Welcome back to the paid edition of the newsletter, and thank you again for being a subscriber.
I keep this newsletter very simple. There is a free article and a paid article every month, and I do no interviews, sponsorships, or anything that would move the content away from the core proposition, which is that we get to explore interesting ideas together. Being a paid subscriber is no small ask, so I try to ensure that you get something long, deep and meaningful each month.
So, truth.
Recently, we were looking at adoption and retention metrics for a feature that had been shipped and enabled for all customers. The status report was accurate as far as it went, because the feature was live, but the adoption was awful.
When we dug in with the team and looked at the UI by walking through common user journeys, we found the sidebar wasn’t sliding out by default, which meant the feature was effectively hidden from users. Shipped was, therefore, true. However, findable by users was not: barely anybody actually used it!
Facepalm…
As you lead larger teams or organisations, the default is to become distant from the details. More people sit between you and the code, the space between you and the user increases, and what’s being built often reaches you as a finished product rather than something that you steered yourself. Left to entropy, this is the natural progression of the management role as scope expands. Yet, what is expected of us now is the opposite. We need to be in the details.
In September’s free article, Nobody needs a human router, I argued that closing this distance is not as hard as it used to be. AI has collapsed the overhead of being in the details, which means the technical judgement you built as an individual contributor, with some adaptation to how you run your role, is now very applicable. Getting your hands dirty again is therefore possible and increasingly necessary, and it doesn’t require you to stop leading either.
But, being in the details rests on an assumption: when you dig in, what you find has to be real. You have to know the truth. But what do we mean by that?
In some organisations the truth of every project, and the quality of the code, survives as it is communicated up to you. In others, it’s filtered, or even stealthily laundered to make sure you don’t see the worst of what’s going on.
This piece is about what organisations do to the truth on its way to the leader, and how to design a system and a way of working that doesn’t distort that truth before it reaches you.
Here’s what we’re going to cover:
The four forces that corrupt truth. We seek some framing from Warren Buffett, who wrote about this very issue many decades ago. How do you actually know what’s going on in your teams? Is it true?
What does “good” actually mean? Jeff Bezos noted that humans are not wired for truth-telling by default, as our nature is to please others. Being able to seek truth in your organisation requires a somewhat unnatural culture.
Structural accountability. We look at ways in which you can create processes that require leaders to see and understand the details by making them accountable for them.
If you’d like to dig deeper, here are some related articles from the archive:
Being in the details set the baseline argument for proximity to the work.
Going direct extended it to communication flow inside flatter organisations.
Writing is culture argued that written artefacts are a great mechanism for surfacing and archiving, which is exactly what this article is about.
Fewer envelopes, more letters applied this thinking to the kinds of problems you work on day to day: are you an envelope or a letter leader?
Alright, let’s dig in.
How do you really know what’s going on?
Given that the higher up you go in an organisation, the harder it is to see the details, how do you fight against this?
The default answer for decades has been to build communication structures around it: delegate more, coordinate more, push information upward through layers, and then trust the processes you put in place to do their job.
This, in turn, has increased the amount of information routing that middle management has had to do: the kind that we explored in the accompanying free article this month.
However, the industry has been refactoring middle management, through layoffs and a flattening that minimises the role of coordination and information passing in the age of AI. This is big companies pushing towards “efficiency” (however you may interpret that), and cutting layers is one of the easiest ways to get there: fewer managers, more direct reports, less money going out on payroll.
For those of you that have worked in big or bureaucratic organisations, you’ll know that layers in the org chart aren’t perfect relays, but distortion machines. Every interface between the leader and the work is a place where truth can be softened or reshaped to look better than it is.
Sometimes this is deliberate, with people protecting themselves or covering up things that aren’t going so well: a project running late because it turned out harder than expected, or a team that hasn’t performed as well as hoped. Sometimes it’s more innocent, such as someone’s innate people-pleasing personality, or just communication mistakes. Either way, this distortion blurs the truth, which means that leaders neither know what is going on, nor do they know what they have to do to fix the hidden problems.
So what actually is truth? For a software company in particular, I would argue the ultimate truth lives in the code. You take customers’ problems as input and ship software that solves them. Your customers are happy if you do this, and they aren’t if you don’t.
Everything else that flows through the organisation is a projection of that code, or some intermediate step to it: designs, prototypes, status reports and board slides. The further you sit from the code, the more you’re reading projections of projections of intermediate steps. None of these are the actual truth.
Truth-seeking often requires acting in a novel way. Take Jim Casey, a co-founder of UPS (as told in this Founders Podcast episode). Travelling between exec meetings, he would ask his driver to stop whenever he saw a UPS delivery in progress, get out, and just go and talk to the courier directly about their work without identifying himself.
These “person on the street” interviews worked because a driver on the ground could tell him things about the state of the operation that a manager’s filtered summary could not. He just went straight to the truth.
Talking to Lex Fridman, Jeff Bezos put it plainly: “any high performing organization has to have mechanisms and a culture that supports truth-telling.” I think of achieving that as building a truth-seeking organisation, which is what all modern leaders (you!) should aim for.
The goal, therefore, is no longer to scale yourself through layers and information passing. Instead, the goal is to reduce the number of interfaces between you and the underlying truth, and make sure that what you see, do, and act upon is real, in the same way that Jim Casey spoke to his delivery drivers, or how Elon Musk walks the Tesla and SpaceX factory floors.
What you’re looking for is the truth. What’s in the way is the challenge to solve, and, like many things worth thinking about, it happens Warren Buffett wrote about this in detail some time ago.
The four forces that corrupt truth
In his 1989 letter to Berkshire Hathaway shareholders, Warren Buffett named what he described as his “most surprising discovery in business”: the institutional imperative. As he put it, “rationality frequently wilts when the institutional imperative comes into play.” Basically, as companies increase in size, intelligent people start to act in less intelligent, herd-like ways. I’m sure we’ve all felt similarly around large groups of people.
No matter how smart the people involved, organisations of increasing size exert an increasing pull that bends decisions away from what’s rational and true. This happens through thousands of tiny actions and decisions every single day.
Buffett named four behaviours that move seemingly smart companies towards irrationality and falsehood:
Institutions resist change. Strategic commitments and projects continue past the point they should be stopped, because inertia has built around them.
Projects expand to absorb available capacity. Budget gets spent and headcount gets occupied, simply because they are there, regardless of whether the work being funded is the right work.
What a leader says, they often get. Buffett’s own wording is that “any business craving of the leader, however foolish, will be quickly supported by detailed rate-of-return and strategic studies prepared by his troops”. This phrasing is kind of weird, so it might need an example.
Imagine a CTO who mentions in passing that she’s excited about a particular technology she’d been reading about over the weekend. Some weeks later, a long design document arrives recommending full-scale migration to it, complete with cost analysis and migration strategy, authored by people who were not excited about it the day before she mentioned it. One senior person’s whim has suddenly manufactured six months of unnecessary work.
When I was at Shopify, this happened all the time. People would say “Tobi said X, therefore we are now building X”. That naturally happens at any company where a leader is trusted and respected, but it’s also how unbelievably large side quests start because you happened to say you liked an article.
