3 ms·

Email patterns can predict impending doom

23 points by muon 17y ago

CWuestefeld 17y ago
I don't think the failure of Enron is typical of that of other organizations. Enron's very life was founded on cheating, and I assume that at least the senior staff mentioned in the article were aware of this. It seems plausible to me that much of the communications leading up to the actual failure were of a cover-your-butt sort, and not so much the attempts at constructive change that we might hope an organization involved in a more ethical enterprise would experience. Indeed, I'd think that in the unethical cases, where there really is something to hide, communication with the wider corporate community would decrease as desperate scheming between those in on the plot would increase. So I don't think that it's fair to assume that the behavior observed in the Enron case can generalize to other organizations.
- stcredzero 17y ago
  If there are insiders with some inkling of approaching disaster, then it makes sense that traffic analysis can reveal a spike in anxiety and contingency planning by members of an organization. If I were dishonest, then I might try traffic analysis as a means to accomplish insider trading. It's very indirect. None of the information assets that traditionally raise red flags are involved. Unlike traditional insider trading, your confederates/information sources wouldn't have to know what you were up to, specifically. It wouldn't work in every case, but if it worked often enough to give you a considerable statistical advantage, then one could still make a financial killing this way. This makes me worry about LinkedIn and Facebook, as well as Google and Yahoo. It would be relatively easy for employees there to accomplish such traffic analysis undetected, then to correlate this with market information. EDIT: A defensibly legal means of leveraging this. Start a holding company that happens to control the IT of all of its subsidiaries. Structure the subsidiaries to be as independent as possible, otherwise. Make decisions to "invest" in your subsidiaries based on the traffic analysis. (This could also include insuring specific ventures against failure.)
  - rapind 17y ago
    An interesting point. You wouldn't even have to be a direct employee. An immoral IT worker could probably gain access to at least some anonymous info by working somewhere on the pipe (ISP etc.) for organizations that route emails externally (encrypted or not, the contents and origin are not necessary). There's most likely some data mining you could do on http stats too, instead of just email. Spooky.
  - gojomo 17y ago
    Or to exercise the paranoid fantasy muscles: Google has this info, but can't plausibly use it via active investing. So they leak it to giant hedge funds that can. Said funds cause other associated groups to indirectly reward Google with either (1) overpaying for clicks; (2) click fraud which shifts more of the surplus from CPC advertising to Google from advertisers; (3) GOOG-supporting trades. Everybody wins!
iigs 17y ago
1) Sample size of one (or did I misread?) -- Enron. I don't have any guesses on what a wider analysis would reveal, but you can draw all kinds of conclusions from data when you have the data and the backstory and no other data to verify against. 2) I wonder if this data is actionable in any useful way (other than leaving immediately before TSHTF). I think the "Hey Guys, we shouldn't be doing this!" emails probably stay down in the noise, and the corner probably turns with the "OH SHIT OH SHIT OH SHIT!" email. Since no company is going to give access to their anonymised on an ongoing basis, I believe the only people that would have access to this would effectively be blocked by insider trading rules, making acting on this data legally dangerous.

23 points by muon 17y ago

I don't think the failure of Enron is typical of that of other organizations. Enron's very life was founded on cheating, and I assume that at least the senior staff mentioned in the article were aware of this. It seems plausible to me that much of the communications leading up to the actual failure were of a cover-your-butt sort, and not so much the attempts at constructive change that we might hope an organization involved in a more ethical enterprise would experience. Indeed, I'd think that in the unethical cases, where there really is something to hide, communication with the wider corporate community would decrease as desperate scheming between those in on the plot would increase. So I don't think that it's fair to assume that the behavior observed in the Enron case can generalize to other organizations.

stcredzero 17y ago
If there are insiders with some inkling of approaching disaster, then it makes sense that traffic analysis can reveal a spike in anxiety and contingency planning by members of an organization. If I were dishonest, then I might try traffic analysis as a means to accomplish insider trading. It's very indirect. None of the information assets that traditionally raise red flags are involved. Unlike traditional insider trading, your confederates/information sources wouldn't have to know what you were up to, specifically. It wouldn't work in every case, but if it worked often enough to give you a considerable statistical advantage, then one could still make a financial killing this way. This makes me worry about LinkedIn and Facebook, as well as Google and Yahoo. It would be relatively easy for employees there to accomplish such traffic analysis undetected, then to correlate this with market information. EDIT: A defensibly legal means of leveraging this. Start a holding company that happens to control the IT of all of its subsidiaries. Structure the subsidiaries to be as independent as possible, otherwise. Make decisions to "invest" in your subsidiaries based on the traffic analysis. (This could also include insuring specific ventures against failure.)
- rapind 17y ago
  An interesting point. You wouldn't even have to be a direct employee. An immoral IT worker could probably gain access to at least some anonymous info by working somewhere on the pipe (ISP etc.) for organizations that route emails externally (encrypted or not, the contents and origin are not necessary). There's most likely some data mining you could do on http stats too, instead of just email. Spooky.
- gojomo 17y ago
  Or to exercise the paranoid fantasy muscles: Google has this info, but can't plausibly use it via active investing. So they leak it to giant hedge funds that can. Said funds cause other associated groups to indirectly reward Google with either (1) overpaying for clicks; (2) click fraud which shifts more of the surplus from CPC advertising to Google from advertisers; (3) GOOG-supporting trades. Everybody wins!

iigs 17y ago

1) Sample size of one (or did I misread?) -- Enron. I don't have any guesses on what a wider analysis would reveal, but you can draw all kinds of conclusions from data when you have the data and the backstory and no other data to verify against. 2) I wonder if this data is actionable in any useful way (other than leaving immediately before TSHTF). I think the "Hey Guys, we shouldn't be doing this!" emails probably stay down in the noise, and the corner probably turns with the "OH SHIT OH SHIT OH SHIT!" email. Since no company is going to give access to their anonymised on an ongoing basis, I believe the only people that would have access to this would effectively be blocked by insider trading rules, making acting on this data legally dangerous.