Seek feedback and iterate

As I sat there in front of my screen developing the spreadsheet/tool that was to be shared with the more than hundred salespeople in the company I realised I had doubts – would this really work? Was this an improvement to what they already had? Or was it more change for the sake of change?

I honestly felt that it was a genuine improvement, but I didn’t know. And having spent so much time already getting to where I was on its development, the last thing I wanted to hear was that I was on the wrong track, and that my work would come to nought.

Also, I was in a state of flow, and getting feedback was an overhead that would break that. Did I really want that?

I got up from my seat, walked over the coffee machine and made myself a coffee while mulling over this: to get feedback or not to get feedback – that was the question.


Development’s fun – I enjoy it. Solving technical problems and shipping something useful is one of the main reasons I entered the tech/data space. But having moved into a managerial role it’s something I do less and less – development’s now a team sport, one in which I’m no longer the star.

That sense of accomplishment when something you create goes out into the wild and receives accolades is something I really miss.

If this piece of development went live, I would well get back that high.


The steaming cup of coffee in my hand relaxed me and made things a little clearer: I made a mistake of having worked on the development as long as I had without getting feedback. I should have followed the same advice I always give my team: don’t work on something for too long without getting feedback, otherwise you may just find yourself spending days or even weeks on end working on something nobody wants to use.

(Thank goodness I also always drill it into them: “Do what I say; not what I do.”)

The longer I went without feedback, the harder it was psychologically to want to seek it. But I knew I had to do it.

I personally had doubts, and this was my baby here. I gritted my teeth, got up from my seat, and started seeking feedback. Like Sun Wukong (aka Monkey King), I reluctantly travelled to collect the sutras, during which I had to bear the pain of hearing things like “it won’t work” and having my “great” ideas turn bad.

It was emotionally draining, but it had to be done.


Hard as it was, I stopped development that day. The week (and weekend!) of frantic development came up to nothing.

Still, there was something I got out of it — like they say, there are no mistakes, only learning opportunities. And for me, it was a reminder to myself to seek feedback early, and iterate.


It was ironic that it was about this time that I was reading the Lean Startup by Eric Reis, one of the pioneering books on iteration and getting feedback. I leave you with this passage that I always use to remind myself before I go too deep into a development or process rabbit hole (text in bold mine):

From the point of view of individual efficiency, working in large batches makes sense. It also has other benefits: it promotes skill building, makes it easier to hold individual contributors accountable, and, most important, allows experts to work without interruption. At least that’s the theory. Unfortunately, reality seldom works out that way.

Consider our hypothetical example. After passing thirty design drawings to engineering, the designer is free to turn his or her attention to the next project. But remember the problems that came up during the envelope stuffing exercise. What happens when engineering has questions about how the drawings are supposed to work? What if some of the drawings are unclear? What if something goes wrong when engineering attempts to use the drawings?

These problems inevitably turn into interruptions for the designer, and how those interruptions are interfering with the next large batch the designer is supposed to be working on. If the drawings need to be redone, the engineers may become idle while they wait for the rework to be completed. If the designer is not available, the engineers may have to redo the designs themselves. This is why so few products are actually built the way they are designed.

The passionate introvert

This TED talk really surprised me.

The content was great, but it was Brian Little’s delivery that really made me go “wow!”

So many times during the talk it felt I wasn’t listening to him talk on the subject of “personality” but rather his grandchildren. His passion was evident, and his joy contagious. I couldn’t help but give him a personal standing ovation at the end.

It is with this sort of passion that we should approach our careers; our lives.

Deciphering Fake

This was supposed to be a post on radical transparency.

But an article bashing radical transparency just left me feeling so outraged with its lies and misleading statements that I just spent the last four hours of my life writing this warning to all of us media-consumers out there: Don’t trust all you see, even if it says “research”, links to academic papers, and cites its sources!


The first time I heard about radical transparency was from G.

And though I hadn’t heard the term before then, it was something I felt that I could really relate to; something I already did.

Not because I thought that it brought the best outcomes, but because my mind was just wired that way.

I’ll tell you why in the post on radical transparency I eventually do write (maybe next week?), but hint: it’s got to do with having an awful brain for lies.


For today, let’s talk about the article that enraged me.

I found it while reading up on radical transparency for the post I had intended to write: Radical transparency sounds great until you consider the research.

I looked forward to reading it just based on its title, as it was perhaps a warning I needed to heed: maybe I ought to be a little less transparent with my dealings with people?

The word “research” also appealed very much to the scientist in me, giving it more weight than it would have had without.

Almost immediately though, within the first paragraph, a red flag was raised.

Here’s what it said:

Radical transparency is an old management approach with new branding. Previously called micromanagement or snooping, this approach supposedly creates higher performance and trust by letting everyone know what’s on the table.

You see, I’m an amateur rhetorician (well, not really, but I am currently reading Jay Heinrichs‘ book Thank You for Arguing) and smelt a rat: I knew radical transparency wasn’t synonymous with “micromanagement” or “snooping”, or even remotely analogous.

My rhetoricsense tingled. Something was up but I didn’t quite know what. So I did a quick search on logical fallacies, and identified what was wrong: the author was guilty of a false comparison!

Snooping, micromanagement, and radical transparency were qualitatively very different things, and there was no “new branding” apparent to me whatsoever.

  • Snooping to me implies trying to find out information others deem to be private and not expect to share;
  • Micromanagement to me implies a person in authority dictating to a worker how to do a job without giving the worker much or any degree of autonomy;
  • Radical transparency to me implies making what may sometimes be deemed private open to everyone, but making sure everyone knows it is no longer private.

I could live with micromanagement, to a certain extent. I could live with radical transparency (I think). But I would probably not be able to take snooping very well.

You can’t really club them together.

Was the author trying to mislead his readers by saying they were the same except for rebranding?

Whatever the case, I continued, albeit with caution.


Then I came across this paragraph, which appeared filled with juicy insights:

But research about human judgement suggests that relying on such data is a mistake. People are terrible at assessing trustworthiness and most skills. Assessments are driven not by real actions, but by appearance and personal situation. On top of these potential inaccuracies, labeling someone as untrustworthy or poor in certain skills has a corrosive effect on collaboration and morale, perhaps one of the reasons why Bridgewater has in the past had very low retention rates that costed the company tens of millions of dollars a year.

The links in the quote above were found on the original article. I clicked on every single one of them to learn more.

(And boy did I learn. I learned that if you take an author’s word for it at face value, despite the authoritative-looking links you’d be hoodwinked quicker than you can say “radical transparency”.)

Here’s my commentary on each of the links in the paragraph shared above:

  • “terrible at assessing trustworthiness”
    • This link brings you to a paper talking about assessing trustworthiness from facial cues. The experiment involved asking strangers to play a game to see if people would invest more money in faces that appeared more trustworthy. If radical transparency involved asking you to rate your colleagues, an hour after you got to know them, on trustworthiness based on how their face looked, then yes, this is relevant.
  • “most skills”
    • This link brings you to a paper talking about the JDS or Job Diagnostic Survey tool, which basically assesses the fit between workers and their jobs. The paper surmises that the tool works, though warns that it is easily faked. But for it to support the premise that “people are terrible at assessing most skills” is ridiculous, because the paper actually doesn’t say that.
  • “appearance” and “personal situation”
    • These two links are paywalled, but based on the abstracts these are related to people assessing people in TV commercials (for the first link) and strangers (for the second). Like the experiment in the “assessing trustworthiness” link above, this is about assessments of people whom you know very little about. Radical transparency isn’t about assessing strangers one-off. Again, I don’t see the relevance.
  • “has a corrosive effect on collaboration and morale”
    • Paywalled. The first sentence of the abstract? “Four studies examined the relation between trust and loneliness.” I’m curious to know what the article is about, but given I don’t know enough I’m not going to judge on this one.
  • “very low retention rates”
    • This link brings you to an interview with an author who wrote about Bridgewater’s radical transparency. The author actually praised its implementation at Bridgewater and was extremely supportive of it. Though it was mentioned that there was a 25% turnover rate, there was no mention of it costing “the company tens of millions of dollars a year”. Also, assuming that it does cost the company tens of millions of dollars a year, could the benefits outweigh the costs? If being radically transparent brings in more than the “tens of millions of dollars a year” that it  hypothetically costs, it’d still be worth it.

I’d always been extremely curious as to the effect of knowing my peer’s salary, and them knowing mine.

I’d even considered moving to a company that did just that for just this reason because I personally thought it was a great idea.

So when I came across the following that the author wrote, it came as quite a surprise:

Publishing individual salaries has negative consequences. While companies should never prevent people from sharing their compensation (and in many states it’s illegal to do so), publishing these numbers for all to see psychologically harms people who are not at the top of the pay scale. Research shows that this directly reduces productivity by over 50% and increases absenteeism among lower paid employees by 13.5%, even when their pay is based exclusively on output.

The first link talks about income disparity and its negative effect on happiness, a common finding in psychological research.

That the author worded it in this way (i.e. “top of the pay scale”) seems deliberately misleading. There’s a lot of dependence on the “reference group” – e.g. a junior employee, despite earning far less than the CEO, would generally not be too concerned. Also, full individual salary disclosure isn’t necessary for radical transparency; compressed payscales and other forms of salary disclosure could be used instead.

The second link was the one that I was more interested in: could salary disclosure really lower productivity and increase absenteeism, even when pay was based on output?

The author said yes.

I read the paper and found otherwise.

What the study found was that it was perceived fairness that had the greatest negative effects, not the disclosure of salary information per se. Where there was wage disparity and output was not easily observable (i.e. there was no way to tell which worker “deserved” the most), those who were paid less than their peers were the most negatively affected, as they would have perceived it as unfair.

And in a world of radical transparency, I’d think that “output” information would also be something that would be freely shared, reducing any perceived unfairness.


I don’t know what led the author to write what he wrote. I was very close to just taking what he wrote at face value, and if it wasn’t for me being a little perplexed and curious at some of the claims that were cited I’d never have uncovered the deceits.

To be clear, I’d just like to add that there is a chance that there was no malice involved, just sloppy research and misinformed conclusions.

But whatever the case, it made me realise how much we take good, honest writing for granted.

We shouldn’t.

And for me, not any more.

Supposedly Irrelevant Factors

I’m halfway through reading one of the best books I’ve read in a long while: Misbehaving, by Richard H. Thaler.

One of the things that most stuck with me was that of “supposedly irrelevant factors“, which refers to something that, in theory, should not affect or influence the thinking of a rational person but does.

Thaler has also written about this in an article for the New York Times. The example that Thaler shared in the article is that of the grading of the notoriously difficult midterm exam that he gives, which he uses to separate his really good students from the rest.

As per the usual practice in academia, the maximum marks you could get in that exam was a hundred. But this posed a problem. Because of the difficulty of the exam, his students were averaging only 72 out of a possible hundred.

Though it didn’t affect the overall grades the students got, since their relative scores were more important than their absolutes (see: bell curve), they didn’t quite like getting such low marks and many complained.

Thaler got worried that the complaints might eventually lead to the loss of his job. So he made a change: instead of having the exam be out of a hundred marks, he made out of 137.

This change had a couple of things going for it: firstly, it made it more difficult to calculate a percentage score; at the same time, it allowed him to give students higher marks, closer to what they would have got on the usual, less challenging exams.

Students were on average now scoring 96 instead of 72, with some delightfully achieving scores above a hundred. Despite the lower percentage scores his students were getting (70 now instead of 72!) his students were happier.

This wasn’t supposed to happen. But it did.

To him, test scores were a supposedly irrelevant factor. To his students, they were anything but.


The concept of supposedly irrational factors really appeals to me because it’s something we tend to forget, especially in a work setting (because everyone’s somehow less human!) and yet something that may have large consequences.

Imagine this conversation between you and your boss:

Boss: Run this report for them every day by noon.

You: Why? They shouldn’t need to see the numbers at that high frequency – no actions can be taken at this late stage that will change anything anyway.

Boss: Just do it.

How would you feel? Would it make you a little more likely to be unhappy? To consider quitting and doing work that feels more worthwhile?

The knowledge of why could be seen as a supposedly irrelevant factor.

Knowing why you’re doing it wouldn’t really change anything – doing what your boss tells you is part of your job. Maybe she knows something you don’t.

But still, it just doesn’t feel right and the knowledge that there might be a purpose that isn’t shared doesn’t make you any less unhappy.

But what if your boss said this instead after you’d asked “why?”:

Boss: Maybe it won’t change anything. And I know it seems pointless from an actionable point of view. But what I know is that it helps calm their nerves; it helps calm their boss’ nerves. It’s not easy being in their shoes – they’re currently under immense pressure, and I’m hoping to support them in whatever way we can.

How would you feel now? If it were me, I’d actually feel even more empowered than before, and that I was making a positive difference in people’s lives.

The simple knowledge of why changes things quite a bit, though it really shouldn’t.

We’re not quite the uber-rationals we think we are.


Some interesting “supposedly irrelevant factors” examples that I’ve come across:

  • Choice architecture, and the default option – we go for default options more often than would be expected, even if the default’s the worst option available
  • Decoy pricing – Classic experiment done by The Economist where the introduction of a third, obviously poor subscription option made the most expensive option much more appealing
  • The Endowment Effect – Owning an item makes it seem much more valuable than it was before ownership came about

Anticipation, proactivity, and the Invisibles

Just read an article via Slashdot on this thing called “Tab Warming” that the Mozilla team is testing for the Firefox Web Browser.

I won’t go into the details, but in essence what Tab Warming does is that it anticipates whether or not you’ll click on a link, and if it does it “paints” the page in the background saving you milliseconds of loading time when you eventually click on it.

Despite the seemingly very small difference in absolute time, (I mean, really, milliseconds?) it has the potential to be the difference between somebody thinking about the page loading versus somebody not thinking about the page loading at all and purely on the content of the page.

And that’s huge. Isn’t the ultimate aim of interface usability to become invisible, after all?

This reminded me of how the best people I’ve worked with are those who pretty much do this all the time: they anticipate what I may need, and even before I’d mentioned it they’re bringing it up and telling me it’s already done.

And to me, despite my never thinking about them (because I don’t have to!), they’re the ultimate stars in our work lives.

To the invisibles: Thank you.

The difficulties of doing “deep work”

These past two weeks I’ve been on leave, staying at home and being a dad to my 2-year-old son.

He’s got quite a standard schedule: the wife and I bring him out in the morning to let him “see the world”, have breakfast, and/or visit the grandparents etc.; he comes back around noon, takes a snack, sometimes a full lunch, then goes to bed for his afternoon nap.

Depending on how tired we manage to get him before his nap, he’ll wake up between 5-6pm. Sometimes though, he surprises us and wakes up at 3pm or earlier. It’s happened enough times for us to unconsciously be in a constant state of high alert throughout his nap, hearing out for his cries.

(I wonder if it’s something akin to gambling addiction, where the release of dopamine is increased when winning is intermittent or unpredictable. Just in this case, it’s more of the opposite in that we’re always in fear that the boy wakes up before schedule!)

What I realised was that during this “high alert” phase, I’ve always found it hard to do anything that requires more than a cursory time commitment, anything that would not be considered “deep work”.

Deep work – the ability to focus without distraction on a cognitively demanding task. It’s a skill that allows you to quickly master complicated information and produce better results in less time.

Deep work requires a commitment of uninterrupted time. Going into a cognitively demanding task and then being interrupted halfway through often means that whatever you’d done up till then is wasted, or at least any progress made set back considerably.

I remember once making good progress on a machine learning project I was doing for work. Suddenly the boy cried and I had no choice but to stop. When I resumed my programming in the night, I found it almost impossible to resume where I left off. What made so much sense just 8 hours earlier made little sense now, and getting back up to speed was a slow and painful process.

What this means is that when I’m in the midst of “expected interruption” I’m gravitating toward activities that are not subject to such a regression. For example:

  • Instead of reading complex works of non-fiction, I’m reading “lighter” books that I can easily dive in and out of, especially great are those where the chapters are short or where stand-alone ideas are wrapped up within a couple of pages.
  • Instead of practicing my technical data science skills or actually writing code, which tends to require a heavy commitment of uninterrupted time, I’m practicing typing on typeracer.com, where within seconds I’m racing against my typing peers and getting an instant hit of dopamine since I win so much but not all the time (see: above link on “gambling addiction”!)
  • Instead of setting my goals for the new year and how I’m planning to achieve them, I’m thinking about what I feel like having for dinner and how to cook it.

Though I always knew this problem also existed at work, I’m now more aware of the impact it might have.

For those of us constantly barraged by “urgent minutiae” or unscheduled projects (i.e. pretty much all of us I bet), the lack of a system or structured approach toward addressing interruptions could lead us to a lifetime of firefighting at the expense of actually doing the impactful, deep work we were brought on board to do.

Personally, these are the things I do to prevent myself from drowning in urgent minutiae:

  • Relagating of e-mails to an hourly or two-hourly affair, which helps you avoid being interrupted mid-thought or while putting the finishing touches on your magnum opus.
  • Scheduling of a “meeting with yourself”, which blocks your calendar and allows you to work, guilt-free, on your most important tasks.
  • Addressing anything that takes 2 minutes or less immediately, which frees the mind of unnecessary clutter, something I picked up while implementing the wonderful Getting Things Done methodology.
  • Focusing on your highest priority tasks while ignoring everything else for the day, which is dangerous but oftentimes necessary.

(PS: To date I haven’t quite found a “hack” for myself at home, though. My kid doesn’t respect my calendar, nor does he bother with e-mails, and he’s just about impossible to ignore.)

(PPS: Above definition of “deep work” found on Cal Newport’s website, whom I *think* coined the term. The term itself came into my consciousness after it was first mentioned to me by S on my team at work.)

Feeling good about one’s work

I was just “thinking about things” when this thought came into my head: To feel good about one’s work, there are two sides of validation: the internal and the external.

  • External validation: somebody tells you, “you’ve done well. This is excellent!”
  • Internal validation: you tell yourself, “you’ve done well; you’d set out to do something well and you did it.”

This past year I’ve had a lot of positive feedback on my work. Still, something was missing.

“This is great work,” they’d say, and I’d smile.

Was it really? I’d think, feeling less than satisfied, tinged with impostor syndrome.

Luckily the opposite was true, too.

The work my team and I do are quite regularly behind-the-scenes “enablement” work. Nobody but us knows; no rah-rah; no fanfare.

They’d say nothing, even after I’d done something I thought was remarkable.

Still, this is awesome! I’d think.

And I’d be satisfied. Almost happy.