The 70% rule to success

Thread Starter

cmartinez

Joined Jan 17, 2007
8,220

ericgibbs

Joined Jan 29, 2010
18,766
hi,
Often I get some criticism over my signature motto, all that I can say, it worked for me.

It can be hard to rein in the enthusiasm of your engineering design team, you have to remind them that all the product has to do, is to meet the clients specification and stay within the design cost budget.

E
 
Last edited:

jpanhalt

Joined Jan 18, 2008
11,087
Hi E,

I worked in a field where the standards were quite high but have no problem with your signature and basically agree wit it. Another way we put it is, "Don't let perfection stand in the way of getting the job done."

That doesn't give a license to be sloppy. For example, in a clinical laboratory environment, a good enough result two or three days sooner was not only satisfactory, it was better in most cases than a more refined result reported later. Either result -- even ones based on an opinion -- had to be close to 100% correct though.

John
 

wayneh

Joined Sep 9, 2010
17,496
The irony of your perfectionist demand to the proposed "magic number" is not lost on me ... :D:p
The article offers sensible advice but there's a big problem with picking a number and thus putting this into practice:
• The percentage varies with industry and probably by the project
• We're not often aware of what percentage is appropriate for any given decision​

It's similar to the problem in statistics of choosing the alpha error (odds of incorrectly rejecting the null hypothesis when it is in fact correct) versus the beta error (odds of missing a real effect by not rejecting the null hypothesis). People publishing papers usually want the alpha error to be below 5% or even 1%, but not everyone using statistics is publishing a paper. Out in the real world, choosing between two suppliers for instance, you need only a modest amount of evidence that one is better than another. In fact you have no choice but to go with the winner of a 51/49 match. You want to choose the better one but you can't afford to collect enough data to make a "certain" decision. You live with high uncertainty.
 

WBahn

Joined Mar 31, 2012
29,979
The irony of your perfectionist demand to the proposed "magic number" is not lost on me ... :D:p
I didn't figure it would be.

While it was largely a tongue-in-cheek response, there was also a component that was in response to all of these pseudo-quantitative claims out there in which they are just trying to make their claims appear scientific and fact-based when in reality they have just pulled a number out of thin air that seemed like a good number to them.
 

WBahn

Joined Mar 31, 2012
29,979
Thus I was told when I first started that, due to the engineers' always wanting to improve the design, as some point in the project, the manger has to shoot all the engineers. :D
The saying that I saw on a poster that captures the same notion is, "There comes a point in the history of every project when it's time to shoot the engineers and begin production."
 

WBahn

Joined Mar 31, 2012
29,979
hi,
Often I get some criticism over my signature motto, all that I can say, it worked for me.

It can be hard to rein in the enthusiasm of your engineering design team, you have to remind them that all the product has to do, is to meet the clients specification and stay within the design cost budget.

E
I make the point even more pointed. I tell them that once the client's specs have been met, that spending a single dime more of the client's money to improve the design is defrauding the client. It's perfectly fine to propose the improvements to the client and try to persuade them that the improvements are worth it, but it is the client, not the engineer, who gets to make the decision.
 

BR-549

Joined Sep 22, 2013
4,928
Then there's were a good team that makes the deadline with all specs.....ready for delivery( group percentages set).....they take off on a well deserved break.......and then the client changes specs.

You have to call them all back in and recalculate the percentages.......because of all the "told you sos".
 

Hymie

Joined Mar 30, 2018
1,277
If I was 30% wrong in my job – within a very short time I would not have a job.

If 50% were to be acceptable – I could just flip a coin.
 
Last edited:

jpanhalt

Joined Jan 18, 2008
11,087
I make the point even more pointed. I tell them that once the client's specs have been met, that spending a single dime more of the client's money to improve the design is defrauding the client. It's perfectly fine to propose the improvements to the client and try to persuade them that the improvements are worth it, but it is the client, not the engineer, who gets to make the decision.
The world has all sorts of clients. Assume your client is a family practitioner in Omaha, and she sends a stool culture for "culture and susceptibility" testing. You know nothing about the patient, such as travel history, symptoms, or confounding variables (like Crohn's disease). How do you know when the "specification" has been met?

That bit of business tripe that cmartinez posted to lead off this discussion deals with business decisions for which there probably is not a right or wrong answer nor a client in the usual sense. There is an outcome that may be good or bad. Mr. Bezos' support for the concept is probably based on that, assuming he was quoted at least 70% correctly. ;)

Unfortunately in application, the C-student MBA's I referred to above will apply it to totally inappropriate situations. One of those situations is meeting a client's expectations. For that, I think the bar is much higher. In my example of the stool culture, the client expects a right answer 100% of the time. I have never had a client who wanted anything less than a correct answer. However, that answer just may not be as precise as it could have been with a lot more work and time.

@cmartinez : I think this is an interesting question, and if I were still working, I can just imagine a management workshop in which that perspective for client relationships is touted. I have been to them. The list of fads is almost endless: Deming, Pareto (80/20), ad nauseam. My comment about tripe applies to those workshops and definitely not to you or your selection of this topic. John
 

strantor

Joined Oct 3, 2010
6,782
Thanks for this. This is what I needed to read today. I've been trying for the past week to settle on components I'm going to spec for a conveyor project. I have already spent 25% more time on it than the estimate I provided to the client. I got to 70% 2 days ago. Then decided to see if I could find better VFDs with more options for lower cost. I did. But they didn't speak the same language as the PLC I selected. So back to the drawing board on the PLC. Found a better & cheaper PLC that spoke the same language as the VFD but didn't meet some of the other system requirements given the other components I had chosen, so I started looking for alternatives to the other components. Done with that, I then realized that in order to be suitable for the application, the new VFD required some add-ons that put the price point above the original VFD, so back to the drawing board on the VFD. It's been back and forth for almost 2 days; I'm wasting time and my indecision is quickly eliminating any cost savings I gained for the client. After I get done here in the doctor's office I'm going home and pull the trigger on the 85% solution I have sitting on the table right now, before I embarrass myself. I've been trying to get to 99% certainty on my spec before I provide the BOM, when I probably won't hit 99% until the last day of commissioning.
 

Thread Starter

cmartinez

Joined Jan 17, 2007
8,220
Thanks for this. This is what I needed to read today. I've been trying for the past week to settle on components I'm going to spec for a conveyor project. I have already spent 25% more time on it than the estimate I provided to the client. I got to 70% 2 days ago. Then decided to see if I could find better VFDs with more options for lower cost. I did. But they didn't speak the same language as the PLC I selected. So back to the drawing board on the PLC. Found a better & cheaper PLC that spoke the same language as the VFD but didn't meet some of the other system requirements given the other components I had chosen, so I started looking for alternatives to the other components. Done with that, I then realized that in order to be suitable for the application, the new VFD required some add-ons that put the price point above the original VFD, so back to the drawing board on the VFD. It's been back and forth for almost 2 days; I'm wasting time and my indecision is quickly eliminating any cost savings I gained for the client. After I get done here in the doctor's office I'm going home and pull the trigger on the 85% solution I have sitting on the table right now, before I embarrass myself. I've been trying to get to 99% certainty on my spec before I provide the BOM, when I probably won't hit 99% until the last day of commissioning.
My only suggestion would be for you to make sure that you've identified and covered all of the critical points of your project. That's what I do, I first catalog things into critical, high importance and low importance. Then I get to work.

Good luck
 

WBahn

Joined Mar 31, 2012
29,979
The world has all sorts of clients. Assume your client is a family practitioner in Omaha, and she sends a stool culture for "culture and susceptibility" testing. You know nothing about the patient, such as travel history, symptoms, or confounding variables (like Crohn's disease). How do you know when the "specification" has been met?
Valid point -- but of course you can only determine if the customer's specification has been met if you have a sufficiently complete specification.

Often times you don't. Almost all of my professional life has been working in environments where not only did the customer not have a complete specification, but the customer wasn't even sure what is was they really needed. But we didn't take it upon ourselves to decide what they needed or how good it needed to be. It was not uncommon for us to have to drag out of the customer what it was they wanted and what was important and the for us to turn that into a specification and then sit down with the customer and go through it in fine detail to determine if it captured what was important and what would be good enough. Even after that, there were almost always aspects of the project where know one really knew what was good enough and, specifically, whether the agreed upon spec really was adequate. So on those points we would push things past the spec if we could do so without risking the schedule, but the customer always knew we were doing this and we had their blessing.

On one occasion, however, we had a customer that specifically told us not to -- we were charging them hourly and they were in a cash-starved situation where anything we spent was something that couldn't be spent on other things related to the project, but that as long as all aspects of the project got far enough, even if it didn't perform as good as hoped, there would be a much larger follow-on project with funds to learn from the actual performance and redesign the chip. This emphatically underscored the position that you are simply not justified taking it upon yourself to go beyond the spec on the customer's dime without their knowledge and approval -- you simply can never know all of the considerations that your customer has to contend with.

Now, there were times that we took it upon ourselves to use our OWN money and time to push the performance beyond the spec for reasons of our own. We might not specifically tell the customer that, but we never tried to hide it and if it happened to be mentioned in a discussion that was perfectly okay.
 

wayneh

Joined Sep 9, 2010
17,496
Almost all of my professional life has been working in environments where not only did the customer not have a complete specification, but the customer wasn't even sure what is was they really needed.
Ditto. When you sell a technical product, part of what the customer buys is “insurance” that you, the supplier, has their back. That’s why there’s such a thing as relationship building. A big part of my job was teaching the basics and enough of the details that my customers could make decent decisions about what products to buy and how to use them optimally. Sometimes they listened, but not always. We nevertheless did our best to ensure they didn’t make stupid decisions. In part, we knew we’d be quick to be blamed if things went south.

This was in biotech, not electronics.
 
Top