Seeing Without Surveilling: Visibility for Remote Managers

In an office, managers get an enormous amount of information for free. You see who's at their desk wrestling with something, you overhear that the vendor call went badly, you notice that the usually chatty designer has been quiet for a week. None of this required a meeting or a report. It was ambient — low-resolution, often misleading, but free.

Remote work shut that channel off, and managers responded in one of two bad ways. Some bought surveillance: screenshot software, keystroke loggers, activity scores, green-dot vigilance. Others became human polling loops: "quick check-in?", "how's it going?", "any update on this?" — a dozen interrupts a day per report. Both approaches damage the thing they're trying to protect, and neither actually answers the questions a manager needs answered.

There's a third way, and it's better than what the office gave you: pull-based visibility. The team's work state lives in artifacts — boards, check-ins, status pages — that you read on your schedule, without interrupting anyone. This post explains why the first two approaches fail, then builds the pull-based system instrument by instrument.

What "visibility" actually means

"I need more visibility" is a vague complaint, and vague complaints get solved with blunt tools. Decompose it. A manager's legitimate information needs reduce to five questions:

1. Is the work on track? Which projects are healthy, which are at risk, and did anything change since last week?
2. Is anyone blocked? Who is stuck, on what, and what would unstick them?
3. Is anyone struggling? Overload, confusion, disengagement — the human signals the office used to leak.
4. What shipped? What actually got finished, so you can recognize it and report it upward.
5. Where is decision input needed? What's waiting on you, specifically?

Notice what's not on the list: what is each person doing right now, hour by hour. No management decision you make actually depends on that information. You don't reassign work, coach, escalate, or plan based on whether someone's mouse moved at 2:15pm. Real-time activity data is the one thing surveillance tools provide and the one thing you don't need. Meanwhile the five questions above — which you genuinely do need answered — are exactly what activity data can't tell you. A monitored employee can be visibly "active" for eight hours while blocked, miserable, and building the wrong thing.

Write the five questions down. Every visibility mechanism in this post exists to answer one of them, and any mechanism that answers none of them is theater.

Why surveillance backfires

The monitoring-software pitch is seductive: dashboards, activity scores, "productivity insights." Here's what actually happens when it's deployed, and why the failure is structural rather than a matter of picking a nicer tool.

You get performance of work, not work. People are excellent at optimizing whatever is measured. Measure activity and you will get activity: mouse jigglers, strategically timed messages, documents left open. Studies of workplace monitoring consistently find that monitored workers shift effort toward observable busyness and away from hard-to-observe deep work — which is precisely the work you most want done. You've paid money to make your team worse at their jobs and better at pantomime.

You convert intrinsic motivation into compliance. The message monitoring sends is unambiguous: we assume you'd slack if unwatched. People treated as untrustworthy respond by doing exactly what's required and nothing more. The discretionary effort — the extra pass on the design, the bug fixed while nobody asked — evaporates first, because discretionary effort is definitionally the part you can't compel.

Your best people leave first. Senior, in-demand employees have options and low tolerance for being screenshot every ten minutes. Adverse selection does the rest: surveillance-heavy shops gradually keep the people who can't leave.

The anxiety tax hits everyone, including the diligent. Green-dot anxiety is real and measurable in behavior: people afraid to step away for lunch, performative "back!" messages, evening messages sent to prove dedication. That chronic low-grade stress degrades exactly the cognitive capacities — focus, creativity, judgment — you hired these people for. And it flows both directions: managers with activity dashboards report more anxiety, not less, because the data stream is endless, ambiguous, and demands interpretation. Was that quiet hour lunch, deep work, or disengagement? The dashboard can't say, so the manager stews.

The polling-loop manager avoids the spyware but lands in a milder version of the same trap. Every "how's it going?" ping costs the report a context switch — task-switching research puts full recovery at 20 minutes or more — and communicates, gently, I wasn't sure you were on it. Multiply by seven reports and a few pings each per week and you're spending hours of team focus to collect stale, verbal, unrecorded status. Push-based visibility is surveillance with better manners and worse data.

The core model: push versus pull

Every visibility mechanism is one of two kinds.

Push: the manager initiates; the report interrupts their work to answer. Pings, surprise syncs, "quick calls," status meetings that exist so the manager can hear updates aloud.

Pull: the work state is written into a shared artifact as a byproduct of doing the work, and the manager reads it whenever they want. Boards, check-ins, status pages, changelogs.

Pull wins on every axis that matters:

Property	Push (pings, syncs)	Pull (artifacts)
Cost to the report	A context switch per ask	Near zero — the artifact is part of the work
Freshness	Stale the moment the call ends	As current as the artifact
Record	Verbal, lost, disputed later	Written, timestamped, skimmable history
Scales with team size	Linearly worse	Flat — reading 9 rows costs what reading 5 does
Signal to the team	"I need to check on you"	"The work speaks; I read"
Serves upward reporting	You re-type what you heard	You link the artifact

The deal you offer your team is explicit and two-sided: "I will never ping you for information that's already in the system. In exchange, keep the system honest and current." Autonomy is what the team buys; visibility is what you buy; the artifact is the currency. This is the same trade that powers accountability without micromanagement — visibility and trust are one system wearing two hats.

The four instruments

Four artifact types cover all five manager questions. Most teams need all four; none of them requires a meeting.

1. Boards: the state of the work

A shared board — kanban, table, whatever fits the work — answers "is the work on track?" and "what shipped?" passively, provided three hygiene rules hold:

• Everything real is on the board. Work that happens off-board is invisible work, and invisible work is what triggers pinging. If someone spent two days on it, it's a card.
• Cards move when reality moves, not in a pre-meeting cleanup frenzy. The board is a live instrument, not a stage set.
• Cards carry owner and date. A board where you can't see who and when is a mood board.

Reading a board is a skill. Don't read cards; read patterns: What hasn't moved in a week? (Stall — possibly a silent blocker.) What's ballooning in the in-progress column? (Overcommitment.) What's sitting in review? (You, possibly, are the bottleneck.) Five minutes of pattern-reading each morning replaces a day's worth of "any update?"

2. Async check-ins: the state of the people

Boards show work; they don't show humans. The instrument for questions 2 and 3 — blocked? struggling? — is a lightweight async check-in: two to four prompts, answered in writing on a schedule, visible to the team. A proven daily or twice-weekly set:

1. What's your main focus since last check-in?
2. Anything blocking you or slowing you down?
3. How are you doing? (a simple green / yellow / red)

The mood question earns its place. People will type a yellow dot weeks before they'll schedule a meeting to say "I'm drowning." Trend it: one yellow is a Tuesday; three consecutive yellows from someone who's been green for months is a conversation you initiate — privately, and soon. Openbook's Check-in room is built around exactly this loop — scheduled prompts, mood tracking, and flags like Blocked or Help that jump out of the digest — but the format works in any tool that makes answers written, visible, and trendable. The craft of designing prompts and cadence is covered in Team Check-ins: Replacing Status Meetings with Async Pulses, and the standup-specific variant follows the same rules.

The non-negotiable norm: blockers get responses. A check-in where someone writes "blocked on the API keys for two days" and nothing happens teaches everyone to stop writing real answers. As the manager, your half of the ritual is reading daily and acting on flags within a day. A check-in system where writing "help" summons help is self-reinforcing; one where it summons silence dies in three weeks.

3. Status and goals: the state of the projects

For anything longer than a sprint, add a per-project status artifact updated weekly by the project owner: a short narrative (what happened, what's next, what's at risk) plus goal-level red/amber/green. This answers question 1 at the altitude your own boss cares about, and it's your question-5 surface too — a good status update ends with "decisions needed," which is where you find out what's waiting on you.

The failure mode to design against is watermelon reporting — green outside, red inside. You prevent it with tone, not tooling: the first amber must be visibly cheaper for its author than a late surprise would have been. Respond to risk flags with help, never with theater of disappointment. The mechanics of honest RAG reporting get a full treatment in Project Status Reports People Actually Read.

4. Shipped work: the state of the output

Finally, make finishing visible. A weekly "shipped" thread, a demo recording, a changelog channel — some ritual where done work gets shown. This answers question 4, feeds your recognition habit (praise specific shipped things, publicly), and quietly rebalances attention toward outcomes. Teams that only track in-progress work develop a strange gravity where everything is always 80% done; a shipping ritual is the antidote.

Your reading routine

Pull systems fail when managers don't actually pull. If the team writes check-ins nobody reads, you've added busywork and subtracted trust — strictly worse than nothing. So treat reading as a scheduled part of your job, not an ambient intention:

Cadence	Time	What you read	What you act on
Daily (morning)	10 min	Check-ins, board deltas	Blockers and Help flags — respond same day
Weekly (before team review)	20 min	Project statuses, goal RAG, shipped list	Ambers needing decisions; wins to recognize
Monthly	30 min	Trends: mood over time, cycle times, stall patterns	Workload rebalancing, process fixes, private conversations

Two habits multiply the routine's value. Acknowledge in the artifact: react to check-ins, comment on cards, answer questions where they were asked. Every visible acknowledgment teaches the team the system is load-bearing. Arrive pre-read: when you do have your weekly team meeting or one-on-ones, don't spend them collecting status you could have read — spend them on the ambers, the decisions, and the human conversation. Meetings shrink; their density goes up.

A note on dashboards: rolling boards and statuses up into a summary view — RAG counts, throughput, open blockers — is genuinely useful at the monthly altitude and for reporting upward. Just keep dashboards pointed at work states, never at people activity. The moment a dashboard has a per-person activity number on it, you've rebuilt surveillance with prettier charts.

The norms that keep it from curdling

Pull-based visibility runs on psychological safety, and three norms protect it:

Never weaponize the artifacts. The history in a check-in or board exists to coordinate work, not to prosecute people. The first time a manager pulls up six months of check-ins in a performance review as a gotcha — "on March 3rd you said you were on track" — honest writing ends, permanently. Performance conversations draw on patterns you observed and discussed contemporaneously, not on forensic artifact audits.

Yellow is protected speech. Anyone who flags risk, admits being stuck, or reports a red mood gets help or gets space — never a penalty, never a public spotlight they didn't ask for. You will be tested on this within the first month, and the team will calibrate on the outcome. Handle mood signals with particular care: respond privately, ask open questions, and let them decline to elaborate. Mood data is an early-warning system, not a diagnosis.

Visibility is symmetric. You write a check-in too. Your projects have statuses. Your commitments sit on the same board. A manager who reads everyone's state while publishing none of their own is running a one-way mirror, and one-way mirrors breed exactly the guardedness the system is meant to dissolve. Symmetric visibility also solves your own upward problem: when your boss asks how things are going, you link the status page instead of assembling a Friday-afternoon deck from memory. Pull-based visibility compounds up the org chart.

Calibrating by person and situation

Pull-based visibility is a default, not a dogma. Two variables legitimately change the settings.

Seniority and tenure. A new hire in week two doesn't need less contact; they need more — but scheduled, not surveilling. The right pattern for someone new is a daily 15-minute sync for the first couple of weeks, decaying to twice-weekly, then weekly, with the pull instruments carrying more of the load at each step. Say the trajectory out loud so the contact reads as onboarding, not distrust: "We'll talk daily for two weeks while you get context, then taper as the check-ins take over." A junior engineer six months in might warrant a standing mid-week pairing block — again scheduled, again framed as support. What juniors should never get is the covert version: the manager who quietly watches their commit timestamps and Slack activity because "they're still proving themselves." If you need more signal, add structure in the open.

Genuine performance concerns. When someone's reliability has actually broken down — silent misses, stale artifacts, weeks of vague statuses — the answer is tighter explicit structure, not tighter watching: shorter checkpoints, smaller commitments, more frequent scheduled reviews, all named as what they are and all temporary. "Let's do twice-weekly checkpoints until we're back on rhythm" is management. Quietly monitoring their activity while saying nothing is surveillance, and it also doesn't work — you'll gather ambiguous data for a month while the actual conversation goes unhad.

Crisis periods flex the system too. During an incident or a launch week, synchronous contact legitimately spikes — war rooms, frequent syncs, live updates. That's fine; urgency is real. The tell of a healthy system is that the spike ends. Teams where every week is treated as launch week don't have a visibility system; they have a chronic emergency.

Objections you'll hear, and answers

"Some people will abuse the trust." Occasionally, yes — and the pull system catches them faster than surveillance does. Someone coasting leaves a distinctive artifact trail: cards that don't move, check-ins that go vague, statuses that repeat last week's words, nothing in the shipped thread. That pattern is visible within two weeks and is far better evidence for a direct conversation than an activity score, because it's about outcomes nobody disputes. Surveillance, by contrast, is trivially gamed by exactly the people you'd deploy it against — mouse jigglers are a solved problem — while taxing everyone else. Design the system for the 95% who are trying, and let the artifact trail surface the rest.

"My leadership demands monitoring data." Then give leadership better data. Execs asking for activity dashboards are usually asking a legitimate question — "is the work on track?" — through a bad proxy. A portfolio view of project RAG statuses, throughput, and shipped outcomes answers the real question more credibly than screenshot counts, and most leaders take the trade when shown both side by side. If your company mandates monitoring software regardless, be transparent with your team about exactly what it collects and how it's used; covert monitoring, beyond being corrosive, is legally restricted in many jurisdictions.

"Writing all this down is overhead." Count both sides of the ledger. A three-prompt check-in costs five minutes; a status update costs fifteen minutes a week per project. Against that: the status meetings you cancel (30–60 minutes per person per week), the pings you stop fielding, the "what's the latest on X" archaeology that a written trail makes instant. Teams that run this honestly for a month almost always find the written system is the cheaper one — and unlike meeting-based status, the writing has residual value: it's searchable, linkable, and readable by the next person who joins.

"My team is five people; we just talk." At five, informal can work — until one person goes remote, one goes on leave, and the sixth joins. The artifacts are also what makes growth survivable: a team that writes at five scales to fifteen without a communication crisis. Start lighter — a board and a weekly check-in may be plenty — but start.

Rolling it out

If your team currently runs on pings and status meetings, migrate in this order — each step funds the next:

1. Week 1: Stand up the board. Get all real work onto it with owners and dates. Announce the deal: no pings for anything the board answers.
2. Week 2: Start check-ins. Three prompts, twice a week to start. You answer them too. Respond to the first blocker flag fast and visibly — this is the moment the system earns belief.
3. Week 3: Convert one status meeting. Take your weekly status sync and replace the go-around with pre-read statuses; keep only the discussion of ambers and decisions. The meeting will shrink by half or die entirely. Bank the time publicly.
4. Week 4: Add the shipped ritual and your reading routine. Put the daily 10 minutes and weekly 20 minutes on your own calendar as recurring blocks, because unscheduled reading is unread writing.

Give it six weeks before judging. The early wobble — people forgetting check-ins, boards going stale — is normal and responds to gentle, consistent modeling far better than to enforcement. What you're building is a team where the work is visible, the people are trusted, and nobody has to perform busyness for an audience — where you see everything you actually need to see, and nobody feels watched.

If you'd rather not assemble the instruments from four different tools, Openbook puts them in one space: boards, scheduled check-ins with mood and flags, project status rooms, and dashboards that read from the boards — with per-room visibility so each team controls what's shared. Explore the features and replace your first status meeting this month.