Manufacturing Strategy for Window Fabricators 4 - Optimised Production Technology




  Download Adobe Acrobat file of complete series to date

Go with the flow

Optimised Production Technology (OPT) provides the methods to deal with the inevitable production bottlenecks. Production managers have has some new tools to play with. A review of the basics by Robin Kent of Tangram Technology.

The goal

OPT states that the goal of a manufacturing business is to make money both now and in the future. This might seem to be rather simple but it provides a framework for all the other decisions involved in the business.

OPT aims to increase 'throughput' (the rate at which the company generates money through sales) whilst simultaneously decreasing both 'inventory' (the money the company has invested in things it intends to sell) and 'operating expenses' (the money the company spends to turn inventory into throughput). If an action does not directly improve one of the three measures then it is irrelevant at best and damaging at worst, do not do it.

The traditional approach has been to optimise every sub-system irrespective of its importance (i.e. to improve the output of the welder) but the OPT tries to optimise the total system to maximise throughput (i.e. if the welder is not limiting the throughput then don't work on it and put the efforts somewhere else). OPT states that the optimum of each sub-system is not necessarily the optimum of the whole system.

Balance the flow not the capacity
Gearing is the obvious bottleneck for the factory. Running cutting at capacity will lead to a build up of inventory in front of welding. Running welding at capacity will lead to a massive build up of inventory in front of gearing. Investment or improvement in cutting, welding or glazing will do nothing to improve throughput, the only meaningful investment area would be at gearing where the ability of the plant to earn money would be rapidly improved.
Gearing must be protected from loss of output for any reason. It is the operation that controls the income of the factory. In reality the choice is never this clear and the bottlenecks move as the product mix changes. Gearing may be the bottleneck when making a lot of doors but glazing will be the bottleneck when making lots of fixed lights.

OPT defines a 'bottleneck' as a resource whose capacity is equal to or less than the market demand placed upon it. This is the constraint that is preventing increased throughput from the factory. Improving the bottleneck will start to optimise the whole system and directly increase throughput and profit. Bottlenecks are easy to spot in the factory - they are the operations with lots of WIP stacked up in front of them. Conversely, a non-bottleneck is a resource whose capacity is greater than the market demand and improvements here will not increase throughput.

The bottleneck concept is best explained in the hiking analogy from 'The Goal'. The speed of a group of hikers must be maximised to get to the campsite by nightfall - but the speed of the group is limited to the speed of the slowest hiker (the bottleneck). If the slowest hiker is at the front of the group this slows down the whole group and increases the time required, i.e. reduces the throughput. If the slowest hiker is anywhere else in the group they still slow the whole group but also increase the length of the group (the inventory).

The only way to reduce the length (the inventory) and achieve the fastest time (the throughput) is to find a way of moving the slowest hiker faster, i.e. working on the bottleneck. An hour lost at a bottleneck, for any reason is an hour lost to the whole system and cannot be recovered. Don't think you can get it back later because you can't. The cost of the lost hour is the total cost of running the factory for one hour because the bottleneck is governing the throughput.

The location and management of bottlenecks is the heart of OPT. An hour saved at a bottleneck is an hour saved for the whole system but an hour saved on a non-bottleneck machine simply increases inventory and does nothing to improve throughput. It is wasted effort.

OPT and JIT both concentrate on quality, lead times and batch sizes but OPT regards the 'river and rocks' analogy of JIT as flawed. In OPT terms, the river is not a flat and evenly flowing stream but has waves of inventory (WIP) moving through it depending on the order situation in the factory. All will be fine until the inventory is at the bottom of a wave. A problem then it is likely to rip the bottom out of the boat and sink the business! This is closer to reality than the JIT model.

The river and rocks model of OPT
OPT considers that the inventory river is not smooth but consists of waves of inventory going through the factory, depending on the order state at any given time. Rocks that may be acceptable when on an inventory peak are not acceptable when in a trough.

OPT and MRPII share a computer based approach with databases of product and machine information for schedule calculation. OPT also requires information on how the product is made, the production route, set-up times and run times. Whereas MRPII assumes that a machine can always work at capacity, OPT accepts that the actual capacity is affected by statistical fluctuations and a dependence on previous operations to supply product for processing. This makes OPT can be more realistic in scheduling than MRPII by taking this into account and also allowing for improvement in times and routing.

The OPT rules

OPT is based on a set of rules:

  1. Balance the flow, not the capacity.
  2. Let bottlenecks determine usage of the non-bottlenecks and do not seek machine utilisation. If a resource is activated when output cannot get through the constraint then all it produces is inventory.
  3. Utilisation and activation of a resource are not the same thing. Activation is when a resource is working but utilisation is when it is working and doing useful work. Producing stock for inventory is not useful work.
  4. An hour lost at a bottleneck is an hour lost for the whole system and cannot be recovered and an hour saved at a non-bottleneck is a mirage.
  5. Bottlenecks govern both throughput and inventory.
  6. A transfer batch is not necessarily equal to a process batch, i.e. because you cut 20 frames at a time on the optimiser saw it does not mean that you have to push them all to the welder at once. Break the process batch (20 frames) down into smaller process batches (1 order).
  7. Process batches should be variable and not fixed.

Bottlenecks beat out the pace like a drum and should be protected from interruptions such as breakdowns, quality, labour shortages etc. This can be achieved by building in time buffers that are a focus for process improvements. The other operations are synchronised to the bottleneck operation and work is pulled through as if it were on a rope.

OPT rarely requires large investment in machinery but seeks to improve the flow of the product and get inventory moving to show that production area as the real profit maker for the company.

For and Against OPT


  1. Easy to start at practical level and easily understood by the shop floor
  2. Quickly targets areas of concern such as bottlenecks, quality and inventory.
  3. Gives quick results and financial feedback.
  4. Suitable for discrete, batch and process industries.


  1. Challenges traditional cost accounting.
  2. Requires accurate database and complex computer process modelling for the full form.


OPT is a philosophy for running the business rather than simply production. It tries to optimise the complete system rather than simply optimising individual operations on a piecemeal basis.

The full version of OPT is a proprietary system owned by a software and consultancy company but this does not prevent using the excellent ideas it contains. OPT is a trademark of the Scheduling Technology Group. The only, but excellent, book on the subject is 'The Goal' by E Goldratt and J Cox.

'Manufacturing Strategy' Series.

The 'Manufacturing Strategy' series is designed to give window fabricators a set of ideas for managing production. The series is being published in Fenestra on a monthly basis and published here after the Fenestra publication. The series is:

Part 1: The Essential Part 
Part 2: The Systems
Part 3: Just-in-Time

Part 4: Optimised Production Technology

Part 5: Work Cells

Part 6: Machines
Part 7: Machines (2)

Part 8: Scheduling

Part 9: Waste (Methods)

Part 10: Waste (Materials)

Part 11: Supply Chain

Part 12: Measurement
Part 13: Things to do NOW!
Part 14: The Cost of Quality

Part 15: The Hidden costs of inventory

Part 16: Environmental management

Part 17: Continuous Improvement

Last edited: 11/03/10

© Tangram Technology Ltd. 2003

Our standard disclaimer regarding Internet data applies.