Automated saw has a setup time of approximately 30 seconds per unique setup
  • 1 Sawyer and 1 lumber catcher for 30 seconds equals 1 man-minute for piece setup
  • A quantity of 20 of 2x4x8’ with an average run time of 5 seconds each
  • 5 seconds is multiplied by 2 crewman equals .17 man-minutes per piece
  • Total time = (.17 * 20) + 1 = 4.4 man-minutes with proper setup.
  • Average time per piece is (4.4/20 = .22)

That is a 65% difference for just 2x4x8’s. There is no conceivable way you can use average time per piece and be consistent with your labor estimations.

Do NOT let some plate/software vendors tell you that piece count is a proper method for labor tracking efficiencies! Their software cannot track via proper man-minutes and they are trying to convince you that it is good enough. For a better labor tracking software alternative see my labor tracking software. Truss Shop Labor Tracking Software.

Now you are asking yourself if it really matters whether you know how long it takes to do the individual activities. Industrial engineers learn the following:

For Most Manufacturing Shops – Gains Expected from Clear Time Standards*

  • No Time Standards = 60% of performance*
  • With Proper Time Standards = 85% of performance*
  • With Proper Time Standards and Effective Incentive Program = 120% of performance*

*All performance figures have a normal deviation range of +/- 10%. Source from case studies with Industrial Engineering Consultants Mitch Fein and Fred Myers. These studies have remained consistent across various other bodies of work and are considered statistically valid.

On average, most manufacturers gain a 42% increase in productivity by simply employing proper time standards. ((85% – 60%); divided by 60 = 41.7% gain). What manufacturer would not like a 42% gain in productivity? By the way, that is the secret of the Houlihan system. My new labor tracking software is an automated Houlihan time standard tracking software system.

Accurate time standards (time measurement) allow you to clearly understand the following:

  1. The shop’s efficiency rates (how well are they producing);
  2. Labor cost per order;
  3. How many units they are capable of producing;
  4. And incentive programs that actually work to motivate the shop to produce more.
  5. With that information scheduling becomes a snap to calculate, even many days and weeks ahead of actual build dates. Staffing also becomes a no-brainer.

Todd Drummond Consulting

  • 10th year as a full-time lean manufacturing consultant.
  • 50+ consultations completed.
  • 20+ years in the truss industry.
  • Truss manufacturing Time Standards:
  • 80 different time elements to choose from for differing conditions
  • All time elements are based on man-minutes in decimal units and are for normal staffing.
  • May use time estimations for Man-minutes, R. E., or S. U.
  • All labor estimations are for “Setup” and “Run/Build” on the following equipment assets:
  • Semi-Auto Fence Line Pull Saw
  • Linear Automated Saw
  • 4 or 5 Blade Manual and Automated Component Saws – all types of component saws
  • Roller and hydro press gantry tables for roof trusses,
  • with or without auto-puck jigging, and laser and no-laser jig setup system
  • Roller and hydro Press gantry tables for floor trusses (any kind of gantry table with various setup systems)
  • Does not include any kind of support activities such as lumber/plate picking, lumber stacking, or other kinds of support activities.

Todd Drummond Consulting

http://www.todd-drummond.com/

todd@todd-drummond.com

603-763-8857