- October 17, 2008
- Posted by: admin
- Categories: Blog, Business Dynamics
A reformatting error in an Excel spreadsheet has cropped up in the largest bankruptcy case in U.S. history, prompting a legal motion by Barclays Capital Inc. to amend its deal to buy some of the assets of Lehman Brothers Holdings Inc.
The firm — Cleary Gottlieb Steen & Hamilton LLP — said in the motion that one of its first-year law associates had unknowingly added the contracts when reformatting a spreadsheet in Excel.
At the time when the error occurred, Cleary Gottlieb was working under a tight deadline put in place by the bankruptcy court to submit the purchase offer from Barclays, according to the motion.
The mistake was discovered Oct. 1 and was first reported by Above the Law, a legal news Web site.
A hearing on the motion has been scheduled for Nov. 5. Spreadsheets have long been one of the most popular ways for corporate users to store and analyze data.
In recent years analytics tools have improved upon the typical spreadsheets. But over the past few years, they have played an increasing role in data breaches because workers are apt to store them unsecured on laptops. In addition, hackers have actively tried to exploit vulnerabilities in Excel.
In the above scenario who do you think is responsible, the software application or the user?
If I were on a jury I would be hard pressed to vote that the user did not have responsibility to review the results of the application to ensure they were reasonable. That would be before I knew the application was part of Microsoft’s Office suite.
Bill Rinko-Gay
Bill Rinko-Gay
Delivery Director at Technisource
Absolutely and without question the responsibility for the error lies in the hands of those responsible for creating, presenting, receiving, and accepting the electronic data (an Excel file in this case).
Blaming bad programming is just an excuse for some level of operator/user error in the process of turning over the data from one company to the other. I highly doubt this “error” is actually a defect in the Microsoft Excel software itself.
The firm is responsible for the error.
Links:
* http://www.brickhouse.com
If the software performed as expected (no mathematical errors or miscalculations), then the fault lies entirely with the user. What truly surprises me is that a company as dependant on calculations and financial modeling as Barclays Capital is making decisions based on user-generated formulae that are subject to an ungodly number of errors. Mis-copying a cell or formula or referencing the wrong cell could, and in this case actually DID, result in a catastrophic issue.
As a programmer and consultant, I look for these things as case studies for why a company needs a controlled piece of software modeled after their business processes and practices to better ensure that these types of things do not happen.
Do we have more data? I mean if it was Excel error for sure or the end user didn’t use some of the feature which would have avoided this.
At the level of the courts, the companies are responsible for the data they provide.
The company in turn can blame the software contractor or the software vendor who caused the error and they can seek recovery in the courts. This is why a software contractor is wise to maintain a business insurance policy to cover them against these contingencies. Often the client demands it.
1. Caveat emptor
2. Barclays should sue their legal representatives for failing to spot the inclusion of ‘assets’ not on previous schedule
3. Haste and greed are dangerous bedfellows.
4. Spreadsheets are intrinsically dangerous. I’ll bet half of the accounting for sub-prime exposures, CDS was done in Excel. Lack of audit trail makes the use of such tools a bad choice for this type of work.
5. The application behaved perfectly. My favorite way of closing such a ticket/defect log is with the line “this behavior is by design”. The tool is not at fault.
Here’s the link to the submission for anybody interested: http://abovethelaw.com/Barclays%20Relief%20Motion.pdf
So your evidence is that “said in the motion that one of its first-year law associates had unknowingly added the contracts when reformatting a spreadsheet in Excel. ”
So you trusted the health of a 100+yo company and billions of assets to first year law associates?
Excel is the defacto-standard spreadsheet software in the accounting professions and I can attest that any CPA knows more about how to use Excel than 99.9999% of computing specialists, including formatting files. Accounting offices regularly have to work with files formatted in all sorts of crazy differences and this should be cake for them to figure out.
Reading between the lines if you brought this up to a non-accountant and non-computing specialist in court they may be able to be fooled by it. I firmly believe this is a contrived excuse.
A bad worker always blames their tools.
It has been so throughout history.
Anyone who relies on a piece of software to provide exactly what they intended, without doing any checks and balances, is a fool, and should suffer the consequences.
The major problem with tools like Excel is that they are so easy to use that anyone can appear to perform complex analysis. It allows a single person in isolation to deliver major analysis where previously a team of people would work together on creating the results. Perhaps the fault is with the court for imposing deadlines that are impossible to meet unless you try to use shortcuts by employing Excel, but that does not make the software a cause or complicit in the fact that a human error ultimately was the core problem. The fact that people rush to deliver on unreasonable task deadlines is a major reason why (1) errors are made and (2) there is no opportunity to detect these errors in time to (3) prevent a disastrous result. That is why there have to be checks and balances for all projects.
The First Year Law Student is to blame; period. I’ve been in IT for nearly 30 years; it’s always the fault of the nut behind the keyboard. And “said nut” will always try to blame the program.
My big push this year with all the departments and individuals I support is a simple message; “You have to own it”. Technology doesn’t own the business, the business owns the technology. “Owning” the technology means that the business has to accept responsibility for the program, its usage, it’s maintenance, and it’s results. We should be mindful that we can delegate “tasks” to computers but not relegate authority to computers. Its up to every manager and every department to insure that people know how to use their technology correctly.
Technology is a “force multiplier” in that it has the ability to do tasks basically at a higher rate than we can. But, it can’t “think” for us. That’s often the problem. One of my favorite quotes is that if you believe in artificial intelligence, then you MUST believe in artificial stupidity. If you are doing something incorrectly, not paying attention, or are unfamiliar with what you are doing; all technology is going to do for you is make you faster and more efficient at being wrong. Applications like Excel and Word have “Spell-Check”, but they haven’t perfected “Intelligence-Check” yet.
You have to own your technology, or it’s going to own you. And that stupidity is going to be far more costly than the investment you should have made in educating your staff how to use their technology in the first place.
You have to own it.
USER ERROR … problems with the use of EXCEL are well documented in the press. It is not the first time excel has been blamed as the reason why a contract has gone wrong … and i bet it won’t be the last time.
If it’s important it’s always best to check.
In this instance if anyone should be blamed … how about looking at the managers of the “first-year law associate” …
Ah excel the bane of modern computing, just read my blog.
Links:
http://angelos-proverbs.blogspot.com/2008/08/antiproductivits-excellsiorii.html
It’s always up to the user to QA their work. As I understand this issue, rows that had been “hidden” in the workbook were un-hidden when it was converted to PDF to be submitted to teh court.
That said, my lawyer friends tell me that Barclays has a strong argument and precedent on their side in undoing this mistake.
Can I use this excuse to get out of my mortgage?
No blame to Excel, its up to the user..
They should have taken my services on this, how to build reports on excel…
Check you work before you sign your name. Sorry, Barclays.
I’m a bit flabberghasted at your question.
Why on earth would Excel (or Microsoft) be responsible because of a “reformatting error”???
Reformatting errors are user produced errors. I could do the same thing in OpenOffice or any other spreadsheet program.
That’s a USER error. One which, I would assume that someone of their stature at Barclays should not make, and they certainly should have checked their information before hand.
Software cannot keep you from being an idiot. No matter how brilliant you are no matter what measures or what you do – you cannot outwit stupidity.
without going into detailed analysis, its nothing to do with the software – surely the problem exists between the chair and keyboard and indeed with the governance and control for audit and checking to make sure this was correct. After all, we are not talking about something insignificant here…
I’ll add my voice to the chorus which says that “this is a user error” – but it’s worth making a further point.
Ever since spreadsheets started, people have a strong tendency to believe numbers that they see in a document which is nicely printed out in a proper typeface. I think it’s because of a long (pre-computers) history of textbook publishing, where setting up a print run was expensive and therefore copy was subject to stringent checking and editing.
Spreadsheets are particularly dangerous in that they allow users to produce results that look highly professional but have had minimal checking.
To my mind, the fault lies with the management of the company, which has permitted major financial decisions to be based on a spreadsheet while omitting to set up (or at any rate to enforce) any robust system of cross-checking.
The correct way of managing this situation is to either implement a system where the checks are built-in (probably not a spreadsheet, but it could be done by add-ins), or to have a highly robust manual checking procedure.
The company is to blame – it’s a PROCESS error and more onus should be placed on companies not to manage critical calculations in Excel.
It really is scary how much of the world runs on MS Excel and it’s pretty well proven the average spreadsheet contains at least 4/5 major errors.
Is it a statistical bug in calculation by Excel product or is it a bug because of the usser. In both cases the product will have intended use and the disclaimers as part of licensing agreement will protect them. In the second case the user incompetence, lack of knowledge, oversight any thing could been the reason for the failure.
oftware is a means, and therefore not accountable. People are.
“User error” is an interesting term, though. People make mistakes.
What care was taken to allow someone to properly check his/her own work? Can the size and complexity of the work be considered within reason?
In the end: people do make stupid mistakes, but sometimes the chaotic approach is condusive to them making mistakes, Excel (& other spreadsheets) are widely used to do mainly manual work, importing from various sources, manually creating formulae etc. This often goes wrong. People live with that.
If they worked the same way in MS-Access, or other means they might make the same mistakes.
The Excel thing has been well covered…. when will come the day the banks take responsibility for anything they do wrong. No sympathy whatsoever. As with most responders.
Caveat Emptor
Take more care
Accept that your rush to conclude affected the accuracy of your work
Ali,
I still think the old saying about computing is true:
S*** in, S*** out.
The ultimate responsibility lies with the user. Relying on the technology without double checking is gross neglect.
Imagine if airline staff would blindly rely on technology and not check if it does what it is supposed to do.
Management responsibility today includes building high reliability organizations, and in this case, management failed.
You’d think that with countless billions of dollars to work with they could have afforded some more sophisticated software…
Both.
The math in Excel has been known to be flawed for quite some time. So there is some culpability on the part of the user.
Microsoft definitely. The flaws have been known and the primary function of the spreadsheet is to provide accurate calculations.
Anyone who uses Excel for their analytics should be shot. Again the flaws have been known and there are enough software options out on the market to make Excel obsolete.
We can’t blame Excel….else we will have to admit its possible for a human(creators) to become flawless !!. At the same time we can’t blame the user as well…..if the software has behaved in expected way 99 times..a law associate can’t be blamed if it behaved unexpectedly the 100 time. Is it a problem with the process !!…..but again we all know critical defects have come up inspite of good processes in place(a bug in the process), apart from the fact they cost a good amount of unproductivity !!. So in the end Titanic sank, bridges have fallen and it is just one of the things one has to live with.
Microsoft cannot be held liable or accountable for any errors or omissions in their products. It says so in the licensing agreement. If the user was not aware of the licensing agreement, then the company could be held liable.
Ooops…!
Sounds like a PEBCAK. Problem exists between chair and keyboard.
Its the blame someone else mentality.
Links:
* http://en.wikipedia.org/wiki/PEBKAC
From the Puget Sound Cloud Snake…”Computers Don’t Make Mistakes. People Do. –Regardless of News Reports to the Contrary”
American Express recently published a letter to many of its card holders blaming the need to adjust lines of credit on a “computer error.” After calling this to the attention of the VP in charge, they identified a “programmer error” and offered a correction–to the letter. I offered to review their IT practicies…no response so far.
Today’s software generates hundreds of thousands of lines of code in microseconds that might all be useless. Mouse clicks replace neuron firings routinely. Outsourcing often replaces internal compenence.
The court can correct the “error.” Should they? Might the “error” have been intentional? Follow the career of the scape goat to find out. When rocket scientists and structural engineers make mistakes, people die. Where’s the justice in that? Maybe the legal community could use a lesson or two in QA from industries that can’t just say “oops.”
Ultimately the user is responsible for the data validity, the software used is only a tool.
The data and analysis need to be quality assured before being used.
Both the application and the users are responsible: the application for misrepresenting the underlying data; the users for using Excel in the first place. Note that when I state that the application is responsible, I don’t mean that Microsoft are responsible, since the end-user license agreement, obnoxious and partly unenforceable as it may be (if scrutinized by the courts), more or less blocks any legal avenue to Redmond. And by “users” and not “user”, I refer to the group of people whose job it was to prepare the body of work in question, not the person pressing the buttons who probably looks like a convenient scapegoat at this point.
It is continually shocking but not surprising that Excel – or any spreadsheet – gets rolled out for analysis tasks in a wide variety of pursuits, especially since it has the tendency to guess what kind of data it is seeing, formatting it as it sees fit. For example, a bioinformatician (to take a completely different domain) might get a CSV file from a collaborator containing a list of gene names and related information, amongst them the name “DEC1”. Unfortunately, upon importing the data, that name might well be auto-converted to “12/01/08” (this is what OpenOffice.org 2.3 does), which would be a weird name that might raise suspicion, but not if all sorts of other “help” were forthcoming from the application. The only way to discover what “help” has been administered is to inspect the original data, leading to the observation that another way of handling the data is surely more appropriate.
As others have pointed out, there are process issues when a serious transaction is prepared and the details not adequately checked. Such processes need to be supported by proper tools and treated like engineering problems. Arguably, the fault for the mistake lies with those who believed that every tabular or number-related task means reaching for a spreadsheet application, or perhaps reaching *only* for such an application and then not verifying their work another way afterwards.
The fault lies neither with the application nor the user.
It lies with the manager that entrusted the task to a very junior colleague in a case where inaccuracies in the result would constitute more than temporary irritation and then didn’t double check the final numbers.
I spend a lot of time extracting data from ERP systems to display in CRM systems or for reporting and always sample the data and do the calculations by hand first to ensure that my formulae give the expected result. The problem with Excel and most software is GIGO, garbage in garbage out.
The case in hand is what in Danish is called fejl 40 – the error is 40cm from the screen.
In the above as with any calculation, you need to be able to rough guess a result to ensure that what comes out the other end is not total gibberish.
Try it yourself:- 2100 * 37 + 3456 = ?
If your guess was nowhere near then always have someone check your figures.
Why are people using excel for this? The problem way immense when that decision was made.
The user / receiver is responsible. When companies are dealing with billions of dollars in assets, don’t be cheap. Get your data verified and re-verified maybe verified again. Make sure their is a full process life cycle when dealing with any aspect of data.