The Feedback Fallacy
hbr.orgWhen I hear people talking up their policy of
> “encouraging harsh feedback” and subjecting workers to “intense and awkward” real-time 360s
... I can't help but think how much a review there sounds like a capitalist version of a Maoist "struggle session" (see https://en.wikipedia.org/wiki/Struggle_session).
The point of a struggle session wasn't to build people up; it was to tear them down. A culture where the way you move up is by tearing your colleagues down is generally a dysfunctional culture. Everybody involved ends up losing sight of things outside their four walls (such as, you know, what customers want), because they're spending all their time squaring off against each other internally.
I would suggest a healthier, more sustainable culture could be imagined by contemplating these words from the Tao Te Ching (Stephen Mitchell translation, chapter 27: http://taoteching.org.uk/index.php?c=27&a=Stephen+Mitchell):
What is a good man but a bad man's teacher?
What is a bad man but a good man's job?
So far, all of the people I know that have been boasting about their "I gave brutally honest feedback" only mean "I will tear you down at every opportunity".
It is always "brutaly honest when it comes to put people down, but weirdly enough, these same person never have positive feedback to give.
Even if it was the case, it would not be worth it. Even as a foreigner weirded out by the "everything is awesome" way of speaking in California, tearing down your colleagues is not how you build a healthy team.
"People who are brutally honest generally enjoy the brutality more than the honesty." - Richard Needham
When I did online dating, I had a few red flags that I used to exit out of a profile. One of them was boasting about being "brutally honest", "honest to a fault", and/or "I tell it like it is."
I would be ok with harsh feedback if that were acceptable also upwards and taken seriously. But most of the time it only goes down or sideways. Feedback up the chain is not being taken seriously.
I recommend this article "The Gervais Principle, Or The Office According to The Office".
https://www.ribbonfarm.com/2009/10/07/the-gervais-principle-...
Horizontal aggression.
I've heard this kindof system of feedback is brutal, and is certainly not for everyone.
I think the point is to aim for apolitical and tactless information transfer. The only chance it'd have of working is if there's a shared understanding that correcting mistakes is more valuable than political standing; and this applies to everyone, especially the bosses.
The saner way to try and get these benefits is to foster psychological safety.
Struggle sessions aren't analogous, as they were held against class enemies (out-group). More akin are the self-criticism sessions that Maoist cadre held. Communist cadre criticized themselves. Which, if you want another analogy stretching backward, would be confession in the Roman Catholic church.
Tangentially, struggle sessions only arose to prominence during the Cultural Revolution. The main criticism of class enemies were the Speak Bitterness sessions of the 1940s, where communist cadre would facilitate a village's peasants criticism of its landlords. Despite the notion that communist cadre were instigating violence, most of Mao's letters were to cadre who were taken aback by peasant's (sometimes fatal) violence toward long hated landlords, Japanese collaborators etc.
Oh.. you're not the only one to find similarities between some corporate cultures and oppression mechanisms from communist regimes.
For example, whenever I had to do my yearly review and write in my opinions about what I did good/bad/could have done differently which afterwards were discussed with my manager and then shared among leadership, I couldn't help but think of a milder version of https://translate.google.ro/translate?hl=en&sl=es&u=https://...
How the author seems to get and present such a distorted view of the original Netflix culture slides (which we modeled quite a lot of our original culture from) is completely beyond me. Nowhere did anybody say that feedback should be a one sided activity, here's the original source for your reference:
https://www.slideshare.net/reed2001/culture-1798664/11-11Com...
The "Communication" slide (p.11) even _starts_ with the skill of listening rather than talking well.
We're living an open, candid, two sided feedback culture in our company and it's by far the best thing that has ever happened to my personal development. The impact on my own perception and management style has been tremendous. Maybe I'm biased by being a founder, but on our most recent anonymous employee survey, structured feedback got an average 8/10 on satisfaction.
I take that just like the original ideas of Scrum or Agile, transparent Feedback culture has apparently been bastardized enough by toxic companies living the letter but not the spirit that it's apparently starting to get into a negative perception... that's a pity but won't stop us from giving and getting transparent feedback regularly.
Having worked in a large company that is famous for its feedback culture, I think the type of feedback culture you refer to is hard to scale. While I think it helps people improve, I’ve mostly seen how feedback is used to hurt people.
+1 to this, at scale it's really important to channel feedback on useful topics. Learning that everyone hates working with your authentication system is really quite different than everyone hating the lone maintainer who keeps it alive in their "spare time".
> The only realm in which humans are an unimpeachable source of truth is that of their own feelings and experiences
Side note, but this strikes me as extremely wrong, based on 35 years of living among humans. Humans have unsurpassed abilities of self-deception, cognitive dissonance tolerance, and building monumental mental models to hide the truth about themselves from themselves.
As one economist put it, your 'self' isn't the CEO, it's the PR department. You produce a PR-approved spin of your own experiences and feelings for your internal, conscious consumption.
I think the error is including experiences.
I am the sole source for what I'm feeling. Whether I admit that or not is another story, but if I think I feel sad, then I feel sad.
Now, whether that emotion is justified is another story and whether the related experience should induce sadness can be up for debate and if I'm interpreting the situation correctly and all of that.
But we can't argue that I'm not feeling sad. I would probably have said that "The only realm in which humans are an unimpeachable source of truth is that of their own feelings and perspectives"
> but if I think I feel sad, then I feel sad.
Our sensitivity to our own emotions doesn't seem to be that great. People talk about happiness sneaking up on you, or having a realization that they're unhappy in a relationship. Presumably the feelings were there, but they weren't conscious of it.
I suspect if I did have dials in my head that displayed the level of each emotion, I'd go insane trying to move the needles. There may be some logic to the arrangement.
This doesn't actually go against what's being said.
Unexamined feelings are still felt. I'm still the only one who can tell you with 100% accuracy what I'm feeling.
Even if you're technically wrong about what you're feeling, there's no way for me to verify that you're wrong. You are still the sole arbiter of absolute truth on that matter.
Because people love to read into things that aren't there. Apparently, I have a very threatening face when I'm concentrating on something and I look kind of angry. I'm not. And it's kind of exhausting to have people try and tell you what you're feeling when they have no clue as to what's actually going on inside of your head.
As a counterpoint, consider how people say depression is anger turned inward, or perhaps the opposite, that anger is self-hatred turned outwards.
Another concept is, when you have chronic pain, you may forget what it's like not to hurt. And then one day you don't and you suddenly have perspective.
What I'm saying, I guess, is that generally no other person can tell you your feelings or emotions though they may try. But that doesn't mean you are necessarily correct in what you think they are.
Sole source, yes, but that doesn't mean you're reliable. For anything more complicated than "sad", and maybe even for that, self-deception and lack of self-knowledge can be issues. For example, have you correctly identified exactly what it is you're sad about? I know I've caught myself believing dumb things about my feelings.
Like I said in my post, everything surrounding it is up for debate. But you can't tell me what I'm feeling. If I say, "I'm sad", that's the end of the story. It's not up for debate.
Even if I'm lying about what I'm feeling, there's no way for you to really know. There's always some room for doubt.
You know that you are sad, but you generally don't know why so you guess. And your own guess about the source of your feelings usually isn't any better than what other people would have guessed about you.
Which is why I said the feeling may not be justified. But my feelings are what they are. Not being able to explain why I'm feeling them does not invalidate the fact that I do feel them.
You cannot tell me what I'm feeling.
Or just modify it to "their own feelings and subjective experiences."
> Adding up all the inaccurate redness ratings—“gray,” “pretty gray,” “whitish gray,” “muddy brown,” and so on—and averaging them leads us further away both from learning anything reliable about the individuals’ personal experiences of the rose and from the actual truth of how red our rose really is.
I don't understand this comment. How does averaging noisy signal, even systematically noisy signal, result in something that is noisier than any individual signal? I would have assumed the average would converge on (real signal + systematic error).
The author is arguing that the real signal is zero and the systematic error is large, so you will always end up converging on a repeatable but useless value. Technically, taking only one sample could have gotten you closer because there is a 50% chance that the random error would have gone in the opposite direction to the systemic error, although the author is wrong to phrase that like it's some kind of advantage, because the other fifty percent of the time the random error will make the total error even worse.
They say later that
> When a feedback instrument surveys eight colleagues about your business acumen, your score of 3.79 is far greater a distortion than if it simply surveyed one person about you—the 3.79 number is all noise, no signal.
Which implies to me that they believe there is signal there, but that it goes away when aggregated?
I think by "surveyed" they don't mean "asked one person for a score" but rather got some overall information from one person including their qualitative feelings and perceptions. There is signal in those as they discuss elsewhere in the article, but the quantitative rating allegedly has no value even when averaging. That's the charitable reading, anyway.
Yeah, I'd like to see what statistical theory they are using here. I don't think it's sound. It's unfortunate since I think the article is otherwise quite good.
If methodology is unsound, there can be negative value and outcomes. Ie. active trading is a consistent loser for most people. Methodology can lead people astray.
? I still don't get it.
Averaging values with random divergence from the truth is useful. Averaging values with random divergence from nonsense is not.
Specifically, I don't get how is one random value is supposed to more accurate than many random values averaged.
I think that's more of a relative impact. When you have just one measurement, you know it's not particularly reliable. When you have a bunch, we are conditioned to think it's more reliable.
So in the latter case, the distance between its reliability and its perceived reliability is greater than in the former case.
I agree with you, that it has to do with perception of reliability. However, the article seems to state that there is an actual greater error with more inputs. That's what I don't understand.
"We cannot remove the error by adding more data inputs and averaging them out, and doing that actually makes the error bigger."
I don't see how it "makes the error bigger". Maybe I'm being too literal and the writer is truly referring to the perception of the results carrying more weight, and therefore having a "bigger error".
An individual rating can tell you something about the individual’s experience.
Averaging the ratings of multiple people tells you nothing since it washes out the individual experiences. I.e. individual samples hold meaning about the samples themselves but aggregation of samples is just noise.
Agreed feedback is tough. One of the problems is that it depends a lot on the setting and team. A person in one team might get some kind of feedback, but in another team the same person could get the opposite advice.
The other problem in our industry being relative inexperience of many people. When everyone in the team has had < 5yrs experience who is really qualified to objectively describe what a person is doing wrong? I know when I was a an inexperienced manager I gave "constructive feedback" that I now recognize was wrong.
I like the simple actionable advice near the end in the "Explore the present, past, and future" section.
The authors' biographies just below that caught my attention:
Ashley Goodall is the senior vice president of leadership and team intelligence at Cisco Systems and a coauthor of Nine Lies About Work: A Freethinking Leader’s Guide to the Real World
Cisco Systems is the kind of giant multinational conglomerate that I would expect to be a (to put it euphemistically) corporate shithole. The kind of place where this kind of forward-thinking management gets lip-service but in the end, for the most part, it's the same kind of Gervais Principle Hunger Games hierarchy we've come to deride and deplore in these type of threads.
Is this not the case?
Ever since I read McKinsey's "in search of excellence" I've been a bit skeptical of these sort of "to run your company better just follow this one weird trick" pieces. The book when I read it was quite convincing, but when I read the results of the companies cited as "excellent" by the book later on, about half had gone bankrupt or otherwise ceased trading!
Good to Great suffers similarly:
Steven D. Levitt noted that some of the companies selected as "great" have since gotten into serious trouble, such as Circuit City and Fannie Mae, while only Nucor had "dramatically outperformed the stock market" and "Abbott Labs and Wells Fargo have done okay". He further states that investing in the portfolio of the 11 companies covered by the book, in the year of 2001, would actually result in underperforming the S&P 500. Levitt concludes that books like this are "mostly backward-looking" and can't offer a guide for the future."
Of course there is the question whether a company's goal should be to be around forever or be excellent for some time.
GE and Jack Welch come to mind. For a long time they were held up as an example of good management but somehow this didn’t result in a long term healthy company.
If you're hearing about good anything, you're hearing about good marketing. ;)
All conversation is a continuation of psychological warfare.
managers spend way too much time on trying to get inside their employees head and fiddle around imho, they really don't have the skills or training. if they spent this time on actually helping do the work in terms of what's needed and wanted so everyone could just go home then the world would be a much better place.
Yes. They'd hire someone like Wayne Dyer to develop treatment plans at a psychiatric hospital. My favorite personal example was when my employer hired a management engineering consultant company to help people who spent their time in meetings to trim fat, i.e., fire people who did the work. One of the ways that was used to help us dumb people understand what a quality product is was to ask us which restuarant has a higher quality product: a gourmet steakhouse or McDonald's. The answer? It was supposed to be a 'gotcha' of course. People inevitably answered the "gourmet steakhouse", but our brilliant teacher laid down some truth that day: McDonald's is of equally high quality because they do very well at providing a consistent product throughout all of their franchises. I told her that there's a difference between consistency and quality. A consistently low quality product is still a low quality product. Management just didn't get it. This was at a time in the early 90s when my employers touted themselves as being a "World Class Hospital". What kind of claim is that when American hospitals were world beaters?
This is such a cliche about consistency anyway. Who says they are consistent? The whole thing about franchising is they are independent businesses, and McDonald's isn't harmed if a given one does terribly and goes out of business.
Certainly after some critical mass, you don't have to be as consistent to retain customers. We all know how first impressions matter at any job.
“...if they spent this time on actually helping do the work in terms of what's needed and wanted“
Is what good managers do.
sadly, they are rare in my experience
They get told to fuck off.. Sort of employees own fault.
Cooperation require peers to do well, as well.
The best I can recommend for managers: https://roadmap.manager-tools.com/product/basics/1
Hey, we're in my territory!
Something to note is this premise is disputed by A LOT of experts in this domain. They've been challenged to a debate on this by Marc Effron (not my favorite guy but does generally have grounded thinking in his work) with proceeds benefiting charity but no uptake yet. They have also not shared their data in a way that enables replication or outside validation (or hadn't lest I checked). They is a pretty robust history of peer-reviewed research that doesn't come to the same conclusion.
I think they are on to something here but I think they are taking too far. For their "Source of Truth" section, most of what they say is true, particularly about rating and assessment of employees. Extending that to say suggestions on future behavior from others is also value-less is wrong. The former is problematic because it's treated as a source of objective truth that personnel decisions are made against, the latter is clearly a single subjective data point people can reflect on and potentially integrate into future behavior.
For "How We Learn" section, much of the words written are true but the conclusion goes beyond what I have seen data justify. Yes, we get better faster at the things we're already better at. Should that mean focusing on strengths is often a better coaching direction than weaknesses, yes. Does that mean we shouldn't do the latter or that it is value-less? No. It does mean we have to find ways to make weakness-oriented feedback happen in a repeatably non-threatening way.
On the "Excellence" section I'm again in agreement with most of it, but have fewer overall critiques. I should think more about this. I still say the conclusions are not natural endpoints for the points he makes.
Then, interestingly there is a table near the bottom. I find it interesting that many of those things you should "Try" rather than "Instead of" are indeed types of feedback and modern organizational development professionals espouse. Language really does matter, that's a great table with great suggestions, and most of those are feedback prompts.
All that said, I truly like Marcus Buckingham and find his work to typically be evidence-based (to the extent work like this can be) and on solid ground. Here is a brief video of him that is on a related topic that I think everyone should really take to heart if interested in this topic. https://www.marcusbuckingham.com/rwtb/performance-management...
In my limited experience as a developer, managers and supervisors need to first open their mouths before we can determine if the feedback is good or not.
Maybe this is some problem on a large institution level. For me, every time something I'd done was not correct or I'd failed, it felt very valuable to me that someone noticed and told me and how they wanted it to be. It reinforced in my mind what was good and that what I thought was not was, in fact, not. I always felt way better coming out of an interaction like that.
The worst feeling is when I feel something is going poorly and no one will tell me they think that about my thing.
The part about pain isn’t actually necessarily true. When I was recovering from surgery, they would ask “on a scale of 1-10, with 10 being the worst possible pain, how much pain are you feeling?” I imagined how excruciatingly bad the worst possible pain would be and rated my pain a 2. When my pain went up to a self-reported 3, the nurses were surprised to find me crying from the pain...
Not everyone interprets the scale the same way...
> Not everyone interprets the scale the same way...
That's the exact point the author was making with this example. A doctor cannot normalize pain ratings across patients, and has to rely on each patient scoring their own pain relative to their own experiences. Your 2/10 could be 7/10 for someone else - all that matters is that you have a measurement stick to compare pain for yourself, so that medical staff know that your 3/10 is more than your 2/10.
The problem is that my 2/10 is rather distressful and doctors assume it isn't, because they assume that self-reported pain ratings map consistently to subjective levels of distress across patients.
The recent Malcolm Gladwell book also talks about how poorly we size up others. Oh and I really like "Excellence is idiosyncratic.", very true.
Feedback is dangerous for many reasons, but mostly because the signal/noise ratio is mediocre.
The problem with noisy feedback is the instability of the closed-loop system, and it can quickly become erratic and ultimately counter-productive.
Thus, we should only use feedback when the signal is clear and loud, mostly at the extremes, when performance is outstanding or catastrophic.
Paraphrased - How should giving feedback work? > Precisely how it will be received most effectively. - How is that? > You must learn that for every person. Here are some ideas.
This is something that I feel like many in engineering have to grow to appreciate (or at least I did, and I see some of the same markers in many of my peers that I had.) not just about feedback and interpersonal relationships, but about everything. There are likely many things that you have an intuitive feel for, but just as many you have to calmly, slowly, and carefully consider yourself, your actions, and their consequences, if you want to be more effective or be better.
In the past, I coached jr. high and highschool boys basketball. Some players got lots of leeway to make mistakes before getting subbed out during games, because they were capable of learning from those mistakes themselves, and coach feedback didn't help their learning process. Other players would make mistakes and immediately get subbed out, mistake pointed out, discussed, correct action proscribed, and shortly subbed back in. They needed the outside feedback to process "that was a mistake, I shouldn't do it again." Some players goofed off in practice and got to sit on the sidelines. Some players goofed off in practice and got to run laps. I had several discussions with parents about why their son got "special treatment" when really it was about me trying to give effective feedback. And I'm not saying I was awesome at this, or always adjusted my approach for every kid in every situation, but when I could, everyone's results were better.
A larger rant I have on this and any topic that circles back to effectiveness is how to respond to "What is the best thing to do in X situation?"
For example, today in a team meeting, my group was discussing way to improve performance in one of our systems. In the past, I've seen caching greatly improve performance over database optimizations, so I'm optimistic about a better caching paradigm, whereas one of my team members is looking at a longer-term code maintenance and simplicity perspective that says, do fewer things better, so optimize our database calls. Long slog to figure out which is better, but we can't just generalize from past experience. What should we do? Precisely what is needed. How do we know that? We'll have figure it out. We have some general guidelines, but we'll have to figure out how to apply them to this situation.
I have two kids. The older one would usually go to anyone when he was a baby and be happy, smile, coo, play, for 10 minutes or so before he would get worried about where mom or dad were. The younger one usually senses that mom or dad might be handing him off, and gets upset and takes 5 minutes before he calms down. But sometimes the older one would cry going from my wife to me, and sometimes the younger one will happily go with the church nursery lady. Why? I don't know. We'll figure it out.
Generalization is great, generalization is helpful, generalization is not right in every case. With people, if you really want to be effective in feedback or anything, you have to figure out how to have an approach that you are generally successful at, how to generally alter it when you need to, and then laser focus that flexibility for the people, relationships, and situations that you really care about.
This is an excellent article that deserves careful study.