(Part 1 of this two-part piece can be read here.)
After the second presidential debate—mere minutes after—I received a ‘robotext’ from Romney’s camp: “Romney Won the Debate!”
My own assessment of the debate aside, it was mildly annoying that I’d have to pay for what I thought was a pointless text. (I would have thought the same text from Obama’s camp pointless, too.) If they’d asked my opinion about the debate, I would have felt differently… and probably would have responded. A survey or political poll is certainly more engaging than a canned message.
But it still wasn’t a personalized text, which takes it up a notch, especially on the creep factor. It didn’t say “Charlie, Romney Won the Debate!” And it’s the ‘personalized’ messages that make people uncomfortable, particularly voice recordings.
But is it really that creepy when you know how they work?
Robocalls come from auto-dialers—automated systems that run through lists of numbers, playing a recorded message or switching the call to a live agent when one of the numbers picks up.
Auto-dialers are designed for cold-calling, a cast-a-wide-net-and-see-what-you-catch approach, like that used by political advocates with the robocalls. They’re limited to simply getting a message across (whether you want it or not).
By contrast, other automated phone systems, like IVRs, are designed for more. In the case of my robotext after the debate, an IVR could have recorded my opinion and added it to a database. That’s primarily what IVRs do—connect callers to information in organizations’ databases (i.e., your bank account balance).
The campaign calls are robocalls, even though some of them are ‘personalized.’ While the use of your first name makes them seem more advanced, there’s a couple simple ways they can do this.
For one, there aren’t that many names in the United States. With a few hundred names, candidates can cover most of the voting public. They could record, say, 500 names (or plug in recordings they’ve already done in previous campaigns) and use those for ‘personalized’ messages. The remaining few would receive non-personalized messages.
Candidates could also use recordings of themselves (for long-time politicians like Romney and Obama, there’d be plenty of recordings) to create a text-to-speech (TTS) app that would build the messages from all the previously recorded audio. The candidate wouldn’t have to record any names at all—the TTS app would build them from audio bits broken off the recordings.
So, in the end, those ‘personalized’ robocalls messages aren’t all that personalized after all. (In other words, you can stop feeling creeped out now.)