I was doing a skip level meeting with one of my managers. She had described a desire to grow, and I asked her what she meant by that. She described a number of valuable soft skills that she respected in others that she wanted to develop herself, in particular how to engage with stakeholders in a confident and knowledgeable way and also how to quickly understand, synthesize, and respond to new information. It also became apparent that she didn’t know how to develop them beyond experience. This makes sense because, as soft as these skills may be, learning them is anything but soft.

What I recommended to her was that she plan on her own what her key messages are and refresh herself on the background material. Then she should talk to her manager about how to approach the meeting. Most importantly, I suggested they debrief after the meeting. She would solicit feedback about her handling of the meeting. Her manager could give her detailed feedback with obvious relevance and an understanding of specific actions to take in similar situations in the future. The reverse would be possible as well; she could ask her manager for an explanation of the insights they had and choices they made.

What’s essential here is the specific detail, and that’s why the immediacy of the follow-up discussion is so important, because everything will be fresh in their minds. The challenge with soft skills is that the underlying principles are all platitudes. “Know your audience.” “Listen carefully.” “Be open-minded.” Everyone knows what they should do in theory, but putting it into practice? People don’t understand what they don’t understand. They need it explained in context and in detail. Experience is certainly a critical component of the process, but without getting feedback and understanding your model, you’re forced to guess at the how and the why, which will dramatically slow learning.

Economists love the idea of “automatic stabilizers.” These are economic policies that go into effect when certain objective economic conditions appear. There’s no political negotiation or partisan fighting. It’s just smart, technocratic policy. And Congress will never do it. The cynical explanation is that they don’t want to give up power. The less cynical but still problematic explanation is that they have excessive confidence in their ability to make a good decision in the moment. That’s a mistake, and it’s a mistake you’re likely to make as well.

We’ve all been infuriated by rigid, stupid policies. They could be hiring policies, development policies, expense policies, etc. There’s something that we need to do in good faith to achieve our company’s goals, but we can’t because of a stupid policy. There are definitely stupid policies, but these should not be arguments against all policies.

Another word for a policy is “heuristic.” A policy is a conditional rule that tells you what to do in specific circumstances. It’s usually not something that’s driven solely from first principles; instead, it’s a pragmatic and possibly arbitrary construct. It can have three major redeeming qualities.

The first is to be simple. Well, not simple but simpler. The policy may be complicated, often more so than any given decision, but it’s likely not more complicated than the aggregate of all of the decisions in a particular category. This has the added benefit of making decisions faster and require less thought.

The second is that it’s clear. Ad hoc decision-making is inevitably going to messier and more opaque. In advance, you don’t know what’s going to happen. After the fact, it can be difficult to fully explain the decision.

Third and most important is that good policy is smarter than people’s ad hoc decisions. You can easily come up with situations where individual policies in specific situations have been stupid. That’s not the standard. The standard is the overall set of policies over all relevant situations. The problem is that we aren’t all that great at thinking on our feet. We get tired, hungry, distracted, distressed, etc. There are many factors that can compromise our judgment, but we’re too complacent about the effect those will have on us.

I’m not saying that policies will always make the best decisions. That shouldn’t be the goal of policies. What policies should do is simplify decision making and avoid the worst outcomes. Yes, you can often make a better decision than the policy. However, given our human fallibility, you can also make a decision that’s worse than the policy. Every decision you force yourself to make in the heat of the moment is one more decision you can make wrong. It’s one more decision that sucks up your mental bandwidth, which means it can also make other decisions worse. Yes, policy can be dumb. Make the policy better if that’s the case. Just don’t get rid of the idea of policy itself. It’ll make you dumber if you do.

We all know not to shoot the messenger. We don’t do a great job living up to it, but we know it in theory. Something people don’t seem to know even in theory is not to credit the messenger. By that I mean attributing success to the person who’s telling you about it. I see my twelve year old son do this, and I also see senior product directors do this. Sometimes it’s because someone is the point of contact for managing an incident. Sometimes it’s an ongoing effort, and you happen to interface with one person more than the rest of the team, often because they report to you.
Either way, since all information flows through them, it can be unclear who’s actually doing it.

What we can do in that situation is unconsciously think of that person as the one doing the work. We might even say that, either explicitly, like “Bill deployed a fix,” or implicitly, like “Sharon delivered the new workflow.” It’s usually a team effort. Of course, if you ask people, they’ll say they know that, but it’s not about what they know when you ask them. It’s about what they know when you’re not asking them. It’s about what other people hear.

When all the things you can do exceed the resources you have available, you have to pick and choose. In almost every case, you want to choose based on which ones have the highest return on investment. You quantify the potential benefit, quantify the cost, divide the former by the latter, and pick the ones with the biggest ratio. Easy. Except when it’s wrong.

The first situation where return on investment will lead you astray is when the investment is small. In those situations, don’t bother. There’s overhead and error involved in calculating the investment and the return. When the cost is low, the effort to quantify isn’t worth it because the most it can save you is only a very small amount. And that’s without getting into what might happen if you’re wrong.

The second situation is similar to the first in that incorporating RoI can only save you a small amount. This is when you’re choosing from a menu of different options that have different benefits and different costs, but the costs aren’t that different. Suppose one service has 99.99% reliability costing $50/year and the other has 99.9% costing $10/year. Suppose for your business you think 99.99% is twice as good as 99.9%. That means that the latter has 2½ times the RoI as the former. That seems like an easy answer, but we’re talking about $40/year here. Technically the first one costs five times as much as the second, but compared to most business expenses, they cost basically the same. When this happens, take cost out of the equation entirely. Don’t bother with return on investment, just focus on return. Otherwise you’ll risk being penny wise but pound foolish.

You’re going to screw up. I’m sorry. But you’re going to. And then you’re going to want to fix it. Sometimes you can, like fixing a major security vulnerability. Sometimes you can’t, like with spoiled work relationships. It’s not usually that hard to tell the difference. What is hard is giving up on the toast you’ve burned but learning how to toast it correctly the next time. If you can’t fix a mistake, sometimes you want to walk away completely. Then you don’t learn and burn the toast again. Or, if you’re trying to learn, you may also believe you can unburn this toast. You can’t. Give up. Cut your losses. And do better next time.

We’ve all heard “that’s not my job.” Even if people won’t say it, they’ll think it. That’s a limiting mindset. It starts from the perspective that what you do is defined by what you are explicitly responsible for. When there’s something that needs doing, you ask “do I have to?” In other words, what can you be blamed if you don’t do or do poorly?

The better way to think is based on what you have the capability of doing. Not “do I have to help” but instead “can I help?” Why is this better? Well, for one, it makes you less annoying. More importantly, gives you away to contribute more. It gives you a way to learn more. Independently doing what needs doing is an uncommon and valuable attribute. If you demonstrate it, you’re going to be a lot more valuable, and you’ll get a lot more freedom to decide how you work.

Inevitably the demands on you will exceed your capacity. It’s not you. You may have tremendous capacity, but it’s finite. The demands are nearly infinite. You have three basic strategies for dealing with this disparity between your responsibilities and your capacity.

Delegate: give the responsibility to someone else.

Degrade: do a less complete, lower quality job.

Drop: leave a task unfinished, a problem unsolved.

These are in order from easiest to hardest. You got to the position you’re in by doing a lot of things and doing them well. Doing things poorly and not doing them at all will go against your instincts and habits. However, they’re also in order from least important. To scale up, you have to be okay with lowering your standards. You have to be okay leaving important problems unsolved. If you can’t, then you’re creating a ceiling on what you’ll be able to achieve.

The transition from individual contributor to manager is one that many people struggle with. They understand that they now work with and through people. What they’re often not told is that their output is now something different. Instead of producing a product, you’re producing a better professional (that produces a better product). Those two things are vastly different. If you skip the professional and go straight to the product, you may get results in the short-term, but you’ll either stagnate or outright fail in the long term.

You manage complex people in a complex organization to do complex things. Invariably something is going to go wrong. Invariably the things that goes wrong is going to be because you made a bad decision. Invariably some of the time someone else knows it’s a bad decision from the start. The only question is when you will discover that it was a bad decision. Before you even commit to it? Or after the damage is done?

If the people on your team aren’t telling you things you don’t want to hear, it could mean that things are perfect and you’re perfect. It’s an incredible universe, and many things are possible. But what’s likely?

What’s more likely is that people aren’t telling you what you need to hear. They may be afraid of consequences. It may be a less punitive and more subtle inhibition, which is that they think it will be a hassle or futile. Either way, you’re not hearing the bad news when you should.

A sanity check is to pay attention to how often people question your decisions, disagree with you, and tell you you’re wrong. I don’t know how often is the right amount, but I have a few of these every week with each of my reports. It happens with important things and with minor things. I don’t know if they’re telling me about all the times they disagree with me, but they do it often enough that I’m confident they won’t let me make a major mistake. That’s not perfect, but it’s pretty good.

Of course, if you’re following my advice, you’re asking your reports, “how am I doing?” And then you find out without having to infer it. Odds are pretty good you’re not doing that, though, because it’s awkward and uncomfortable. Paying attention to the bad news will at least get you part of the way there.

The purpose of interviewing is predicting future job performance. Applying that well means focusing as much as possible on gathering information. You want to discover everything you can about the candidate that is relevant to the job.

Only when you’ve gathered as much information as possible do you shift into evaluation mode. If you evaluate too early, you’ll get less curious. You’ll be likely to confirm your biases. You may be too narrow in considering the role for the candidate.

What you’re doing is mapping the shape of a key. You identify the points and the grooves. Then you try to see if it fits the lock. If you map the key well, you can test it on multiple locks. It may fit several, one, or none. But if you only map the key by testing it against the one lock, you won’t fully profile it, and you won’t be able to test it against other locks.