The Email You Should Have Sent: Communication That Actually Works at Work

Pax

Pax

April 4, 2026 · 14 min read

You’ve been staring at this email for twelve minutes.

You’ve written it, deleted the second paragraph, rewritten it, softened the opening, hardened it again, removed an exclamation point, added one back somewhere else, and you’re now wondering whether “just to clarify” sounds passive-aggressive or if that’s just you projecting because you already know you’re annoyed.

If you’ve ever burned a quarter of your afternoon composing three sentences to a colleague, you don’t need me to tell you that workplace communication is hard. But it might help to understand why it’s hard in ways that kitchen-table arguments and friendship conflicts are not. Because the difficulty isn’t a personal failing. It’s structural.

Why Workplace Communication Is Uniquely Hard

Three forces make professional communication a distinct species of difficulty.

Power asymmetry. You cannot be fully honest with someone who controls your livelihood, and pretending otherwise is naive. When your manager asks “what do you think?” they are often asking a question with an invisible frame around the acceptable answers. You feel that frame even when it isn’t there. Research by organizational psychologist Amy Edmondson at Harvard shows that psychological safety (the belief that you won’t be punished for speaking up) is one of the strongest predictors of team performance, a finding later validated at scale by Google’s Project Aristotle. And most teams don’t have it. Not because managers are tyrants, but because hierarchy exerts a gravitational pull on honesty that neither side fully sees.

Permanence. Spoken words fade. Emails don’t. Every message you send becomes a document — forwardable, screenshottable, discoverable. You aren’t just talking to the person on the To line. You’re talking to every possible future reader, including the one in HR. That awareness, even subconsciously, makes people write defensively rather than clearly.

The tone vacuum. You’ve probably heard that 93% of communication is nonverbal. That number comes from Albert Mehrabian’s research at UCLA in the 1960s, and it’s often misquoted. What Mehrabian actually found was narrower: when a message is ambiguous, listeners rely on tone of voice and facial expression to decode meaning. His study was about emotional messages, not all communication. But email is almost always ambiguous. Without tone, without facial cues, your reader fills the gap with their own assumptions. And people under workplace stress tend to assume the worst. The neutral email gets read as cold. The direct one as aggressive. The friendly one as sarcastic. You’re not just fighting the message. You’re fighting the medium.

The Feedback Sandwich Is Dead (What It Was Trying to Do Isn’t)

For two decades, management training programs taught the “feedback sandwich”: say something positive, deliver the criticism, close with something positive. The idea was that the compliments would cushion the blow.

The degenerate version of this (hollow praise, real criticism, hollow praise) deserves its bad reputation. When you open with “Great job on the Henderson presentation,” your employee’s brain doesn’t hear a compliment. It hears a wind-up. They’re bracing for the “but” before you’ve finished the sentence. The opening praise feels manufactured because it is manufactured. It’s there to serve the structure, not because this was the moment you needed to express admiration. The criticism arrives pre-discounted. And the closing compliment lands like a pat on the head. Sheila Heen and Douglas Stone lay this out in Thanks for the Feedback: receivers filter for the negative regardless of packaging. The bread doesn’t hide the meat.

But here’s what gets lost in the backlash: the sandwich was solving a real problem. A lot of managers, especially under time pressure, default to only surfacing what’s broken. They walk into a review with a list of deficiencies, work through it, and leave. The employee walks out feeling like their entire contribution has been reduced to a problem report. The sandwich, whatever its flaws, imposed a discipline: before you tell someone what’s wrong, you have to do the work of noticing what’s right. That’s a cognitive exercise for the person giving feedback, not mere packaging.

So the issue isn’t structure versus no structure. It’s whether the person giving feedback has done the internal work of caring about the recipient’s growth. The sandwich was a crude tool for approximating that care structurally. Direct feedback without that underlying orientation is just efficient cruelty.

What works better is direct, specific, future-oriented feedback delivered by someone who sees the whole person. Name the behavior (not the person), name the impact (not your judgment), and name what you’d like to see next. But don’t stop acknowledging strengths — just stop performing them on cue as anesthetic. Genuine recognition, given separately and on its own terms, lands far harder than praise that exists to set up a “but.” Kim Scott’s Radical Candor framework calls this “caring personally while challenging directly” (worth reading if you manage people).

The mechanical sandwich: “You’re such a strong writer, but the client proposal had several errors that made us look unprofessional. Keep up the great work though!”

What actually works: “The client proposal had three factual errors in the pricing section. I want to flag this because the client noticed, and it affects their confidence in our numbers. Can we set up a review step before the next one goes out?”

The second version is more direct and, paradoxically, kinder. It names a specific problem, explains why it matters, and invites a solution. There’s nothing to decode. Nothing to resent. And it leaves room for genuine recognition to happen on its own, when it’s actually earned and actually meant.

Five Emails, Rewritten

This is where the theory meets your inbox. Each scenario shows the version most people send, explains why it misfires, and offers a rewrite.

1. Disagreeing With Your Boss’s Decision

What most people send:

Hi Jordan,

Sounds good, let’s go with that approach.

Or, on a bad day:

I’m not sure that’s the right call. The data doesn’t support it and I think we’re going to run into problems down the line.

The first version is dishonest: you’re going to spend three months executing a plan you don’t believe in while quietly resenting it. The second is honest but positional. It draws a line and dares your boss to cross it. Most bosses will, because that’s what hierarchy does when it feels challenged.

The rewrite:

Hi Jordan,

I’ve been thinking about the new rollout timeline. I’m on board with the goal of launching faster — I think that’s the right priority. I have a concern about the resource allocation in weeks 3-4 that I’d like to think through with you. Specifically, if we pull the QA team early, we might hit the same issue we ran into with the Kendall project last quarter.

Before we lock in the plan, I’d like to walk through that scenario with you. I want to make sure we’re setting the team up to hit the date without the fire drills.

Why this works: it leads with alignment on the shared goal, raises the concern as a question rather than a declaration, and cites a specific precedent instead of a vague fear. You’re not saying “you’re wrong.” You’re saying “I want to help us get this right.”

2. Giving Feedback to a Peer Who Isn’t Pulling Their Weight

What most people send:

Hey, just wanted to check in — I’ve noticed I’ve been handling a lot of the deliverables for the Maxwell account lately. It would be great if we could balance that out going forward.

This is so hedged it barely registers. “Just wanted to check in” signals discomfort. “A lot of the deliverables” is vague enough to be deniable. “It would be great” is a wish, not a request. Nothing changes.

The rewrite:

Hey Sam,

I want to flag something on the Maxwell account before it becomes a bigger issue. Over the past three weeks, I’ve handled the status reports, the client calls, and the budget reconciliation solo. I don’t think that’s sustainable, and I don’t think it’s the split we agreed on.

Can we look at the task list together this week and re-divide? I want to make sure we’re both set up to deliver without either of us burning out.

Why this works: it’s specific (three tasks, three weeks), names the impact (unsustainable), references a prior agreement, and proposes a concrete action. You’re not saying “you’re not doing your job.” You’re saying “our arrangement isn’t working and I’d like to fix it with you.”

3. Responding to a Passive-Aggressive Message

Your colleague sends:

Per my last email, the deadline was Friday. Just want to make sure we’re aligned on expectations going forward.

Your first instinct is to match the tone. To reply with your own “per my last email” and a cc to your mutual manager. Don’t.

What most people send:

As I mentioned, I was waiting on the data from your team, which I received Monday. Happy to discuss expectations around dependencies if that would be helpful.

This is passive-aggressive judo. It redirects blame, deploys the weaponized “happy to discuss,” and escalates without appearing to escalate. Neither of you will trust the other next quarter.

The rewrite:

Hi Alex,

You’re right that I missed the Friday deadline — I want to own that. The delay was partly because the dataset arrived Monday, but I should have flagged the risk earlier instead of waiting.

Going forward, I’ll flag any dependency delays by Wednesday so we can adjust together. I’ll plan on that unless you want to discuss.

This lands because it breaks the passive-aggressive cycle by being direct. Owning the miss disarms the other person’s need to escalate. The forward-looking commitment gives them what they actually wanted: not a fight, but a guarantee it won’t happen again.

A note on this one: “Own your miss” is easier advice to give than to follow, and the risk of following it is not equally distributed. Research on gender and racial dynamics in performance evaluation consistently shows that admitting fault can carry disproportionate consequences depending on who you are and what your workplace is like. If your environment punishes vulnerability rather than rewarding it, the calculus changes. The principle (break the cycle, focus forward) still holds, but the tactical execution might look different. Use your judgment about what’s safe in your context.

The next two scenarios are the ones people tell me they struggle with most. They require a kind of honesty that has nothing to do with the other person and everything to do with you.

4. Saying No to More Work Without Seeming Uncommitted

Your manager sends: “Can you take on the Riverside analysis? Need it by Thursday.”

What most people send:

Option A: “Sure, I’ll make it work.” (You won’t. Or you will, at the cost of something else falling through.)

Option B: “I’m really slammed right now, I don’t think I can take that on.” (This reads as unwilling, even if accurate. No manager wants to hear no without understanding why.)

The rewrite:

Hi Robin,

I can take on the Riverside analysis. Right now I’m working on the Quarterly review (due Wednesday) and the Chen onboarding materials (due Friday). If Riverside needs to be done by Thursday, which of these should I deprioritize, or would it help if I handed one off?

Want to make sure we’re putting the hours where they matter most.

The move here: you said yes and made the tradeoff visible. You gave your manager the information they need to make a resource decision, which is their job. Most of the time, they’ll either extend the deadline, reassign something, or tell you which ball to drop. You look like someone who manages their capacity honestly, which is far more valuable than someone who says yes to everything and quietly drowns.

5. Addressing Being Talked Over in Meetings

This one is delicate because it’s about a pattern, not an incident, and patterns are easy to deny.

What most people send (after the meeting, frustrated):

I’d appreciate it if you could let me finish my points in meetings. It’s happened several times now and it’s really frustrating.

This is vague enough to be dismissed (“I don’t think I do that”) and emotionally loaded enough to make the other person defensive. It names the frustration but not the stakes.

The rewrite:

Hi Taylor,

I want to bring up something I’ve noticed in our last few team meetings. A few times when I’ve been mid-point, the conversation has moved on before I could finish — including Tuesday’s discussion about the vendor contract, where I had concerns about the liability clause that didn’t get heard.

I don’t think it’s intentional, and I know our meetings run fast. But I want to make sure my input is landing, especially on decisions that affect my projects. I propose we build in a quick round-the-table check before we close out agenda items.

Why this works: it cites a specific instance (Tuesday, vendor contract, liability clause), making the pattern undeniable without being accusatory. And it proposes a structural solution rather than asking the person to change their behavior through willpower alone. Structural solutions outlast good intentions.

The 24-Hour Rule

I once watched someone send an email at 11:47 PM that cost them a promotion. Not because the content was wrong; every fact in it was accurate. But the tone, written after a ten-hour day of being second-guessed, turned a legitimate concern into an indictment. The same email, written the next morning, would have been a career move. Instead it was a career lesson.

Not every email needs the treatment above. Some emails are about logistics, deadlines, updates. Send those immediately. Speed is a professional virtue when the content is operational.

But when the email is about a person (their performance, their behavior, a conflict between you), apply the 24-hour rule. Write the draft. Save it. Sleep on it. Reread it in the morning.

There’s a reason this works. After a full day of decisions, trade-offs, and social navigation, you have less patience for nuance. Whether you call it decision fatigue or just being tired, the effect is real and familiar: your 11 PM draft is almost always sharper, more self-righteous, and less effective than what you’d write fresh.

The heuristic is simple: if the email is about a task, send it. If it’s about a person, wait.

You’ll lose very few things by waiting a few hours. You’ll avoid a lot of things you can’t unsend.

When to Stop Writing and Start Talking

Some conversations should never happen over email at all.

If the message requires tone to land correctly, pick up the phone. If the content could be screenshotted and forwarded out of context, and the prospect makes your stomach drop, have the conversation in person. If you need to hear the other person’s voice to know whether the two of you are okay, no amount of careful prose will substitute for a five-minute call.

Consider this real pattern: two colleagues spend four days exchanging increasingly careful emails about a project disagreement. Each reply takes an hour to compose. Each one is read three times, with growing dread. By Thursday, both are convinced the other is acting in bad faith. They finally get on a call. Within ten minutes, they discover they actually agree on the goal and were arguing about a misunderstood assumption in email number two. Four days of work tension, resolved by one short conversation where they could hear each other’s tone and ask clarifying questions in real time.

Email is excellent for delivering information. It is terrible for resolving ambiguity, building trust, or navigating emotion. Know which one you’re doing.

A useful test: if you’ve rewritten the email more than three times, it probably shouldn’t be an email.

The Instinct You Already Have

If you’ve ever rewritten an email six times before sending it, you already care more about this than most people in your office. That instinct, the one that says how I say this matters, is an asset, not a liability.

You pause. You rewrite. You wonder how it will land. That’s not overthinking. That’s the work.

The email you should have sent isn’t the one with better vocabulary or a smarter subject line. It’s the one that says what you actually mean, directly enough to be understood and carefully enough to preserve the relationship you’ll need tomorrow.

Start with the next email in your drafts folder. The one you’ve been avoiding. Does it say what you mean? Or does it say what your frustration suggested?

Micro-action: Pick one email in your drafts or sent folder from the past week. Rewrite it using the principles above: specific, impact-focused, future-oriented. You don’t have to send the new version. The practice is the point.

Speak kindly. Listen closely. Paws often.

— Pax

Thanks for reading. If you'd like Pax delivered to your inbox, subscribe on Substack — new posts arrive there a day or two after they appear here.

Further Reading

Books referenced in this post:

Related posts: