Hi Thomas,

Consider the reducing risk - this has to do with make a decision most often. There is uncertainty around the facts - like how reliable the design is or is not. It is something and time will reveal the answer, it’s just the information we have today that is causing the uncertainty. So, if we something to peel back the curtin of uncertainty a little, say an ALT, we don’t change the failure rate, we learn what it is and how likely the design is to last a suitable long time with a sufficiently low failure rate or not.

The value is not the failure rate it is the information. So, how do you account for this value?

Before the ALT and based on our organizations history we have launched 25% of prior product with higher than desired reliability. This means 75% of the time we got the decision right and enjoyed the benefit of a successful product launch, 25% of the time we paid higher than expected warranty costs. This was in the case when we decided to launch.

another way to consider this is the other half of this decision is to delay the launch - which happens on half the programs over the past 5 years in our organization - again this is driven by uncertainty and expected the product is not reliable enough, so we decide to delay, and incur additional development time costs, or launch and run into the above uncertainty.

In either or similar situations the failure rate is not changing, we just don’t know it. So do ALT or some other activity to learn a bit about what the failure rate really is.

After ALT we can reduce the uncertainty - this is not an exact science and I’ve typically just asked the decision maker for their understanding of the risk - before ALT 75% certainty we’re good, after ALT with it’s clearly presented information, she may say we’re 85% certain of being right, it’s a low failure rate.

So the actual failure rate is either low and all is good, we get the profit from sales we expected…. or it is higher than expected and we incur less profit due to warranty and related expenses being higher.  So, we are shipping 10k products at $1k each and expect with a low failure rate a profit of $2.5m - the decision based on our certainty would expect 0.85 * 2.5m… as there is a chance of a higher than expected failure rate and say, a loss on the product launch of $1.5m… warranty, rework, lost sales, recall, etc… again this based on our assessment of the uncertainty is 0.15*(-$1.5m)  tally these to number to find the expected benefit based on the decision. Similar tally done prior to the ALT and the difference is the value of the ALT’s information.



In the Finding Value book there is a checklist or set of questions to ask someone about how the value was created for them in a particular situation - the idea it might be less risk around a decisions or saving time by avoiding having to relearn or redo something, etc… The primary idea of these homework questions and you’ll see in making proposals in the exams and project is being able to articulate value clearly and not only from a failure rate reduction.

The decision makers we work with do not always value a lower failure rate as the primary goal - it may be shipping on time which often entails reducing uncertainty around decision making concerning shipping or not.

So, yes, this process can be difficult if only focused on failure rate changes - yet well worth the effort to master.

give it another go - think about what is of value to the decision makers in our organization or by customers

Cheers,

Fred


On Jun 25, 2019, at 5:03 AM, Thomas Rands <tom.rands23@gmail.com> wrote:

More attempts at questions from yesterday. What else can I do to come up with different sources of value that are related to reliability but not related to failure rate? Based on the feedback I feel like a lot of my suggested answers have no path to a correct solution so I keep having to start again and come up with new ideas without using the specific examples spelt out in the lectures or finding value document.



List 5 different ways reliability is of value to a business. For each of the 5 ways, what specific information is needed to quantify the value? Show the formula necessary for a calculation of the value.
Decrease warranty budget
If we spend extra time to design the system for reliability so that it is 2x reliable and 2x cheaper to repair than we can reduce the warranty budget and decrease yearly expenses to improve the company’s financial statements.
Old warranty budget = average number of failures/unit * average cost to repair * number of units
New warranty budget = (0.5 * average number of failures/unit) * (0.5 * average cost to repair) * number of unit
The cost of shifting to 2x reliable requires investment in the form of more engineering hours, additional reliability activities, potentially more/less expensive components, less time on the market (and potentially others)
Cost of investment = engineering hours * engineering hourly rate + cost of analysis activities + cost of design activities + cost of components (may be negative) + days not on market * expected sales per day * unit sales price.
Decrease in expenses in year 1 = old warranty budget – new warranty budget + cost of investment
Decrease in expenses in year 2+ = old warranty budget – new warranty budget


Reduce risks/uncertainty
Without conducting a reliability assessment like ALT, the current design might be reliable enough to begin production/sales, but this is uncertain and if the design is not reliable enough not than we may be required to conduct a lot of redesign and recall of products. If we conduct the reliability activity, then we can decrease the uncertainty about whether the current design is reliable enough.
Cost of unreliability = engineering effort to redesign + cost of altering production line + cost of changing vendors/components + cost of recall
Cost of reliability = engineering effort to conduct assessment + cost to conduct assessment
Potential value = cost of unreliability – cost of reliability


Improve market share
If we advertise a product as functioning within certain operating conditions such as stress, temperature, pressure, depth or altitude but we are increasingly seeing field failures that indicate the product was used in those condition. In this case we could create value by consulting users to determine the most profitable improvements based on where/why the product is used and number of potential new customers, then increase the reliability of the product in a wider range of operating conditions, and then advertise the new product to a widened audience in order to improve market share.
Costs = engineering effort to improve operating range of product + production costs of modifying the design + cost of user consultation/surveys + cost of analysing user data/defining new use cases + cost of marketing to potential new customers
Benefits = number of potential new customers * product sale price
Value = Benefits - Costs

(this one seemed like a bit of a stretch but I wanted to try to show how I might calculate the value of increasing the use cases of a product)





For a hypothetical company HH, Inc, net revenue $60 million/year with 50k units sold each year, calculate 3 specific examples of increased value by moving reliability maturity from stage 3 to stage 4.



Supply Chain Management. Cost savings associated with improving supply chain management by using vendor reliability data to select vendors. Higher reliability of vendor parts means less warranty claims (assuming that the vendor is not charged for the warranty claim). Example will be for 1 part, let’s say a temperature sensor, but could be extended to cover x parts. We are going from the current vendor to the top vendor, which we only determine by using vendor reliability data.
Assume that before the change we are using the same vendor we always have, who sells their sensor at market rate ($1) and has a reliability of 90%.
When we decide to assess vendors based on reliability data, there are 5 possible candidates. A reliability team of 3 people spends 5 days assessing the vendors using their data, their daily rate is $200/day. The best candidate has a pump with a reliability of 99% and charge 1.2x market rate ($1.20). An additional 2 days are spent conducting tests to confirm the pump is as reliable as the vendor data suggests.
Assume that average cost for a warranty claim is $600 (based on a unit price of $1200).
Cost of using old vendor = 50000 * 0.1 * 600 = $3,000,000
Cost of changing to new vendor = 50000 * 0.01 * 600 + 3 * 7 * 200 + 0.2 * 50000 = $314,200
So, the cost savings of changing to the new vendor, based on reliability, is $2,685,800.



Proactive DFR. If we choose to spend more time to increase reliability of design, there are less failures so less man hours are spent fixing defects after the product goes to market.
We assume that a physics of failure analysis is conducted which identifies a failure mechanism that can be removed by altering the design. The failure mechanism would have contributed to a 0.5% failure rate (reduced to 0 after the DFR activity). The analysis activity required 3 people and took 10 8-hour days. Assume that the repair cycle for a defect takes 2 hours on average and requires 2 people.
BEFORE DFR. Time spent fixing defects = 50000 * 0.005 * 4 * 2 = 2000 hrs/yr
AFTER DFR. Time spent on proactive DFR = 3 * 10 * 8 = 240 hrs
So, 1660 engineering hours are saved in the first year, and 2000 hours each year after that, which is almost equivalent to one full time person conducting repairs (2085 hours assuming 260 8-hour work days).