< Previous Chapter | First Chapter | Plot Summary | Next Chapter >
Stamper got right to the point: “So, give me the full story.”
“So, first off,” said Keith, “that question was a mistake ...”
“Oh c’mon Keith, you’re not a mistake kind of guy,” said Stamper. Keith avoided mistakes like the plague and Stamper knew it. She liked that about him.
“Yeah, I know, it’s not that kind of mistake. I put that question in the Daily Pulse on purpose. It’s a data point that I think we should track.”
“What’s a data point we should track?”
“Frustration with being tracked.”
“Now wait a minute! I haven’t heard anyone say they were frustrated and I’ve certainly never heard anyone say we’re tracking people,” Stamper asserted with some heat. “We’re trying to create a system where the human experience is optimized, where people only get the direction and feedback they need to be their best.”
“Yeah, and that’s why it’s so important that we gather comprehensive data.”
“We are gathering comprehensive data! Have you seen the reports I give to the S-team ... well, of course you have, you prepare most of the graphs and illustrations!”
“Yeah,” said Keith, “It’s a mad amount of data, but haven’t you noticed what’s missing?”
Keith couldn’t know that insinuating that Stamper had “missed something” was like waving a red flag in front of a bull.
“Oh, so I’m missing something? You know I invented this program, right? I think I’d fucking know if there was something missing. So why don’t you tell me Mr. Wizard!”
Stamper heated up fast. She didn’t like being questioned. And she couldn’t know that confrontation with authority was the last thing Keith wanted. He turned back sin hasty retreat.
“Kate, I’m so sorry, I did not mean to insinuate anything at all. I and everybody else on the Human Dynamics team have the greatest respect for what you’ve created and I’m sure you understand things about it that I have no idea about. Will you please accept my apology?”
“Of course silly,” she smiled, deftly reversing the emotional tone of her previous comment. She cooled down quickly too. “I’m such a hothead!”
“Now,” she continued with a smile, “will you please tell me more about this mistake?”
“Okay, thank you, and again, I’m really sorry, I didn’t mean anything ... ,” Keith would have done anything to start over. “What I meant was it was a mistake to put that question out there without first thoroughly vetting it with you. I just didn’t want to bother you until I had more data.”
“But what kind of data are you looking for with a question like that?” said Stamper, Cascade’s words about not fishing for negativity ringing in her head. “You know that the S-team has made our work a top priority for the 3-year plan and we’re looking to showcase success.”
“I want to showcase success too,” said Keith. “But at the same time, I’m hearing little rumbles of resistance that I want to get a handle on ...”
“Rumbles? Resistance? What the hell are rumbles? I haven’t seen anything in the Pulse rollups or the tap data that speaks to resistance. C’mon Keith, you can do better than this.”
“I know, I know, it’s vague and I don’t like that either! But I’ve looked back through all the Pulse questions we’ve asked, and we’ve never offered a single option for people to ... uh ... question the program. We’ve never asked any questions that would reveal anything but shades of positive acceptance. But there’s something there, I swear it—in little anonymous comments that come in, in the way people talk about taps in meetings, like little jokes and stuff, and very rarely in the free-form answer fields attached to Daily Pulses.”
“Those are anecdotes, Keith, anecdotes! You know we can’t pay attention to that. Whenever there’s something new, somebody is going to grumble about it.”
“I know, and I agree—and that’s why I wrote that question. I wanted to see what people would say to a question like that.”
“And?”
“Not a single person selected that answer. Zero!”
“See! That’s good! It’s because people really like it.”
Keith swallowed hard. How could he tell her what he suspected? How could he say that the fact that NOBODY had selected that answer was itself a statistical anomaly? Well, he had opened this door, mistake or not, he had better step through it.
He didn’t want to push back on Stamper. He wanted to just go along, to keep working with this data, drawing connections, preparing these amazing graphics that he had been building recently, ones that showed how accurately they were able to predict behavioral trends, that by tapping people in advance of their behavior they could curtail that behavior, and that this contributed to overall productivity.... This was such a good story to tell, and it was some of the most satisfying work he had ever done.
But he couldn’t ignore what he was seeing. He couldn’t pretend that the discontinuities in the data correlations he was starting to see didn’t mean something, and he couldn’t only look for correlations under the streetlight. He had to know if the clues he was seeing in the shadows meant anything. He had to know if the ice cream was melting because the freezer was broken.
And with that, Keith started to spell out for her, in as much detail as he could remember, all the various little signals he had been getting about the dark side of the Human Dynamics program: the conversations he had overheard, the cryptic Pulse replies, the care with which people spoke when they learned that he worked with Human Dynamics. This is what the reports had been missing—any evidence that all these taps had unintended consequences, that they might in fact be creating behaviors that were, well, not very Amazonian.
So Keith took a deep breath and started in:
“I’m fine with ignoring a few little complaints here and there,” Keith said, “really I am. And I don’t want to over-index on anecdotes. But given what I’ve been seeing and hearing, my question should have pulled a couple responses. Heck, somebody should have simply clicked it by accident! But zero—zero! That’s just improbable.”
“Maybe it’s because people really don’t feel that way!” said Stamper.
“Let me tell you why that can’t be the case,” said Keith. “There’s this one team where we’ve started to see above average tap rates, and I’ve been keeping my eye on it. It started with this one guy on the team, fairly new to Amazon—in fact, he started here the very week we put the taps into effect.”
“This guy has had above average tap rates from the beginning, but on all his other metrics he scores really high. His productivity score is super high, his manager ratings are good, his teammate ratings are off the charts, all that. So, he’s an interesting anomaly.”
Stamper simply nodded.
“So I start to look into this guy, right, like I pull all his data and all his interactions with Huddy”—these were the logs of any chats that came about as a result of taps—”and he shows just this really bizarre pattern. It’s like he prompts a tap, then pushes at it a little harder, seeing if he can get more taps or prompt an interaction with Huddy. And he pushes back—like he argues with Huddy.”
“So he’s got backbone, that’s good, right?” says Stamper.
“Is it? Or is he supposed to disagree and commit?”
“There’s a balance,” replies Stamper. “Is he becoming a problem?”
“I don’t know!,” said Keith. “Honestly, I’ve jumped onto some of his Huddy interactions myself. I asked to direct some of his human overrides to me, just so I could see what the deal is.” They had built in the capacity to override the AI chatbot they called Huddy, inserting a human into the conversation if it got too complicated for the chatbot (which was pretty rare) or, more often, if some on the Human Dynamics team wanted to get a better sense of the interplay between employees and Huddy.
“You didn’t interact with your own voice did you?” asked Stamper.
“No, of course not, I used the Huddy voice masking,” said Keith; this was the default, but it was possible but highly discouraged to use your own voice; they didn’t really want people to associate the directions with any individual. “But here’s the deal: he didn’t seem belligerent or angry, just skeptical. It’s like he was probing to see what he could get away with.”
“Well listen, we’re not hiring sheep—we should expect questioning, right? After all, our goal is to hire and develop the best,” she said, citing another LP. “Keith, honestly, I’m not sure I’m seeing the issue.”
“Well, here’s the deal: the longer he’s been here, the more the tap rates of his teammates start to increase. I’m seeing a correlation between time with this guy and tap rates. Weird, right? So I’m asking myself, what’s going on? We don’t have any data that tells us that he’s encouraging resistant behavior—like the chat logs don’t show any anti-HD tendencies—but his mere presence seems to be increasing the tap rate on his team. Look, here’s my data,” says Keith, and he pulls up a graph on his phone to show Stamper the trendline data that shows that the tap rate on this guy’s team is 12% higher than average. He then points to some individual Pulse data that shows that people who had low tap rates before coming to work on this team suddenly start to exhibit distinctly higher tap rates once they interact with this guy. And yet, the whole team shows good overall performance. Keith had prepped all this in advance of the conversation.
“Keith, honestly, I can’t make sense of this data looking at your phone! Can you net it out for me?” says Stamper.
“I don’t want to be alarmist, but I think I’m seeing a cell of … uh, god, I don’t know what to call it,” said Keith.
“Try,” urged Stamper.
“Look, I don’t want to be dramatic,” Keith sighed, and then he just committed: ”But hell, let’s call it ‘resistance’ or just maybe ‘reluctance’ toward our work, and it’s centered in a top-performing team. That’s what led me to insert that question into the Daily Pulse. And yet that question gets no response, nothing, not even from this team that I think is showing reluctance. That’s why getting zero responses to that question bugs me. I wonder if they’re mocking us?”
“Well fuck me,” said Kate, using one of her favorite phrases. “Why would anybody reject what we’re doing in Human Dynamics? We’re just trying to make it easier for people to get their work done.”
“I don’t think you should ask me,” says Keith. “I think you should ask the guy who seems to be single-handedly raising the tap rate of everyone he comes in contact with.”
“Fine,” says Stamper, “I was hoping to avoid this but I guess I can’t. Who is it?”
“I was hoping to avoid it too,” said Keith. “I think you know him. His name is Christopher Dourado.”
Stamper didn’t know whether to laugh or cry.
< Previous Chapter | First Chapter | Plot Summary | Next Chapter >
Disclaimer: This is a work of fiction. I’ve made up the story and the characters in it. While certain businesses, places, and events are used to orient the reader in the real world, the characters and actions described are wholly imaginary and any resemblance to reality is purely coincidental.
“What’s a data point we should track?”
“Frustration with being tracked.”
Great line. funny, thanks for morning chuckle,
= .
I like Keith. He's got his finger on the human element. The story is headed in a great direction!