I’ve written quite a bit recently on communication, and in particular on how to help someone understand the idea you’re trying to convey. This is an incredibly valuable skill to develop, but we should note that it’s not the only skill tied to “communication” – for instance, being able to communicate concisely is it’s own virtue.
Today, I’d like to discuss a subtly different aspect of communication – how to ensure that what you’re saying is contentful. In other words, to ensure you are saying something that actually carries information.
I’ve previously mentioned the mathematical definition of “information”, but to rehash, information is that which reduces one’s uncertainty. In other words, something containing information must convey something which, one way or another, is surprising or otherwise difficult for the recipient to have predicted. After all, if they’d have predicted you’d say it so easily, what is learned by you actually doing so?
Of course, it’s easy to go overboard policing this. For instance, consider seeing your next-door neighbour and saying “how’s it going?” Your neighbour could have predicted this – after all, it’s how you always greet them – but, you might argue, it was still important to say it. After all, if you’d not said so, you would have been snubbing them!
And this is true! But the insight is that, in this case, your goal wasn’t to communicate any new information – it was to leave the world exactly as it was before, with you and your neighbour on friendly terms. There was no need to convey new information, because you didn’t want to change the state of the world. These sorts of phatic expressions serve a valuable social purpose, but I would argue they don’t really count as communication in the original sense, so I’ll be glossing over them for the rest of this post.
Beyond phatic examples, most speech conveys some information – and how much is driven by how much uncertainty it resolves. When I order a sandwich at the deli, I’m only conveying a little information – the deli clerk could have guessed I’d order one of a small number of things. This enables us to develop shorthand – perhaps they label their menu, and next time I just order the number 7. In contrast, one would have considerably more difficulty replacing all their discussions with coworkers with numbers – there’s far more range of what could be discussed, meaning more information is contained in those conversations (usually) than in a typical conversation with your deli clerk.
This may seem like a pretty trivial insight, but I’ve found it to be an incredibly valuable framing to help improve my communication in business. It’s often crucial to convey a lot of information in a limited amount of space or time; and the simple heuristic of “would the listener be surprised to hear me say these words” has proven a powerful way for me to optimize my messages.
Let me share some examples.
When discussing your company’s great culture, it’s often tempting to use high-handed language like “world-class team” or “making the world a better place.” These aren’t inherently bad things to have in a culture, of course, but the problem is that everybody says these things. Even if your culture really does involve making the world a better place, saying so in exactly those words will fail to convey information, because a reader would have predicted you’d say this regardless of whether or not it’s true.
(What, exactly, should you do instead? Well, if I suggested something here, it would lose some of its power when you use it – because the expectation of your readers that you’d say that, regardless of your actual culture, will have increased! So I can’t offer any magic bullets. But methods that have worked in the past include making public commitments to charitable giving, or permitting reports tell-all access to provide an unbiased perspective – which often communicates a lot before the tell-all report even comes out!)
When giving feedback, the goal is typically to help your colleague improve themselves. In order to do this, though, you need to share real information about what you feel they’ve done poorly at. I see this typically fall through in two ways. The first is simply when your colleague already knows what you’re telling them. For instance, a teammate struggles through giving a talk, visibly frazzled, and then you give them the feedback that they need to improve at public speaking. However well-intentioned this feedback is, it rarely helps, because your teammate is already aware of this shortcoming. (Of course, if you’re instead offering specific new suggestions for how they could train themselves, or emotional support for friend who has gotten too self-conscious, it could still be valuable to chat with them – but in this case you’re doing much more than giving simply feedback!) Thankfully, I’ve rarely seen feedback this minimal in practice – though when I have, it’s usually because the speaker is trying to fulfill some sort of mental “quota” of “how much feedback am I giving,” rather than because they genuinely were thinking of the person they’re supposed to be helping. (In this case, there is still information conveyed – it’s just “I’ve bothered to give you feedback”, rather than any actual useful insights.)
I think the other failure mode of feedback is more interesting, though. This is when you have real feedback to give, but fail to be sufficiently direct. Everyone has a model of themselves in their own head, and when we receive feedback, the first thing we do is compare that feedback to our model. If it’s all in alignment, we tell ourselves “great, I don’t have to change anything,” but if it differs from our self-impression, we have to admit to a shortcoming and strive to do better. Although most people recognize the value of the latter kind of growth, we all have defensive instincts which mean we search for the first kind of interpretation. So when someone gives feedback like “It seems like you struggled a little bit in that meeting,” they are risking failing to convey information. If the recipient responds to the feedback with “Oh, that’s interesting, I though I was doing well, tell me more,” then you’ve successfully shared information. But if they hear it and think “I know, the guy who was in charge of the meeting was incompetent, the whole thing was disorganized, it was so annoying”, then you’ve failed to actually convey any information that you think part of the problem might be from them.
To me, this is one of the hardest things about giving feedback. I always prefer to give constructive criticism as softly and openly as possible, but I find that it’s easiest to convey that through my attitude and caveats, rather than through softening the language of the actual feedback. This way, I can do my best to continue to serve as an ally of my colleague, but still maximize the chance that the information content of my feedback remains relatively undamaged.
When you interview a candidate, you might be focused on communicating about your culture as we discussed above – but the candidate is also focused on communicating information to you, about how they’ll perform well within your organization. Unfortunately, if every candidate tells you about how they are “really excited by your mission”, then none of them are succeeding at conveying any information about their relative fit – even if some of them actually are genuinely more excited than others.
This is an example of a particularly tough situation to convey information – the fact that most companies rely primarily on verbal interviews, and it’s trivial for some people to lie, means this situation goes full-on anti-inductive. Any sentence that is generally known to signal your fit with the company will immediately be co-opted by less honest applicants, making that sentence no longer trusted, meaning it no longer works, even for you.
Some companies attempt to avoid this through things like coding interviews or brainteasers which are harder to fake – in these cases, though, it’s important to remember that the valuable information being conveyed isn’t the code or the answer, it’s what that it says about the candidate that they got to that solution that way. I’ve met many candidates who technically failed to find the answer to a problem, but their process was strong enough (and their miss over something sufficiently superficial) that I advocated for them anyway.
Regardless of your tactic, it’s important to remember whenever you ask a candidate something that you’re looking for information, not just words. Craft your questions in an open-ended way that give candidates freedom to answer authentically and completely – even if some of the content they discuss isn’t directly relevant to the job, you’ll still learn vastly more about them than if you asked questions like “Is X on your resume?” or “What’s your biggest weakness?”
I could write a ton here about how investors are skeptical, and companies are incentivized to lie, so things like the classic “hockey stick” graph are taken with heaps of salt…but that would basically be a rehash of the earlier discussion about culture. Instead, I think it’s worth describing a different failure mode which comes up in fundraising. Before most investor conversations, the investor will ask for a pitch deck or some other materials. Based on that limited info, they’ll come up with an expectation of what your business is – at far more depth than what was actually written on the deck, since they’re filling in the gaps based on their own experience.
This can be good – with smart investors, it means you can skip the basics and go straight into the juicy stuff. But if you’re working with an unusual technology, business model, or market, there’s also a risk that the investor will expect the wrong thing. In this case, it becomes similar to the feedback discussion above – you need to make sure that you’re conveying any new information clearly enough that it throws into relief the fact that their model is wrong; rather than saying something fuzzy which risks them keeping their current model.
(Modulate faced this very acutely during our first fundraise. We used the phrase “voice clone” a lot, which isn’t inherently a bad phrase to describe what our tech does, but which had already been used significantly by text-to-speech companies. This led several investors to the incorrect belief that we were another TTS company, resulting in significant confusion when we began to elaborate on a business model and market which wouldn’t have fit a TTS org. This was a great way for us to learn the crucial lesson that our messaging needs to not only be concise and accurate – if we want to convey information about how we’re special, our messaging also has to be unique enough to differentiate us from other superficially similar companies.)
What do you think? Is this a useful framework to keep in mind for clear communication, or just a rehashing of an obvious concept? In other words, could you have predicted all the examples I gave above – or did I successfully communicate some new information?