The Story of Lenny, the Internet's Favorite Telemarketing Troll

Lenny is a decade-old chatbot designed to troll telemarketers that has developed a cult following online. It’s remarkably convincing, but is it actually effective?

|
Nov 21 2018, 1:00pm

Image: Shutterstock

If you’ve noticed an uptick in the number of telemarketing calls you’ve received in the past few years, it’s not just your imagination. According to the New York Times, automated and scam calls have increased by nearly a third just in the last year to around 3.4 billion monthly calls. The surge in telemarketing prompted the FCC to issue new rules allowing telecom companies to block some robocalls earlier this year. During this congressional session, US lawmakers have met about the issue several times and introduced 12 different bills aimed at curbing telemarketing abuses.

Clearly, these measures have done little to stem the rising tide of telemarketing spam so far. Fed up with the government’s lackluster response to unsolicited calls, a frustrated IT worker decided to take matters into his own hands nearly a decade ago and in the process managed to create the internet’s favorite telemarketing troll.

In 2009, the anonymous IT worker began using an interactive automated message to handle telemarketing calls to the company where he worked. He called the chatbot “Lenny” in homage to his elderly neighbor who used to collect plastic shopping bags in a “giant man-height stockpile” in his backyard.

According to a Reddit post by the person who claims to be the voice and creator of the Lenny chatbot, he sought to create a “telemarketer’s worst nightmare.” This, he decided, would be “a lonely old man who is up for a chat, proud of his family, and can’t focus on the telemarketer’s goal.”

The final result was a chatbot that consists of 16 stock phrases played in order. The first four phrases are scripted so as to encourage the telemarketers to begin their sales pitch and the last 12 phrases are played in a loop until the telemarketer hangs up. Lenny is powered by an interactive voice script, a software program that listens for one-and-a-half second pauses in the conversation so that it knows when to say the next phrase in the loop.

To those in the know, Lenny’s persona is hilarious. He has a thick Australian accent, a bit of a lisp, and talks excruciatingly slow. When a telemarketer calls, users can forward the call to Lenny, who answers the phone and eagerly assents to whatever the telemarketer is calling about. As the conversation progresses, however, Lenny’s responses get increasingly off topic. At one point he begins telling the telemarketer how proud he is of his family and then later he has to get off the line to go silence some ducks that can be heard quacking in the background.

After nearly a decade of existence, Lenny has garnered something of a cult following. Lenny is available on a public server so anyone can forward their telemarketing calls to the chatbot. There’s a dedicated subreddit chronicling Lenny’s interactions with telemarketers, and hundreds of audio recordings of Lenny have been uploaded to YouTube, often attracting hundreds of thousands of views. I spoke with the person currently maintaining Lenny’s server, as well as a researcher who has studied Lenny’s effectiveness as an anti-spam tool to better understand the internet’s obsession with this virtual kindly old man.

THE RISE OF LENNY

Lenny wasn’t the first chatbot to fight telemarketing spam with automated trolling. Its design was inspired by Asty-Crapper, a chatbot created around the same time as Lenny. Many similar systems exist, such as the Jolly Roger Telephone Company, but these are mostly paid subscription services. Lenny was unique insofar as it was the first such chatbot whose recordings were released freely on the internet so that anyone could implement Lenny using software for hosting telephone servers, such as Asterisk.

Lenny was publicly released in 2011 and since then a community of Lenny fans have cropped up around this telemarketing chatbot. These Lenny disciples swap recordings of frustrated telemarketers and scammers who sometimes stay on the line for up to an hour trying to sell this non-existent old man everything from domain names to medical alert systems. Sometimes the telemarketers lose it when they discover Lenny is a recording, but a surprising number of the calls end with the telemarketer coming up with a polite excuse to drop the call.

Today, the publicly-available version of Lenny is maintained by a former programmer who goes by Mango. Mango is not Lenny’s original programmer. He told me that shortly after he learned about Lenny in 2013, the chatbot’s server dropped offline. Not content to let this valuable and hilarious tool disappear, however, Mango used previously published recordings of Lenny’s interactions and wrote a voice detection algorithm of his own so that he could host Lenny on a publicly accessible server for others to use. He’s been Lenny’s sole maintainer ever since.

Mango told me his dedication to Lenny mostly stems from seeing people get scammed far too often. Lenny was a way to fight back.

Read More: We Talked to the Hacker Who Flooded IRS Scammers With Robocalls

“I live in a town with a higher-than-average population of seniors,” Mango told me in an email. “It seems that every other week I hear about another of my friends sending thousands of dollars in iTunes gift cards to someone overseas or paying ‘tech support’ to remove non-existent viruses simply because someone on the phone told them to.”

Mango is emphatic that Lenny should only be used to troll telemarketers and not to trick unsuspecting people. When he first set up the Lenny server in 2013, Mango said they were overrun by prank callers. While he understands the impulse to prank your friends with Lenny, Mango said that these calls ended up tying up the system so that Lenny couldn’t be used against actual scam calls.

After setting up an algorithmic screening system, Mango said the prank calls dropped almost to zero, but the number of people putting Lenny to legitimate use has remained steady. According to Mango, the public Lenny server receives about 300 calls a day.

“Most calls are quite short,” Mango said. “This could be due to several reasons, such as scammers that require you to press a number to speak to an agent or telemarketing calls that are simply dead air. Only 1 to 2 percent of calls are longer than 10 minutes.”

IS LENNY ACTUALLY EFFECTIVE?

Lenny’s popularity is indisputable, but what is less certain is its effectiveness at disrupting the telemarketing ecosystem. Although Lenny’s creator appeared to want nothing more than a way to avoid the hassle of dealing with telemarketers, if Lenny-like systems were deployed at scale it seems like they could drastically decrease the profitability of telemarketing systems.

This was one of the takeaways from research presented last year at the Symposium on Usable Privacy and Security by the French security analysts Merve Sahin, Marc Relieu and Aurélien Francillon. According to Sahin, an expert on telephony fraud, she and her colleagues learned about Lenny while trying to develop a telephone “honeypot” of their own to collect real-world data on telemarketing scams.

“Every minute that Lenny is on the phone with a scammer is a minute that they don’t have to scam an actual confused elderly person"

Instead, Sahin and her colleagues ended up analyzing nearly half of the 500 recordings of Lenny posted to YouTube by Mango. Through this analysis, they discovered that the average length of the telemarketers’ calls with Lenny that have been uploaded to the internet is 9 minutes and 43 seconds. Although several calls approach nearly an hour in length, over three-quarters of calls forwarded to Lenny last only two minutes.

Remarkably, 72 percent of telemarketers studied by Sahin make it through two loops of Lenny’s recording before abandoning the call. Only 11 of the 200 telemarketing calls analyzed by Sahin and her colleagues resulted in the telemarketer realizing they were speaking to a recording. Seven other telemarketers thought Lenny had Alzheimer’s or dementia and tried to get him to put his nurse on the line. “Moreover,” the researchers write, “several spammers aggressively try to interrupt Lenny by shouting phrases like ‘sir please stop’ or ‘listen to me’ or even by clapping hands.”

“Technically, Lenny is a very simple chatbot,” Sahin told me in an email. “However, Lenny is actually a very sophisticated and well-designed chatbot in terms of the conversational quality of the recordings.”

Read More: The History of the Call Center Explains How Customer Service Got So Annoying

Sahin pointed out how Lenny’s recordings are very flexible in the sense that they can be reasonably applied to a number of telemarketing prompts. More importantly, however, Lenny is also able to control the conversation and force the caller to adjust to Lenny’s narrative. This is why so many telemarketers end up staying on the line for ten minutes or more, despite interacting with a looping recording.

Clearly, Lenny is effective a fooling telemarketers, but does this actually put a dent in the profitability of telemarketing?

According to Sahin and her colleagues’ research, automated telemarketing calls cost about four cents a minute, but using human operators can cost up to a dollar a minute. Even when this human labor is moved overseas to call centers in the Philippines or India, telemarketers still pay about 20 cents per minute to call. Although humans telemarketers are more likely to successfully close a telemarketing call, FCC data suggests that two-thirds of telemarketing calls are automated.

Based on the number of Lenny users reported by Mango, Lenny probably isn’t going to single-handedly stop telemarketing. At most, it’s an entertaining way of removing telemarketing calls from your life. Even if Lenny was deployed at scale, this would likely just push telemarketing firms to rely more heavily on robocalling to save money. This would turn our phone lines into a network for bots to talk to other bots, which is what has happened to most of the internet.

Although Sahin doesn’t see Lenny as the solution to telemarketing, she said it was a novel solution to a difficult problem. Mango also doesn’t see chatbots as the end-all solution to unsolicited calls. This, he said, is something that telecommunication providers must handle themselves.

“I think the best solution to the problem would be federal governments forcing the carriers to discontinue service to their problem customers,” Mango said. “Until that happens, stopping scams requires a multifaceted approach: a service like NoMoRobo that selectively blocks suspected unsolicited calls, education campaigns, and bots like Lenny that waste the time of scammers.”

“Every minute that Lenny is on the phone with a scammer is a minute that they don’t have to scam an actual confused elderly person,” Mango added. “I run Lenny because I hope that the volunteers who forward their scam calls and I have, in some small way, saved a few people here and there from becoming a victim.”