Not too long ago, 62 employees at a major consultancy found themselves getting called into a room in pairs, neither person having any prior relationship to the other, for what they were told was a role-playing exercise. Researchers asked them to sit across from each other. Participants then learned they weren’t assigned to be collaborators, but adversaries — opposing sides engaging in a mock negotiation to buy or sell a biotechnology plant. They had six minutes to haggle over the price, and heart-rate monitors would track the ups and downs of the argument.

When the negotiations were finished, each side gave feedback about his or her opponent’s performance. Some participants were told to give the feedback unprompted. Others were instructed to ask for feedback. Quietly, the heart-rate monitors listened.

Here’s what the researchers found: If you want to put people on edge, tell them they will receive some feedback. But if you want them to send them over the edge, tell them they’ll be giving feedback. Subjects in the study felt far more anxious about offering feedback than receiving it, which might explain why so much workplace feedback — particularly in the United States — amounts to a series of polite statements, with few suggestions for improvement.

“There’s a strong culture of being very nice to people, and it’s hard to be critical of someone in those conditions,” says Tessa West, the New York University psychologist and NeuroLeadership Institute senior scientist who led the study, alongside researcher Katherine Thorson, also of NYU. “How good is the feedback going to be if the person feels this strong normative pressure to be nice during the interaction? It might just be overly nice and not constructive because they feel weird about the feedback experience.”

Simple as it may seem, feedback — that ubiquitous necessity of organizational life — has proven to be an axis on which organizational culture turns. Research is suggesting that by switching from giving feedback to asking for it, organizations can tilt their culture toward continuous improvement; smarter decision making; and stronger, more resilient teams that can adapt as needed.

Why Feedback Matters

Feedback isn’t just a ritual of the modern workplace. It’s the means by which organisms, across a variety of life-forms and time periods, have adapted to survive. To University of Sheffield cognitive scientist Tom Stafford, feedback is the essence of intelligence. “Thanks to feedback we can become more than simple programs with simple reflexes, and develop more complex responses to the environment,” he writes. “Feedback allows animals like us to follow a purpose.”

It’s no coincidence the words organism and organization share a Latin root. Just as feedback enables the former to flourish, so it does for the latter. The single-celled amoeba that relies on feedback from its marine environment can more easily find bacteria to munch on, and the salesman who risks losing his job owing to missed targets — metrics, too, are a form of feedback — knows he must change his approach, finding better leads or making more of the customers he has. The same is true for the underperforming department that faces restructuring and rethinks how it collaborates. In all cases, feedback is what keeps organisms, and organizations, alive and well.

Even within organizations, feedback can take many forms. Key performance indicators (KPIs) and other quantitative data are perhaps the most recognizable kind of feedback, especially during performance reviews, but conversational feedback — for example, a quick chat over coffee — counts too. Indeed, just as leaders should think carefully about the KPIs that guide behavior on their teams, they should consider the patterns of verbal feedback that guide their teams to improve.

Research has found roughly 87 percent of employees want to “be developed” in their job, but only a third report actually receiving the feedback they need to engage and improve. The reason for the gap is hardly a mystery: Typical feedback conversations are about as pleasant as a root canal. Managers dread them because it’s often unclear what kind of feedback the employee wants or needs, and employees dread them because even light criticism can feel like an assault on their status and credibility. Indeed, West and Thorson’s new study found that receivers’ heart rates jumped enough to indicate moderate or extreme duress in unprompted feedback situations.

Management gurus have devised a range of tactics to repair these broken interactions. Mostly, they restructure how feedback is given, and apply little thought to what the research literature advises. Under the popular sandwich model, a manager carefully slips a criticism in between two compliments, hoping not to threaten the employee while still offering guidance. Other variations include the start, stop, continue method, which encourages employees to start doing one set of behaviors, stop doing another, and continue doing a third. In our own research of 35 such models, no organization was confident its feedback model was effective at creating lasting behavior change. And tellingly, only one gave tools to the feedback giver, not the receiver.

Of course, some organizations forgo these methods entirely, opting to spend large sums of money — US$1,273 per employee, by some estimates — to build and deploy learning initiatives that seek to improve behavior and performance en masse. These initiatives stand partially, if not entirely, in place of feedback conversations because organizations assume it’s easier to get everyone up to speed at once than to let it happen organically.

A growing body of research argues against all of these approaches. This research compels organizations to heed the wisdom of West and Thorson’s negotiation study: Developing a culture of asking for feedback may be the most cost-effective way to develop healthy, ever-evolving work cultures.

Mock Negotiations, Real Insights

After the participants in West and Thorson’s study finished their negotiations, each person was randomly assigned to one of two conditions: Either they would ask for feedback or just give it outright. Afterward, the researchers asked people how they felt during the interaction. Did they feel anxious? How difficult was hearing feedback about their negotiation skills?

When the investigators analyzed the data, they found a curious effect among the people who gave feedback unsolicited: They were rated as being much friendlier than those who were asked to give it. Not only that, the feedback itself was judged to be more positive. It was only when West and Thorson looked at the givers’ heart-rate reactivity and saw it was jumping around erratically that they deduced people were actually terribly anxious during the interaction.

“They’re looking really friendly,” West says, “but they’re feeling really uncomfortable.”

Psychologists have come to label this phenomenon “brittle smiles.” It happens when people try to adhere to a “culture of niceness,” as West calls it, even though they really want to speak or act more candidly and critically. So they overcompensate. They smile too much and become overly positive in their speech.

To West’s mind, asking for feedback is the best way to avoid brittle smiles and the culture of niceness. “When you ask for feedback, you’re licensing people to be critical of you,” she says. “It may feel a little more uncomfortable, but you’re going to get honest, more constructive feedback.”

This permission, it turns out, is hugely important for putting both parties in a psychological state that’s ready for negative news. Without it, the brain begins to revert to a state that isn’t conducive to growth, and that finds its roots thousands of years in the past.

The Science of Why Feedback Is So Miserable

Though most of us no longer have to fend off predators, our brains are still exquisitely attuned to threats — both physical and social. It’s a vestige of how survival has largely depended on appeasing group members. Among our ancestors, eviction from the group led to a dangerous, isolated existence in the wild.

Modern humans base their decisions on many of the same pro-social, consensus-building impulses. We make polite chitchat at work, even in our most antisocial states, so others will see us as friendly. We avoid talking to the attractive stranger at the bar because something deep and ancient in us registers the possibility of rejection as a matter of life and death. When neuroscientists conduct brain scans of people exposed to social threats, such as a nasty look or gesture, the resulting images look just like the scans of people exposed to physical threats. Our bodies react in much the same ways. Our faces flush, our hearts race, and our brains shut down. No matter if we’re giving a speech to thousands or coming face-to-face with a jungle cat, our body’s response is the same: We want out.

Feedback conversations, as they exist today, activate this social threat response. In West and Thorson’s study, participants’ heart rates jumped as much as 50 percent during feedback conversations. (Equivalent spikes have been found during some of the most anxiety-producing tasks, such as public speaking.) In their self-assessments, participants reported feelings that mirror what just about everyone has experienced personally: nerves, uncertainty, and anxiety. All this physiological stress has the unfortunate effect of draining a person’s mental resources.

“Giving all that negative feedback that wasn’t asked for, you might feel like you just told someone a bunch of stuff they care about,” West says, “but they just shut down and stop listening to you.”

Even if people retain the information, there’s no telling if they’ll agree with it, because social threats can create cognitive dissonance. People are inclined to flee the actual room or space where they are threatened; similarly, cognitive dissonance motivates them (pdf) to “flee” the threatening idea itself. People have been shown to more often reject disagreeable information, such as criticism, as patently untrue when they are in a threat state. The goal is self-preservation. If they can convince themselves the critique is false — My boss has no idea what he’s talking about! — they can also avoid a bruised ego.

Scientific research suggests that feedback conversations, if they are to be productive, must begin with the goal of minimizing threat response.

The Rewards of Asking for Feedback

Asking for feedback is the path to get to minimal threat response, because it appears to offer both the receiver and the giver much more psychological safety than a giver-led approach. This safety is crucial during feedback discussions because our brains will be in a much better state (pdf) for performing complex cognitive functions.

One of the strongest models for understanding social threat and reward is what psychologists call the SCARF model. The term stands for status, certainty, autonomy, relatedness, and fairness, each component referring to a domain of social interaction that can create a threat or reward state in participants. If a meeting has no clear end time or there is no well-defined agenda, attendees may feel certainty threats — negative feelings due to a lack of clarity. Alternatively, if the host clearly lays out the structure and schedule of the meeting, people may feel certainty rewards. In another instance, employees whose manager constantly checks in and meddles with every last detail would rightfully feel threats to their autonomy. They might feel more rewarded if the manager appointed people to lead individual projects, giving them a greater sense of control.

When people ask for feedback, they feel greater autonomy and certainty because they are in the driver’s seat — they can steer the conversation where it’ll be most useful. Givers, in turn, feel more certainty because they have clearer guidelines for the kind of feedback they should give. The information will be more relevant to the team member and less threatening to his or her status, ultimately making the entire discussion feel more equitable and fair. It might be uncomfortable, but manageably so.

“We’re not promising it’s going to feel good right away,” West says. “But it will be better for you in the long term.”

Getting to a Culture of Feedback

West believes what she found in that negotiation room should compel all organizations to adopt an asking model of feedback. However, she realizes how threatening it may seem to actively seek out criticism — perhaps even more nerve-racking than just happening to receive it. Unless an organization has an extremely well-oiled growth mind-set, in which employees absolutely relish the chance to get better, people are unlikely to go searching for anything that might reveal room for improvement.

The key is to start small.

“It’s like going on a diet,” West says. “You don’t want to cut out everything that’s delicious. You have to gradually replace the unhealthy with the healthy.”

At the office, leaders can begin by asking for feedback on low-stakes topics, such as the temperature in the office or how people felt about yesterday’s lunch. The point is to get people used to giving feedback that was asked for. When leaders take the first step, they signal to the wider organization that asking is important, and the low-stakes questions help build a sense of trust and agency in their team members. People are given an opportunity to feel heard, which boosts their status, makes them feel more included, and gives them a greater sense of autonomy. West says it also empowers them to give better feedback, replacing brittle smiles with more honest critiques.

If organizations keep up this behavior, they should have little trouble amping up feedback to tackle larger challenges, West says. But, she adds, organizations would be wise to roll out the initiative according to three criteria, in an effort to get feedback that is less biased, that promotes a growth mind-set, and that cements the habit as part of the corporate culture. Those three criteria are asking for feedback broadly, explicitly, and often.

In typical feedback conversations, one direct report learns of ways to improve from one manager. Even if the person asks for that feedback, it’s bound to be influenced by the manager’s unique experiences, assumptions, and mood. Indeed, behavioral economists have found that something as simple as eating or not eating lunch before making decisions can skew them in one direction or the other.

Getting broader feedback from higher- and lower-ranking people across departments can reduce the chances that feedback will be biased. The critiques or compliments will better reflect the person’s actual performance rather than the mental state of the person they asked.

The second sign of good feedback is that it is explicit. Our research has found that if people ask for more specific feedback, it’s bound to be richer and more informative than if they just ask “How am I doing?” or “What can I do better?” One simple reason, based on West and Thorson’s research, is the finding that giving feedback creates much more anxiety than getting feedback. Managers deal with incredible uncertainty about what kind of feedback is appropriate, and also about how to deliver feedback in a way that doesn’t create a threat state in their employee.

Here West says it’s up to employees to equip their managers with the right kinds of questions — a help-them-help-you approach to feedback, she says. These can include “Could you please give feedback on my presentation skills?” or “Should I have spoken up more in yesterday’s meeting?” The tactic helps managers avoid what relationship psychologists call “kitchen sinking.”

In kitchen sinking, “You say one thing that sucks, and then you pile everything else on that sucks,” West says. When employees ask for explicit feedback, they give their manager clearer boundaries.

An added benefit of asking explicitly is that employees can choose the level of construal at which they’d prefer feedback. Construal level covers the spectrum from abstract to concrete, and research has shown people have individual differences in the levels they prefer. For example, if someone wants to improve their presentation skills, a high-construal question might be “What were the goals I should have considered when presenting?” and a low-construal question could be “Did I talk too fast?” The first deals more with the why, the second with the how or what. If employees can tailor their feedback request to their preferred level of construal, they’ll be more likely to process and retain the information.

Finally, employees should ask for feedback often — for two reasons. In the short term, frequent feedback allows people to course-correct more quickly than sporadic talks. They can avoid errant thinking and unnecessary problem solving. Frequent feedback requests also shorten the time between events and feedback, so a manager’s memory of recent events is fresher and less tainted by bias.

Making Feedback a Habit

The feedback habit is important for both parties. If employees ask for feedback only every so often, they risk wasting valuable energy and discussion time to gain information that merely collects dust. The conversations might feel good, but learning won’t be taking place. With more regular interactions, askers get more comfortable asking, givers get more comfortable giving, and both gain experience in seeing how to fill the opposite role when the time comes.

Of course, there will always be times when managers must give feedback unsolicited, such as when team members make inappropriate comments or act on impulse, hurting others’ feelings, or worse. The beauty of regularly asking for feedback is that people become emotionally well-equipped to give and receive their feedback in these cases, too. In West and Thorson’s study, only one participant demonstrated this give-and-take ability. When she needed to critique the other person, unprompted, she said, “Can I give you some feedback on your eye contact?”

“She turned it into an ask,” West observes.

In work cultures where asking is the norm, she says, givers can ask permission to give explicit feedback; receivers can understand the giver’s intent; and both can enjoy more accurate feedback, fewer perceived threats, and stronger learning. It all goes back to West’s call for people to get comfortable with the uncomfortable, for the sake of personal and organizational growth.

“If both people have license to be critical, it’s actually going to be good,” West says. The culture can become one of reciprocity, not niceness, which means people will still feel incentivized to give honest feedback, but do so respectfully, since the roles might be reversed someday. “You might be a little more sensitive in how you deliver critical feedback, because you know it’s going to come back to you.”

Companies that drag their feet and uphold a culture of niceness may feel better from day to day, in other words, but it’s the ones that embrace some creative discomfort that make better decisions, and prevail in the end.

I’ve spent part of the summer catching up with a few folks that I’ve talked with for this series in years past.  And although it’s been less than a year and a half since I last spoke with Opus Research founder Dan Miller, in the fast-moving area of conversational interfaces like chatbots and voice assistants, it feels more like a lifetime.

Dan, whose company is gearing up for their Conversational Commerce Conference next month, coined the phrase “conversational commerce” in 2011.  So I wanted to catch up with him to ask where we are today with conversational commerce on the heels of the reportedly low voice shopping numbers for Alexa on the eve of the conference.  I also got his take on what impact the Cortana-Alexa integration might have on adoption of voice assistants in business.

Below is an edited transcript of our conversation.  To see the full conversation, check out the video below or click on the embedded SoundCloud player.

* * * * *

Small Business Trends:  As you come into this year’s event, where is voice in the landscape of things? How far behind is it? Is it catching on with folks? Or is it really still mostly all about chat bots, at this point?

Dan Miller:  That’s complex. I’ve written about the parallel universe in that, you don’t want to measure which modality is in front of, or behind the other. Because if you take an individual-centric approach, a customer-centric approach, that all companies say they’re aspiring towards, you’re really talking about putting tools under the control of the individuals. And I bring this up in the context of, is voice better than text? Are people going to use bots as opposed to talk to their personal assistant? It’s so obvious that it’s not either, or. It’s a matter of convenience, and at least this is what we’ll be talking through in a number of sessions; how to offer consistently correct answers to your customers or prospects at scale. And that means across all modalities.

There’s just the observation that today, a brand should invest in the resources that today, they’re powered by your CRM system, by your knowledge management, by existing back office stuff to help inform live agents. That should be consistent with what would be presented by a voice-enabled IVR [Interactive voice response]. It’s the same information that would show up on a website, often as a chat bot.

Web chat is the fertile gateway into offering conversational bots. In other words, you’ll see a lot of “bot platform” specialists come in and say, hey, give me your transcripts. Give me your chat transcripts for my platform to ingest. I’ll identify the major areas that people are calling about, the categories. And within that, we’ll know how they talk about those things. So I can build a conversational bot. At that point, do I care whether it’s rendered in text or rendered in voice? Yes, because you want to respond in the appropriate modality. But wouldn’t having those answers apply, whether it’s over a voice channel, live person to live person, over a voice channel through a speech-enabled IVR?

Well, I’ll tell you what the new wrinkle is. What about these smart speakers that you talk to, that have moved the voice channel to be everywhere that’s in the house, is it in the 40 million, 60 million? Whatever many million houses now have their intelligent endpoints that you can talk to. That means you can’t ignore voice. You can’t say, oh, voice is behind or ahead of. You can say, hey, when somebody talks to my skill on Alexa, or to me through Google Assistant, I want them to get a consistent, and a response that I’m proud of, that is engaging, and all that sort of stuff.

Small Business Trends:  So, since you are doing a conference called Conversational Commerce, one of the hot things that cropped up last week was this report from the folks at The Information, on voice shopping via Alexa. Quite honestly, the numbers are pretty low. I think it was two percent of those that have these devices, who have actually done some kind of voice shopping. And then only 10 percent of those folks did it a second time.

So there was a lot of cackling. There’s a lot of, “This stuff is, this is just nothing. Nothing to see here folks.” I would love to get your take on it. Is it already lights out for voice shopping because of what we saw last week?

Dan Miller:  Absolutely not. I haven’t written up my comment on this. Here’s what’s in my head though; you have to think of Jeff Bezos’ master plan here for Amazon. And you can’t just think of voice in a vacuum or isolated, or how it relates to the other modalities. There’s a strategy that will connect online, offline activity, you know, “Order here, pick up at Whole Foods.” So the first step is getting those sensors out there. And Alexa’s just one of many sensors that Amazon has managed to have close to your person. Voice is one of the many modalities for ordering.

But first thing’s first, get people using them in ways that are organic to their lives. If the thing’s a speaker, people think of them as a speaker. They get their news, they treat it as a radio in many respects. They’ll get the weather. There’s longstanding areas that people are happy to get the most recent news on. And even though that assessment of how people were using it, looked like so few of these culminated in a transaction, they still defined a relationship that people were forming with Alexa. They’re talking to Alexa like a person. And you don’t buy something from every person you walk up to. But you’re conversing, and that’s step one.

And then I think everything’s going according to plan. I think you’re going to see more information about local activities, because that showed up in the usage. In as much as those activities culminate in a purchase. I think you’ll find that they’re engaged in much more commercial activity, even though it’s not shopping, that culminates with buying and ordering.

I’m in the percentage that doesn’t mind saying, “Hey Alexa, we’re out of paper towels.” And she figures it out. Oh, last time you ordered these, and next day delivery, like last time. I mean, the mechanism is there to make it really easy.

Small Business Trends:  There’s an educational component that I don’t think has been fully addressed. And I also think the use case that you just outlined about, reordering paper products or things that are easy for us to do because we know exactly what we want. That’s a perfect use case currently. But the vast majority of use cases start with somebody searching for something, or looking for a recommendation. And right now, I don’t think the process flow is there yet.

Dan Miller:  Actually, that’s very perceptive. I used to talk a lot to “dialogue designers,” and I had this idea of when I thought about the protocol for invoking, for waking up Alexa, and then telling it to open a skill, and then carrying on a conversation with a brand, is not a dialogue that’s very pretty. It’s at least a “trialogue”. We’re not going to have success developing a conversation providing command line stuff to your smart speaker. It’s kind of wrong-headed.

Now is the game over? The fact that people aren’t shopping through this, is it game over for voice commerce? We’re at a point in the hype curve, because the expectation was, people are going to be shopping with this. So at this point, there will be disillusionment. It’s early days, that’s what we keep telling ourselves. There’s a lot of missing pieces, both in the dialogue design. And another topic that we chat about, consistency across what I call walled gardens.

I remember at our first Intelligent Assistant conference, four years ago, we had the people that developed Siri, the people that were working on Cortana, I don’t think that Alexa had made the scene. But at the end there’s that reporter from the local TV station, asked the guy from Microsoft, “Does Cortana talk to Siri?” He said, “Yes, I’m sure they do.”And, I mean, we’ll work out the details…

There’s an expectation that my personal, virtual agent, which I don’t have quite yet… I want it. My personal assistant, this entity that’s on my shoulder, that hears what I say, and goes and figures out how to do it for me. That’s going to be the thing I carry on my conversation with. Which means that, yes, Cortana should talk to Alexa and that sort of thing. And we sort out those gaps. Or we could start by sorting out, but you identify the gaps and you start filling them.

It would start with not making me have to have that first order decision, which will become more subconscious. But what device do I want to use? The one I have. What agent will I talk to? The one that is associated with. What do I want to do? Well, that’s contextual. That’s of the moment.

Small Business Trends:  All right, so two quick questions. I’m assuming you view this [Cortana-Alexa integration] as a win-win for both Microsoft and Amazon. Is there any, either one that actually, you feel, is the big winner in this? And then the other piece would be, what does this do to enterprise adoption of voice? Does this move the needle, or is it just this is a nice, but it’s still waiting and see?

Dan Miller:  Well, first of all, everything’s got to be a win-win-win. This shouldn’t be about the two suppliers, it should be about who’s consuming this stuff. And I use that advisedly because I’m not a big fan of the term, “consumer.” Because it just makes you feel like you’re just there, devouring things, which isn’t what happens. So, in a way, it should be a bigger win for Microsoft, just that I would think that Alexa in many ways, has exceeded the expectations of its developers and maintainers. And that Cortana was sort of languishing in my opinion.

Small Business Trends:  I totally agree with you.

Dan Miller:  So if this can bring activity from any point of ingress, that takes advantage of the formidable stuff that Microsoft has in its coveted resources, in its tools, in its developers’ support programs, this is like super win.

And then, whether it incorporates voice or not, both companies have invested significantly in some of the best brains in the business, to build voice-based conversations. Well, I mean, it’s too ironic, voice is a natural way to interact with resources within a company. It’ll be everywhere, and it’ll work better, and sound more lifelike. All this stuff will happen. Its impact on, like I said, in the parallel universe, it’s not like there’s a pie where voice succeeds at the expense of chat. But you will find more and more instances where it feels much more natural to just talk to things. You know, it’s the Star Trek vision. It’s like, “Computer.”

Small Business Trends:  I’m guessing the pie can expand as voice gets used more, and chat continues to be used a lot. So it’s not like, it’s a zero-sum game here. It sounds like things can grow and grow and grow as both these technologies get more introduced and used by consumers, or customers.

Dan Miller:  Maybe as a closing thought, because what you just said inspired in my … We’ve been getting briefings from a lot of companies that are treating and recognizing that voice conversations, be it, in calls like ours, in company conference calls, that the recordings, or it could be transcripts, but the voice-based conversations themselves are a tremendous asset.

In this saga of big data in analytics, it’s an element of unstructured big data that some new tools for doing analysis to do something as simple as, when we hang up, and you sick one of these agents, assistants on, just to evaluate what we talked about, it can almost, within a matter of minutes, deliver a summary. Identify what the topics were, what the main points were, identify in the voice files, where they were brought up and who said them. Yeah, I think we’re moving into a golden age of voice. Voice is an asset. Voice is the basis for better customer care, better employee productivity, all that stuff.

This is part of the One-on-One Interview series with thought leaders. The transcript has been edited for publication. If it’s an audio or video interview, click on the embedded player above, or subscribe via iTunes or via Stitcher.