Can’t Believe I Just Sent That

How to prevent embarrassing, misfired e-mails.

Last week, Google announced what one of its managers referred to as a “Gmail embarrassment reduction pack.” The Web app will now give you a window of five seconds to cancel a sent e-mail. The “undo send” feature joins the little-used but funny “mail goggles,” which require you to answer math questions before sending late-night mail on the weekends, and the very useful “forgotten attachment detector.” The final element of the embarrassment-reduction pack is the built-in Gmail feature that signals when the message thread that you’re reading has been updated. This theoretically prevents you from being the fourth person to drop the same witticism in an officewide e-mail jokefest.

Google, busy organizing all of the world’s knowledge, has apparently given itself a more difficult mission: saving us from our idiot selves. One of the leaders of this mission is Gmail product manager Keith Coleman, who explained the thinking behind “undo send”: “When we looked at the mistakes, we realized they were being discovered right after send. The user would see the wrong guy in the ‘To’ field and rush to close the browser window.” Sending an e-mail, he said, has the paradoxical effect of clarifying what you actually meant to say: “When you hit send and the thing becomes real, that triggers people to realize what they were really thinking.” In our haste, we often don’t know what’s in an e-mail until it’s gone.

Is five seconds really enough? You can catch obvious mistakes, but, in my experience, the “I probably shouldn’t have written that” regret sets in around the two-minute mark. Coleman says other time options will be added, but more along the lines of 10 seconds than 10 minutes. The problem with longer hold times is that people don’t like their e-mail to be “slow.” While “undo send” is undeniably useful, it’s not the ultimate face-saving tool. That would be “e-mail recall” (without anyone knowing that you recalled it). Let’s say you accidently forward the details of your boring law internship at Skadden to half the partners in the firm. With “undo send,” you have five seconds. The magical “recall” button would let you extract a misfired e-mail out of anyone’s inbox at any time. Coleman said that this was impossible to implement for mail sent to a non-Gmail address but that it could potentially be done for Gmail-to-Gmail communication, provided the recipients haven’t read the e-mail yet. But don’t expect this anytime, ever, as it would create usability and server issues that are “nontrivial.” Not to mention the creepy, stalker-standing-outside-the house vibe of taking a message out of someone’s inbox rather than keeping it in a holding pattern in your own.

As we strive to answer more and more e-mail—and get annoyed with it—the blunders increase. Reply-all gaffes are legendary and sometimes just legend. In January, the State Department’s e-mail system was brought down by excessive reply-alling, and this MetaFilter thread has a great but unverifiable story of how a five-word e-mail—”Free bananas in the kitchen!!!”—crippled a global company’s servers for several days. (Each division would wake up and start complaining about getting the stupid/pointless e-mail.) Despite these examples, we seem to be getting better at reply-all, as an etiquette of replying only to the sender is catching on (though it would be a good idea to make the reply-all button the size of an M&M). To me, the most potent threat at the moment is the auto-completing e-mail address.

Auto-complete is insidious because it’s just helpful enough. You don’t have to remember anyone’s e-mail address, and it would be tedious to disable the feature and go back to the old days. But it’s so easy to type a few letters, hit return, and ruin your year. Just ask Steve Shanks, an athletic director at a Catholic High school in Iowa who complained about the “[l]ong-ass singing and a long-ass homily” of one of the priests and slighted the girls basketball team in an e-mail that was meant for his brother Joe Shanks, but instead was sent to Joe Katich, a coach he had fired. Or, to give the most prominent example, the lawyer for Eli Lilly who wanted to e-mail her co-counsel Bradford Berenson details of a negotiation but instead sent them to Alex Berenson, a reporter for the New York Times. The result was a front-page scoop revealing that Eli Lilly was talking with the government about a billion-dollar fine for improperly marketing its anti-psychotic drug Zyprexa. (And so much for the effectiveness of those lengthy legal disclaimers at the bottom of corporate e-mails.) *

Despite any gaffe-proofing of our e-mail systems, the problem remains that e-mail is too disembodied and abstract. Unless you tweak things, a message from your boss looks the same as one from your fantasy baseball league commissioner. You don’t feel the presence of a thousand people when your mouse hovers over the reply-all button. The next breakthrough will be better visualization. The auto-complete problem could be combated by making e-mail into “facemail” that would show photos of any recipients in the compose window. The commercial version of Gmail already has a feature where the “To” field turns orange if one of the recipients has an e-mail address that is not in your domain. The most charming visualization that I’ve seen is one that turns e-mails into little pulsating microbes. (I especially like how old messages become hairless and inert.) Someday soon, a design and programming team will find a workable, intuitive way to detect which messages may be private and confidential and give them heft and gravity. In the meantime, type carefully.

*Note: I can’t confirm that auto-complete was the culprit in these cases, but it seems highly likely given the name similarities. (Return to the text.)

P.S. If you have any great e-mail-gaffe stories, please e-mail me. I find them addictive.