_Trying out the product in the real world
Testing a design to see if it works is not just a single thing that happens. There are different aspects to testing that mean different things to different people. These are detailed at length in the article 'A Test, an Experiment and Two kinds of Trial', but in essence it is broken down to the following:
The Test
This is a term that should be reserved for the engineer or the designer responsible for getting the project to go. A design specification is made. The engineer produces something and performs a ‘Test’ to show that it conforms to this specification, regardless of how appropriate it is in the field. If it passes ‘The Test’, it is not ready for sale, it is ready for ‘An Experiment’.
The Experiment
An Experiment is the first real-live introduction to the field to see if the experience in the test lab can be replicated where the product must operate in the real world. It may involve attaching the product to a vehicle and driving it around, or introducing the item into a much larger product and see how well the two parts work together. If the experiment fails, its back to the lab for changes, more Tests and another Experiment. This step is strictly ‘Eyes Only’ and customers should not get involved.
Trial 1
This Trial step resembles an extended Experiment. The Experiment has shown that the product will function, but it is still pretty green and you must seek information about its longer term reliability. Will it overheat? Crack with repeated use? Become susceptible to interference in certain situations? These are questions that must be answered in the Trial 1 stage. Often it is difficult to replicate real-world conditions without the involvement of a customer. If so, make it plain what the trial is for and what you hope to achieve. Make it known how long you expect the trial to last. At the successful end of Trial 1 you can declare the item ‘A Stable Product.’
Trial 2
This is a customer oriented trial. It assumes that the core design of the product is now stable and beyond reproach. Any problems that do occur should be traceable to a flaw in the manufacturing process or the installation process. If a design problem appears, then the Trial 1 step was not done properly.
Trial 2 is largely an acceptance trial. The customer will say ‘Sure it does what you said it will do, but ‘Is it right for me? Does it solve my problems? Should it do something extra we had not yet thought of?’ Have good documentation and a check list describing all of the functions. If there are changes to specifications as a result of this trial, then that’s ok, just document the changes and go back to the engineer. If there are failures and breakdowns during this trial the customer will rapidly loose confidence in the product and the company who made it.
Production Testing
Of course the process described here relates mostly to the development of first working models. Testing in production is different, in that the emphasis is on ensuring each assembled unit performs key functions on a check list. It is not necessary to test every aspect of every feature. If all inputs respond correctly to stimulus and all outputs are driving their essential loads, then it is pretty reasonable to expect that this unit would then be as good as the heavily tested prototype.
Click here to return to CREATING A NEW PRODUCT
Testing a design to see if it works is not just a single thing that happens. There are different aspects to testing that mean different things to different people. These are detailed at length in the article 'A Test, an Experiment and Two kinds of Trial', but in essence it is broken down to the following:
The Test
This is a term that should be reserved for the engineer or the designer responsible for getting the project to go. A design specification is made. The engineer produces something and performs a ‘Test’ to show that it conforms to this specification, regardless of how appropriate it is in the field. If it passes ‘The Test’, it is not ready for sale, it is ready for ‘An Experiment’.
The Experiment
An Experiment is the first real-live introduction to the field to see if the experience in the test lab can be replicated where the product must operate in the real world. It may involve attaching the product to a vehicle and driving it around, or introducing the item into a much larger product and see how well the two parts work together. If the experiment fails, its back to the lab for changes, more Tests and another Experiment. This step is strictly ‘Eyes Only’ and customers should not get involved.
Trial 1
This Trial step resembles an extended Experiment. The Experiment has shown that the product will function, but it is still pretty green and you must seek information about its longer term reliability. Will it overheat? Crack with repeated use? Become susceptible to interference in certain situations? These are questions that must be answered in the Trial 1 stage. Often it is difficult to replicate real-world conditions without the involvement of a customer. If so, make it plain what the trial is for and what you hope to achieve. Make it known how long you expect the trial to last. At the successful end of Trial 1 you can declare the item ‘A Stable Product.’
Trial 2
This is a customer oriented trial. It assumes that the core design of the product is now stable and beyond reproach. Any problems that do occur should be traceable to a flaw in the manufacturing process or the installation process. If a design problem appears, then the Trial 1 step was not done properly.
Trial 2 is largely an acceptance trial. The customer will say ‘Sure it does what you said it will do, but ‘Is it right for me? Does it solve my problems? Should it do something extra we had not yet thought of?’ Have good documentation and a check list describing all of the functions. If there are changes to specifications as a result of this trial, then that’s ok, just document the changes and go back to the engineer. If there are failures and breakdowns during this trial the customer will rapidly loose confidence in the product and the company who made it.
Production Testing
Of course the process described here relates mostly to the development of first working models. Testing in production is different, in that the emphasis is on ensuring each assembled unit performs key functions on a check list. It is not necessary to test every aspect of every feature. If all inputs respond correctly to stimulus and all outputs are driving their essential loads, then it is pretty reasonable to expect that this unit would then be as good as the heavily tested prototype.
Click here to return to CREATING A NEW PRODUCT