Of interest if you shop at Krogers

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

http://www.cincypost.com/business/y2k051199.html

Kroger ranked as laggard for Y2K fixes By Nick Miller, Post staff reporter

When it comes to preparing for the millennium computer bug, Kroger Co. is one of the Y2K laggards among the nation's 100 largest non-financial companies, according to a new report by Weiss Ratings Inc.

That's in stark contrast to the story told by officials at Cincinnati-based Kroger. The operator of supermarkets across the nation has even testified before Congress that its level of readiness should ensure that Jan. 1, 2000 will be just another day of business.

Palm Beach Garden, Fla.-based Weiss also included in its ratings three other companies based in Greater Cincinnati - Ashland Inc., Federated Department Stores and the Procter & Gamble Co. While Kroger received a ''below average'' rating from Weiss for its Y2K readiness, Federated and P&G received ''average.'' The firm said it had insufficient data to rank Ashland.

Weiss said that none of the nation's 50 largest non-financial companies received the ''high'' rating to indicate a substantial level of compliance.

Although it's been known that small and mid-sized companies are struggling with Y2K compliance, that has not been the case for larger companies, Weiss said.

''The poor progress made by so many of America's largest companies comes as quite a shock, implying potentially serious disruptions in the operations and profits of at least some of these companies,'' said Martin Weiss, chairman of Weiss, a company known nationally for its ratings of insurance and financial companies.

Weiss lamented a ''continuing lack of public disclosure'' on Y2K by the nation's largest corporations.

But at Covington-based Ashland, the petroleum and specialty chemical products company has had a task force working on Y2K for four years. It has completed about 95 percent of its readiness program and expects to spend more than $8 million to make its systems operate normally after the new year.

After all, said Stan Lampe, Ashland spokesman, it only makes sense that any company with millions or billions of dollars at stake would address the issue. ''I would be very surprised to find a Fortune 500 company that has not been working on this problem seriously for the last several years,'' said Lampe, in Houston this week discussing Ashland's exhaustive Y2K preparations at a conference.

Ashland has kept most of its systems testing, equipment replacement and other Y2K work as internal as possible. Lampe said that's typical in Ashland's industry because of proprietary technology and formulas.

Weiss officials said much of the company's rating system was based on how much money companies plan to spend in their Y2K budgets and the percentage of those budgets already spent.

Lynn Marmer, Kroger's group vice president of corporate affairs, said no one at Weiss talked with the company about its Y2K program.

Publication date: 05-11-99

-- Arlin H. Adams (ahadams@ix.netcom.com), May 12, 1999

Answers

Percentage of budget spent is a completely unreliable indicator of Y2K readiness. Weiss should be sued for publicly lying with statistics.

-- Polly (skippy@innermongolia.com), May 12, 1999.

More info about Kroger's compliance efforts is at this link:

http://209.208.199.147/Y2K/y2kfind.asp

-- Kevin (mixesmusic@worldnet.att.net), May 12, 1999.


Polly,

What measure of Y2K compliance would you suggest if, as you claim, percentage of budget spent is so unreliable?

Here at NASA, costing against budgeted funds is a key metric for determining progress toward goals. Because we are forced to prepare detailed budgets that show how we will spend our money through the course of a fiscal year (including spending 'spikes' for contract initiation, procurement of key hardware, etc.), it is one of several key measures of success. Not the only one, but a pretty darned important one.

The key to using percentage of planned funds spent as a reliable indicator of progrees is having a good budget in place that takes into account the variation of spending over time. I'd be curious to see what the typical spending profile is for a large software project. Does it start out slow and then ramp up rapidly near the end of the activity? Any takers?

If the nation's largest corporations haven't done a good job planning their Y2K remediation efforts and developing budgets appropriate to their plans, then what they have actually spent on Y2K fixes will NOT be a good indicator of their progress. On the other hand, if they have well-thought-out plans and budgets, I would argue that percentage of budget spent is indeed one of the best indicators of where a specific company stands relative to Y2K remediation efforts.

-- Brian E. Smith (besmith@mail.arc.nasa.gov), May 12, 1999.


NASA is an agency that the Y2K bug can have. It's the biggest waste of taxpayer money next to toilet seats sold to the Pentgon. Could feed the world with money wasted down that hole.

Nothing personal of course Mr. Smith.

-- Anti-NASA (SickofWaste@taxedTOdeath.com), May 12, 1999.


Brian:

I would agree that % of budget spent is an rough indicator of where they are now but it is NOT an indicator of where they are going. One point does not make a trend so any type of snapshot analysis like Weiss and Triaxsys and others are doing is inherently flawed. It also tends to equate companies with similar numbers, wheras a 50% of budget spent number indicates something very different for a company just starting their work versus one who has been working on it for several years. I don't believe there is any single number that you could use to accurately 'rate' the potential for companies to be ready for Y2K or not, which is what they are really shooting for.

I think a better (but still insufficient) number would be % spent versus planned-to-date rather than total budget. That would answer whether, at this time, they were ahead of, behind, or on target but it still would not be a reliable indicator of whether or not they would finish on time. I think the only reason that so many are using this % of total budget spent metric is that it is easy to calculate from publicly disclosed numbers. Any other metric would require a much more indepth analysis of the SEC disclosures and no one wants to waste their time doing it.

-- RMS (rms_200@hotmail.com), May 12, 1999.



Sorry, RMS, but you have at least two dots. Most companies claimed to have started in the 1994-1996 time frame (zero spent). Some in the 1997 timeframe. I suppose you could say that if you gave someone a 20% time assignment to 'find out what this is all about' couldn't you say that you had officially started then???

If a business hasn't completed their project then they haven't done it, money or no money spent. No great resounding chorus of 'success stories' out there yet. It's May 1999. We should be at 95% of the companies completed AND running on their new s/w at this very moment. But it is not so. Typical behavior for this distribution and time scale is that MOST of the companies will not finish on time.

-- David (C.D@I.N), May 12, 1999.


David:

Sorry, RMS, but you have at least two dots.

OK, you have a zero point and another point, that still does not make a trend unless you assume that work is directly proportional to money spent and that money spent is perfectly linear with respect to time. I took the numbers from the Triaxsys report, put them in Excel, and they were not linear. They actually fit an exponential curve very nicely and showed 100% reached in mid 3Q99. Now, I would not stake my life on that prediction but it is just an example of how the same numbers can be extrapolated to very different conclusions, depending on what your assumptions are. While Weiss and Triaxsys never say they assume a linear progression, it is implied. Which is curious because the past expenditureshave not been linear so why would you assume that future expenditures would be?

No great resounding chorus of 'success stories' out there yet. It's May 1999. We should be at 95% of the companies completed AND running on their new s/w at this very moment. But it is not so.

This argument seems to be generally accepted as fact but I have yet to see a reasonable explanation of why. Why should 95% of the companies be doen by now? When you read their SEC disclosures, most have mid to late 1999 completion dates. Why would you create an artificial deadline a year ahead of the actual hard deadline? Most companies have allowed themselves some slack time in their project plans so unplanned activities or problems did not push them past 12/31/99 but I don't know of many that were expected to be 100% compliant (meaning remediation AND testing/validation) by now. That 12/31/98 'deadline' has been often quoted but that did not include testing and validation nor was there a plan for a full 12 months of testing. It sounded good to say, "We'll have all of 1999 in case something goes wrong" but the reality is they don't. Does that mean they will not be done by the real deadline? I don't think that you can draw that conclusion based only on them missing a self imposed artificial deadline but some people would disagree.

To look at it differently, there is a huge thread about the "Y2K prophets" being wrong so far and whether or not you should believe them any more. The general consensus from the non-pollys in that thread is that the 1/1/99, 4/1/99, 4/9/99 etc. dates were not meaningful and the only real test will be 1/1/00. If you accept that argument, then why is it importnat for companies to be 100% complaint so far ahead of the only real test?

-- RMS (rms_200@hotmail.com), May 12, 1999.


OK - lets have an extreme example. Company X starts their Y2K survey two years ago. They budget $100,000,000 bucks for it. They fix their internal software problems for $10,000,000 - and then find an embedded system problem that is huge. They have to replace the entire thing - but it is all located in one place. The cost of the new parts is $80,000,000. The board decides they will order the parts to come in Nov 1 (insuring the order so it WILL be there or someone pays through the nose) and will not pay till it comes in. The plant will shut down Nov 15, they will replace everything and start running again sometime in Feb, when all work is done.

Now look at that from a business viewpoint - they have left two weeks for parts to come in before closing - they have no exposure on cost as the order is insured to come in on time - preplacing the order early guarentees no "parts shortage" - and the money is at work elsewhere until the order actually comes in and is paid for. Perfectly reasonable - have seen many plant closings for maintainence work exactly this way.

From the Y2K pessimistic viewpoint - They won't spent over 10% of their budget till NOVEMBER - DOOM. They won't be compliant until FEBRUARY - DOOM. They can't get parts delivered in November - DOOM.

You have to look at this with a bit of a balanced viewpoint - taking a single point and trying to paint a picture of an entire operation just does not work!

-- Paul Davis (davisp1953@yahoo.com), May 12, 1999.


In the Triaxysis report are data listing average progress for the sample of 674 companies as follows:

date%nbsp%nbsp%nbsp%nbsp percent progress toward completion of Y2K projects

end of Q198%nbsp%nbsp%nbsp%nbsp 27

end of Q398%nbsp%nbsp%nbsp%nbsp 43

end of Q199%nbsp%nbsp%nbsp%nbsp 56

When I plot and extrapolate these data, a linear curve fit shows that the average completion percentage for this group of companies as of this coming New Year will be less than 80%. If I use a 2nd-order polynomial curve fit (one that passes through all three data points), the completion percentage as of Jan. 1, 2000 drops to 70%. While percentage of total budget spent is not a completely reliable prediction method, the trend evident in the data for these large companies is disturbing, especially since the rate of spending seems to be falling off not increasing with time. As the saying goes, "Follow the money!"

-- Brian E. Smith (
besmith@mail.arc.nasa.gov), May 12, 1999.


Let's try that again!

In the Triaxysis report are data listing average progress for the sample of 674 companies as follows:

date      percent progress toward completion of Y2K projects

end of Q198      27

end of Q398      43

end of Q199      56

When I plot and extrapolate these data, a linear curve fit shows that the average completion percentage for this group of companies as of this coming New Year will be less than 80%. If I use a 2nd-order polynomial curve fit (one that passes through all three data points), the completion percentage as of Jan. 1, 2000 drops to 70%. While percentage of total budget spent is not a completely reliable prediction method, the trend evident in the data for these large companies is disturbing, especially since the rate of spending seems to be falling off not increasing with time. As the saying goes, "Follow the money!"

-- Brian E. Smith (besmith@mail.arc.nasa.gov), May 12, 1999.



Dear Anti-NASA, It's a little known fact that the technologies developed by NASA for this country's aeronautics industry (Boeing/McDonnell Douglas, Lockheed, etc.) have led to a range of advanced domestic passenger aircraft that lead all U.S exports in terms of total sales. The U.S. aeronautics industry is the single largest contributor to the positive side of the U.S. balance of payments ledger.

The first "A" in NASA stands for aeronautics (National Aeronautics and Space Administration). It is the contributions made by the aeronautics side of NASA (not the Space side) that have created a positive export cash flow for our country and hundreds of thousands of jobs here. Without the technical expertise of NASA in the aeronautics arena, Boeing and all of its numerous domestic suppliers would not be as competetive in the world market as they are today.

-- Brian E. Smith (besmith@mail.arc.nasa.gov), May 12, 1999.


Don't let him flame you - you're doing fine. NASA ain't too bad either - though I believe many things could be de-engineered to reduce cost & weight, but if reliablility suffers when that happens? (See the recent launch failures....)

What worries me here is that NO Fortunate 500 companies have declared themselves ready - with 500 in the list, you'd figure by now (mid-May) there would be 5%-10% (25 - 50) actually finished just by applying simple bell curves. If NONE are complete, the middle of the bell curve must be assumed to mid-Sept to mid-Oct, and the number of top 500 companies NOT ready perhaps as many 15-25%. Guesses, sure. Does anybody have better numbers?

That's a huge impact - these are much more serious than the mythical 80% compliant assumed by most y2K writers in the past - because they were implying these 20% failures would be in the small & medium enterprises (SME's), not the big boys.

Alos - note that Kroger could have simply publicized their real expenses or percent complete by system, budget, schedule, or any other real value. They didn't - they stonewalled and criticized the messenger. If they (Kroger) really had actual values, they must have had some reason to conceal it.

If, on the other hand, they don't know their real figures, then they can't do much except make the same vague promises that the government does.

-- Robert A. Cook, PE (Kennesaw, GA) (cook.r@csaatl.com), May 12, 1999.


Paul -

your use of reductio ad absurdum is inapropriate here. Please show at least one or two examples of companies which have completed their repairs on half or less of their proposed budget. can't show me two? yeah, I thought so. Now, care to discuss some major corporations which have had to increase their proposed y2k remediation budget, like, oh, say Apple...hmm...what does this tell us? (hint: they're more likely to be pursuing repairs aggressively if they're asking for more money, but not so likely to be anything close to completed if they're still wandering around with half of their budget unspent. I challenge you to provide concrete examples to the contrary.

Arlin

-- Arlin H. Adams (ahadams@ix.netcom.com), May 12, 1999.


Brian...good post.

Anti-NASA. I use equipment first developed for monitoring the astronaut's vital signs and the telemetry for transmission of same. Yes, it probably would have been developed by someone at a later date...but it sure has helped me save lives in the meantime. The other question is how far along would we be if NASA had not stirred imaginations over the years? I would not care to speculate because, just like y2k, all we have are opinions.

-- Lobo (atthelair@yahoo.com), May 13, 1999.


Apple Says Costs Increased to Address Y2K (Yahoo! News/Reuters)

The bulk of this short article -- "Apple Computer Inc. said its estimate of its total external spending associated with Year 2000 issues has increased to $13.2 million compared with $4 million previously, due to increased costs. Apple said in its 10Q filing with the Securities and Exchange Commission that it has so far incurred $6.4 million in incremental external spending, associated with Year 2000 issues, through March 27, 1999. It now estimates that it will spend an additional $6.8 million, primarily as a result of increased costs, to address Y2K compliance of certain Apple branded products. This estimate does not include the costs that could be incurred by the company if one or more of its significant third party vendors fails to achieve Y2K compliance,' Apple said in its filing. As its Y2K plan unfolds, Apple also said the actual future incremental external spending may prove higher. Apple said its Y2K compliant products will recognize the Year 2000 as a leap year and will not produce errors processing date data in connection with the year change from December 31, 1999 to January 1, 2000."

This is an interesting example of how one major firm has under- estimated its Y2K remediation costs by over 330 percent. I'm sure there are other companies that have overestimated their costs, but I haven't seen any concrete examples in the headlines.

If the average company's Y2K spending is lagging their projected budget, finding new problem areas that haven't been planned for will only make matters worse. It will be interesting to see if throwing more money at Apple's Y2K problems is fruitful. Noted experts like Capers Jones have stated that putting additional staff on a late software project usually makes it later.

Made with Macintosh (Y2K compliant, I hope!)

Comments anyone?

-- Brian E. Smith (besmith@mail.arc.nasa.gov), May 13, 1999.



Brian:

Just one comment on your extrapolations. I believe the 56% number was end of 4Q98, not end of 1Q99. It is not real clear but the story says it is based on SEC disclosures made between January and April 1999 so, since it takes anywhere from 30 to 60 days after the end of the reporting period before you see an actual report, I would say that these numbers are only through the end of 1998. Also, the title of the one graph is something like "Status Entering 1999". If you move the 56% number back to 12/31/98 and redo your extrapolation, you will get very different numbers. Also, you may want to reconsider using a 2nd order polynomial. These are cumulative percentages and thus the "function" must be monotonic and increasing so a 2nd order polynomial does not fit.

Robert, how did you arrive at the conclusion that the distribution of Y2K compliance dates would be normal? If the amount of work to be done by each company was the same and they each started at the same time, then I would expect to see a bell shaped curve based on the random variations in how different people solve the same problem. But in this case, I don't see how you can make any assumptions about the distribution function a priori. Here you have companies of widely varying sizes and widely varying degrees of automation and computerization so trying to use sampling statistics does not seem to be valid in my opinion.

-- RMS (rms_200@hotmail.com), May 13, 1999.


Just a side note on Kroger's. They recently bought QFC and Fred Meyers on the west coast. A friend of mine who works for a QFC told me today that at their QFC, the entire computer system (registers and inventory control systems etc) is being replaced. They had been told no vacations in June because system was to be ready for a 2 week training period in June. Yesterday they were told, never mind, the system wouldn't be delivered until September and the training period would not be set up until late August. Wonder if this is indicative of their total remediation effort or just this one store. ? - if this is a $1mil + store, will they be ready in time?

-- Valkyrie (anon@please.net), May 13, 1999.

To the contrary, the large number of companies, their different sizes, degree of automation (and differences in automation between different industries, and between the different branches (divisions) in each Fortune 500 company), and the different start dates assumed for each different Fortune 500 company require the use of statistics to analyze.

These are also large enough that they MUST be automated, if only to manage peopel, processes, material, and money. They are big enough to have automated early - hence are especially vunerable to early y2k errors, early computer systems, and early "loss of source code" problems.

There are, obviously, 500 of these companies. Some (a very few, 2-5%) must finish much earlier than the rest, more will finish a little earlier than the rest (10-15% ?), more yet again (30% ?) will finish "just" ahead of the last 50% (hence the easy conclusion that the first 50% will finish ahead of the last 50%). Some large number will finish "a little later" than average (20-30%), somewhat less (10-15%) will finish more later than average, and some few will either not finish at all or will be extremely late.

Regardless, they will not all finish at the same time. The degree of this "spread" around the average completion time is unknown - with 500 companies of different sizes all starting at different times facing different problems with different levels of resources, I'd assume it will be at least 3-4 months between the "first" to finish and the "average" finsh date. Another 3-4 months between the "average" finish date and last to finish date.

First, can you disagree with this "bell curve" of average completion date? There is no way they will all get done at the same time.

So, this leaves to be determined: how large a spread (std deviation) is to be assumed, and whether the distribution is symetric about the average completion date, and when the "average completion date" will be acheived.

We know absolutely what the population is (500). We know that none are complete yet. We know that all 500 are especially vunerable to year 2000 problems, internal, external, and from infrastructure and overseas problems.

So can we predict the peak (somewhat) as soon as 10% (50) are declared "compliant"? to a certain degree, yes. At least we will know that those 50 are complete. When 250 are compliant, then we can easily determine how many will fail to finish in time.

Well, at least when 50 are compliant, we know that somebody will be compliant, somewhere.

Its mid-may, and none are compliant yet.

-- Robert A. Cook, PE (Kennesaw, GA) (cook.r@csaatl.com), May 13, 1999.


Just a data point, Robert.

Disregarding for the moment what you mean by "compliant", the survey conducted by InformationWeek here, lists between 10 and 20 percent of large companies (> $1 billion) as having completed their Y2k projects.

Granted, the survey was not all encompassing of the Fortune 500. But I think you are wrong in the assumption no one is done yet.

-- Hoffmeister (hoff_meister@my-dejanews.com), May 13, 1999.


I'm still not sure why you insist the distribution must me normal. I see no justification that for that assumption. Obviously they will not all finish at the same time but you get very well have a bi-modal or tri-modal distribution or any other type.

Your argument that some "must finish much earlier than the rest" and some "will either not finish at all or will be extremely late" is only valid if the distribution is normal and I see no justification for that.

First, can you disagree with this "bell curve" of average completion date? There is no way they will all get done at the same time.

Are you saying that the fact that they will not all get done at the same time means that the distribution will be a bell curve? Can you provide a little more detail? There seems to be a few steps missing there.

Statistics are useful to make predictions when you are confident that your sample is representative of the population. Since there are no samples, how can you make statements like "When 250 are compliant, then we can easily determine how many will fail to finish in time"? There are a whole lot of shaky assumptions built into that inference.

-- RMS (rms_200@hotmail.com), May 13, 1999.


Dear ANTI-NASA,

The last time I checked, for every dollar that went in to the program, we got back (in spinoffs, etc) that dollar, PLUS another 80 some cents (see "Expanded Universes of RAH, testimony to Congress, Robert A. Heinlein" for the figure in the 1980's.) I suspect that it has risen considerably, but even if not, this is the ONLY FEDERAL PROGRAM to give back more than it gets.

Chuck, who hasn't seen (with one notable exception) a NASA event he didn't like.

-- chuck, a Night Driver (rienzoo@en.com), May 16, 1999.


Don't get too hooked up on your high school statistics book - if "bell curve" scares you as being too "exact" a curve, try:

Earlier Than Average, Average, and Later Than Average.

The classic "bell curve" is valid for almost all natural distributions of something about a mean (average) value. This will skewed by several factors: accidental inccurate reporting (the company reporting compliant, but didn't know it missed something or some critical system), deliberate inaccurate reporting (the company declared itself compliant to meet an artificial deadline, knowing it was not yet ready), and finally the absolute deadline of the date itself - thus, there is a tendency to "push" the schedule as close to 12/31/1999 as possible to allow as much schedule time as possible.

Bi-model, tri-model - sure, they could occur. Especially if artifically induced by several deadlines from the bureacrats in one or more heavily regulated industries. But the Fortune 500 is a very disverse group, and even if a section is strongly regulated, every corporation is so widespread it iw itself be faced with a bell curve od divisions getting ready: some early, most ontime, some behind the deadline.

If 10 are ready (compliant) now in mid-May, then when 250 have declared themselves compliant - forget verification or any else, just declare themselves compliant as a first step - then we know what the "first half" of the bell curve looks like. Assume 250 have declared themselves compliant by end of July (10 weeks from mid May): then in an additional 10 weeks (mid-Oct), we can expect the remainder to declare themselves ready.

Okay - that's not too bad. Mid Oct leaves a couple of weeks to continue ironing out problems, maybe this won't be too bad.

Assume though that 250 don't declare themselves compliant until mid Sept (18 weeks from now), if the distribution of those self-declared compliancy dates shows that the std deviation is large (the dates are spread out - now all bunched up at the end of Aug, then that definitely shows signs of trouble. 18 weeks from mid-Sept is Feb, 2000.

That would indicate that 10-20% of the Fortune 500 will not be ready, and thus economic disruption from internal failures much greater in small and medium businesses that didn't prepare, or that rely on the biggies for their existance, and in other big businesses than do business with the failed large companies.

Above all, don't let declared compiancy and self-reported dates lull you - the only absolute requirement is an ability to continue making a profit in 2000, regardless of false declared compliancy, undeclared true compliancy, and unforseen non-compliancy.

Add in a measure of just plain stupidity too.

-- Robert A. Cook, PE (Kennesaw, GA) (cook.r@csaatl.com), May 16, 1999.


Brian,

OT: I'm a big NASA fan. FWIW (little).

-- Drew Parkhill/CBN News (y2k@cbn.org), May 17, 1999.


Moderation questions? read the FAQ