July 10, 2012 at 5:02 pm
GSquared (7/10/2012)
Steve Jones - SSC Editor (7/10/2012)
jay-h (7/10/2012)
SQLRNNR (7/10/2012)
If you can show that there is data to support a decision, it is more likely to be well received than one based on the experience, instinct, or hunches of any person.
I agree with that statement. Even though experience should have significant import, it's the numbers that matter - or so it seems.
But experience can tell you when to trust the numbers, and how much weight to give them.
As long as you look at the numbers. Too many people ignore them
At least partially because they are too often better off ignored.
Weather forecasts for example. . . .
Weather is apparently manipulated, so IMO it is not surprising that predictors based on pre-manipulated weather do not work.
July 12, 2012 at 3:59 am
One of my favourite advocates of an evidence based approach is Ben Goldacre and his Bad Science [/url]blog. He is a doctor, so fairly focused on clinical trials and interpretation thereof, but often covers (UK) government or newspaper foolishness. It's quite a geek-out with extensive coverage of recently for instance Benford's Law and it's application to detecting electoral fraud.
July 12, 2012 at 6:39 am
Revenant (7/10/2012)
GSquared (7/10/2012)
Steve Jones - SSC Editor (7/10/2012)
jay-h (7/10/2012)
SQLRNNR (7/10/2012)
If you can show that there is data to support a decision, it is more likely to be well received than one based on the experience, instinct, or hunches of any person.
I agree with that statement. Even though experience should have significant import, it's the numbers that matter - or so it seems.
But experience can tell you when to trust the numbers, and how much weight to give them.
As long as you look at the numbers. Too many people ignore them
At least partially because they are too often better off ignored.
Weather forecasts for example. . . .
Weather is apparently manipulated, so IMO it is not surprising that predictors based on pre-manipulated weather do not work.
How'd you know weather is manipulted? Fnord You must forget that Fnord :w00t:
- Gus "GSquared", RSVP, OODA, MAP, NMVP, FAQ, SAT, SQL, DNA, RNA, UOI, IOU, AM, PM, AD, BC, BCE, USA, UN, CF, ROFL, LOL, ETC
Property of The Thread
"Nobody knows the age of the human race, but everyone agrees it's old enough to know better." - Anon
July 12, 2012 at 7:55 am
One of the things I found out early on was was that decision makers often ignore data in favor of their pre-existing desires and prejudices.
I created a fairly detailed analysis once that showed that replacing our main VAX server with a new server would save about $200,000 over the first 3 years, and that the net cash flow would be positive from day one. Basically, it would result is a savings in cash from day one, and would generate a large labor performance improvement because the old system was so overloaded that people were spending far more time doing their work than they should be.
I thought this would be a no brainer because the reduction in monthly maintenance cost on the current system would be greater than the lease cost for the new system.
My proposal was immediately bogged down with emotional counter arguments: “Shouldn’t we be moving to UNIX?” “Why don’t we just let end-users do everyting on PCs with spreadsheets?” “This is just IT wasting money that we should be spending in my department”
We eventually went forward with the proposal, but not until there was a major turnover in decision makers.
Since then, I always try to make sure that I understand the emotional context of the decision makers and make sure my proposals fit their prejudices along with having hard financial analysis support. Of course the danger of playing to their emotions is that I may get them to make a financially bad decision, so I try to make sure the data really supports it, and that I am not pushing my own prejudices.
July 13, 2012 at 6:48 am
Michael Valentine Jones (7/12/2012)
One of the things I found out early on was was that decision makers often ignore data in favor of their pre-existing desires and prejudices...
When people want a conclusion they can often find 'hard data' to support it. One by careful 'categorization' of data you can provide evidence that a project will pay off. If there's not enough immediate payoff, keep looking at secondary and tertiary 'benefits' until the numbers come out right.
I remember reading an article some time back about a city who wanted to justify promotion of an 'arts district'. They just included more and more of the area's commerce into the equation (as a benefit of the district) until they got the numbers they wanted. The fact that those benefits might not come from the project, or would come regardless of whether the project was implemented were conveniently ignored.
...
-- FORTRAN manual for Xerox Computers --
July 13, 2012 at 11:03 am
jay-h (7/13/2012)
Michael Valentine Jones (7/12/2012)
One of the things I found out early on was was that decision makers often ignore data in favor of their pre-existing desires and prejudices...When people want a conclusion they can often find 'hard data' to support it. One by careful 'categorization' of data you can provide evidence that a project will pay off. If there's not enough immediate payoff, keep looking at secondary and tertiary 'benefits' until the numbers come out right.
I remember reading an article some time back about a city who wanted to justify promotion of an 'arts district'. They just included more and more of the area's commerce into the equation (as a benefit of the district) until they got the numbers they wanted. The fact that those benefits might not come from the project, or would come regardless of whether the project was implemented were conveniently ignored.
I see this type of "analysis" more often prompted by the marketing boys and their "I need a proof that we are number one in the world in something, whatever it is."
January 20, 2016 at 12:45 am
Whilst gut feel shouldn't be ignored neither should the data. There will be prejudice but we must remember that this prejudice also will have a corresponding knock on effect on the the likelihood of success i.e. if someone wants something to not be true then they will be a blocker, sometimes without even being conscious that they are doing it.
Gaz
-- Stop your grinnin' and drop your linen...they're everywhere!!!
January 20, 2016 at 1:17 am
Whilst gut feel shouldn't be ignored neither should the data.
There's the situation when you have the hard data in front of you and your gut feel is they're wrong. I'm sure everybody has been there when they've got results in front of them and something hasn't felt right. Maybe the figures appear heavily skewed in one direction or higher/lower than expected. In those instances it's worth listening to your gut and double checking because everybody makes mistakes.
How to post a question to get the most help http://www.sqlservercentral.com/articles/Best+Practices/61537
January 20, 2016 at 3:04 am
BWFC (1/20/2016)
...it's worth listening to your gut and double checking because everybody makes mistakes.
Instinct is a great alert system but can generate false positives frequently so BWFC's advice to double check the data makes sense i.e. if in doubt check your findings.
Gaz
-- Stop your grinnin' and drop your linen...they're everywhere!!!
January 20, 2016 at 4:11 am
isn't this just the sort of thing that just might have been going on in well known car manufacturers that led to bad emissions data being published?
January 20, 2016 at 5:47 am
kate.fletcher 80760 (1/20/2016)
isn't this just the sort of thing that just might have been going on in well known car manufacturers that led to bad emissions data being published?
If you mean VW, that wasn't bad data, but bad decisions. They chose to falsify the data to meet emissions standards.
January 20, 2016 at 6:53 am
One reason some managers operate on gut feel vs. data is that is many cases, you don't really have the data you need.
Case #1: Not really understanding your metrics
A call center uses talk time data to give a prize to the agent with the lowest talk time, then fires her a month later because she was hanging up on all her calls instead of handling them.
Case #2: Not gathering metrics correctly
An insurance processing center sets up a metrics team to create a baseline for how much work an underwriter should do in a day; meanwhile the underwriter managers hand-picked their best people and easiest work to create the baseline, because they want to look good. When the metric system rolls out, 99% of their people fall below the baseline.
Case #3: Errors in the system
A timeshare company fires salespeople who do not perform to a certain minimum level.
A code change introduces a bug in the report, and they fire two solid sales people, normally in their top 25% percent.
They try to re-hire them back after the bug is discovered. One refuses, the other demands (and gets) a bonus.
Sometimes the data you have isn't the whole story, and honestly you do want managers that can recognize this and look at the bigger picture.
That's the not the same as the executive who insists Crystal Pepsi is going to catch on any day now.
Knowing when to follow the data and when to look beyond it is a valuable skill, and a skill that will actually become more important the more data we collect.
January 20, 2016 at 2:35 pm
"Gut Feel"? There is a pill for that.
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
January 25, 2016 at 4:21 am
As part of any new reporting project, I try to convince managers and project sponsors to perform some basic data profiling of their data source to understand if reality matches their belief. Often I find strongly held beliefs about the quality of their data are challenged when facing metrics from their data source; but sometimes too, it is vindication of organisational policy and standards that yields good quality data.
This then segues into a conversation about the value of reporting on quality data.
January 25, 2016 at 4:34 pm
I guess "It Depends" on what you call a "gut feel". I was recently asked to provide hard evidence that a particular stored procedure was the source of a performance problem. They refused the "anecdotal" evidence of the (recorded) major blockage during the day occurring (causing the folks on the floor to complain) when that stored procedure is running and nothing else is. They called the plateau of CPU and wild/high disk IO as recorded in perfmon "anecdotal". They were especially firm in calling the many user complaints that occur at that same time "anecdotal". So, my "gut feel" is that they don't have a clue or don't want to spend the time fixing something that needs it.
Someone needs to believe in "Scotty" when he says, "CAPTAIN, THE ENGINES!!!".
--Jeff Moden
Change is inevitable... Change for the better is not.
Viewing 15 posts - 16 through 30 (of 31 total)
You must be logged in to reply to this topic. Login to reply