← All posts

Best Practice: Time Management, Scheduling & Timesheets

Time is the one resource an MSP sells by the hour and only gets to sell once. An engineer's day that isn't scheduled, logged or billed correctly doesn't come back. It's lost margin, a client who feels under-served, or an engineer heading toward burnout, often all three at once.

Professional-services firms routinely lose 10-25% of billable time to poor tracking practices: work done but never invoiced, because it was never captured accurately in the first place. That gap isn't a billing-software problem, it's a discipline problem. It starts with how shifts and rotas are planned, continues through how time gets logged day to day, and ends in whether overtime and after-hours work are priced and compensated fairly.

This guide covers four connected disciplines (rota and coverage design, calendar and capacity planning, timesheet accuracy, and overtime/TOIL handling), plus after-hours billing, and how each is built into MSP Pulse.

Designing rotas and coverage

A rota's job is to put the right skill in the right seat at the right time, not just to fill a grid. Build it from the response commitments outward, not from whoever happens to be free.

Start from response targets, not headcount. Coverage hours should be set per priority, not as one blanket window. A P1 commitment that promises 24/7 response needs a genuine on-call rotation behind it; a P3 commitment inside business hours doesn't need anyone awake at 2am. Deciding coverage before deciding headcount avoids the common trap of rostering people onto hours nobody actually promised to cover.

Publish rotas well ahead of time. Two weeks' notice is the common standard: enough for people to plan around it, and enough runway to catch a gap (a week with no senior engineer on a shift, a skill nobody rostered) before it becomes a live problem rather than a retrospective one.

Build in genuine overlap at handoff points. A shift change is where context gets lost: an open ticket, a half-finished diagnosis, a client who's already frustrated. A clean handoff needs both time and a habit, a short overlap window plus a written note, not a verbal "nothing much going on."

Rotate on-call fairly and cap its frequency. Being on-call is a cost even when nothing happens: broken sleep, a phone that can't be ignored, evenings that aren't really free. Spreading the load evenly (and tracking who's carried it recently) keeps it sustainable rather than something a junior engineer quietly absorbs because they're least able to push back.

Keep a backup responder for every primary. On-call rotations fail at the moment someone genuinely can't respond: illness, a dead phone, no signal. A named backup, escalated to automatically after a defined wait, is cheap insurance against an incident sitting unacknowledged.

Make shift swaps easy to request and easy to audit. Rigid rotas get worked around informally, off the record, which is worse than a formal swap process because nobody can see who's actually covering what. A lightweight request-and-approve flow keeps the rota and reality in sync.

In MSP Pulse: each person's normal working week is set on their Employment tab, and people who don't work fixed hours can be switched to a variable rota. The Rota tab on the Calendar page is where owners and admins build it: shifts are drafted, then published, and a published shift becomes that person's schedule for the day. On-call is its own kind of shift and can be given to anyone, whether they work fixed hours or a rota. The calendar shades non-working hours, so a gap in coverage is visible at a glance rather than discovered when a ticket goes unanswered. Response Targets resolve board → agreement → tenant default, with coverage hours tracked separately from priority (P1 defaults to 24/7, lower priorities to business hours), so the rota question ("who's covering this window") and the commitment question ("what did we promise for this priority") stay linked but distinct. Out-of-hours work that happens outside the published schedule surfaces for review rather than disappearing into a normal time entry, covered under overtime and TOIL below.

Calendar and capacity planning

A rota says who's meant to be working. A calendar says what's actually filling their day, and the two only stay useful together if leave, meetings and ticket work all live on the same view.

One calendar, not three. When leave is tracked in one system, scheduled visits in another, and ad-hoc ticket work nowhere at all, nobody can answer "who's actually free this afternoon" without cross-referencing three sources. A single shared calendar that shows working hours, approved leave, and scheduled or in-progress work is what makes capacity visible rather than guessed at.

Don't let unscheduled work vanish from the picture. A ticket someone worked on without ever booking a slot is still time that was spent. A calendar that shows only what was deliberately scheduled understates real load, so surfacing logged-but-unplanned work on it keeps the picture honest.

Plan to roughly 70-85% utilisation, not 100%. Billable utilisation (billable hours divided by available hours) is the core capacity metric, but experienced resource managers deliberately plan below full capacity. The unbooked portion is what absorbs the unexpected: an emergency callout, a longer-than-quoted job, someone out sick. Planning every hour as billable removes that buffer entirely, so a single disruption overloads the whole team rather than being absorbed by slack that was built in on purpose.

Approve leave against the rota, not in isolation. Leave requests approved without checking who else is already off, or what coverage that week needs, are how a team ends up short-staffed on a day nobody flagged as a risk. Leave approval and rota coverage need to be the same decision, not two separate ones made by different people at different times.

Distinguish availability from utilisation. A fully staffed week can still show low utilisation if demand is thin, and a short-staffed week can show high utilisation while still under-delivering. Both numbers matter, and conflating them leads to the wrong fix: hiring when the real problem is pipeline, or pushing harder when the real problem is headcount.

In MSP Pulse: the scheduling calendar overlays non-working hours, approved leave and the published rota on one view. Time logged against a ticket with no matching calendar block appears as a "ghost" entry, so unplanned work shows up instead of silently inflating someone's day. "Mark visit complete" pre-fills a time entry straight from a scheduled visit, keeping the schedule and the logged work in sync instead of two systems quietly drifting apart. Leave types, balances and requests go through an approval step, and approved leave appears on the same calendar, so approving time off and seeing its effect on coverage happen in the same place.

Timesheet discipline

Timesheets fail quietly. Nobody decides to under-bill a client; it happens ten minutes at a time, when an engineer finishes a quick fix and moves straight to the next ticket without logging it, meaning to "catch up later." Later rarely arrives with the same accuracy.

Log time as the work happens, not at the end of the day or week. The single biggest driver of timesheet accuracy is proximity. An entry made while the work is fresh captures the right duration and the right detail; an entry reconstructed from memory on Friday afternoon is a guess dressed up as a record. This is as much a habit problem as a tooling problem: the easiest path for the engineer needs to be the accurate one.

Separate billable from non-billable, deliberately. Every hour an engineer works is real time, but not every hour is chargeable: internal meetings, training, admin. Logging non-billable time too (not just skipping it) is what turns utilisation from a guess into a measured number, and what shows whether a low billable percentage is a demand problem or an internal-overhead problem.

Write entries you could invoice from, not just a duration. "1h - server" tells a client nothing and an auditor less. A short, specific description (what was done, for which issue) is what makes an invoice line defensible if a client questions it, and what makes six months of history actually useful when the same fault recurs.

Review entries on a fixed cadence, not just before invoicing. A quick daily or weekly scan catches a missing entry while the engineer can still remember what happened; catching it a month later at invoice time means either guessing or writing it off.

Freeze the rate at the point of logging. Whatever rate card, agreement or role-based rate applies when the entry is made is the rate that entry keeps. A later change to rates, roles or agreement terms must never silently rewrite hours already logged. This is as much a trust principle as a technical one: an engineer's logged hour and the rate it earned should never move under them after the fact, and a client's invoice should never shift for work already done.

In MSP Pulse: a time entry captures the ticket, the labour role, a start and end (or a duration), and a written note. The resolved rate is written to the entry the moment it's saved and never recalculated, even if the underlying rate, agreement or labour-role configuration changes afterward. Work against a quoted task is handled the same way at the point of entry: the entry resolves whether it bills time-and-materials or against a locked fixed-fee rate automatically, so the engineer never has to make that billing judgement. Standard entries resolve their rate from the labour role plus location and time modifiers, with office-hours versus after-hours differentiation available where configured (see below). Every entry stays editable while unbilled and locks once it's invoiced, keeping the audit trail intact.

Overtime, TOIL and out-of-hours work

Out-of-hours work is where good intentions go to die. An engineer picks up an evening call because the client's in trouble, and three months later nobody can say whether it was ever compensated, let alone how.

Require pre-authorisation wherever the work is plannable. Overtime that's agreed in advance, against an actual need, is manageable; overtime that's discovered after the fact in a timesheet is a surprise cost and often a surprise obligation. For genuinely unplannable work (an emergency callout), require it to be logged and reviewed promptly instead.

Write the policy down, including who gets a choice. Whether extra hours become paid overtime or time off in lieu (TOIL) shouldn't be an ad hoc conversation each time it comes up. A written policy should cover who qualifies, how TOIL is calculated against hours worked, whether the employee or the employer decides which applies, any cap on accrual, and a deadline for using accrued TOIL before it expires or is paid out.

Keep TOIL and standard leave balances separate. Mixing TOIL into a general leave balance makes both harder to audit and makes it easy to lose track of what's actually owed. Best practice treats TOIL as its own ledger.

Never let TOIL or overtime substitute for genuine rest. Regulations in most jurisdictions set minimum rest periods regardless of how extra hours are compensated. A policy that lets someone work a late callout and then start a normal shift a few hours later is a compliance risk as well as a wellbeing one.

Resolve it by decision, not default. Overtime shouldn't auto-convert to any one outcome. The right resolution depends on the situation, the policy, and sometimes the employee's own preference. A deliberate approval step, not a rule running silently in the background, keeps that judgement where it belongs. The entry and its review happen regardless of outcome; only the last step, which outcome it becomes, is a judgement call.

In MSP Pulse: out-of-hours time entries (logged outside a staff member's working schedule) surface for owner/admin review rather than flowing straight into pay or billing. Each is resolved individually to paid overtime or TOIL, with no default outcome, so the decision stays deliberate. Paid overtime is only available for people marked as overtime-eligible, and MSP Pulse hands the hours to payroll rather than storing any pay rates. TOIL goes into its own bank, separate from leave balances: you can set an expiry and a maximum bank size, the engineer books time off against the balance themselves, and the balance can never go negative. TOIL taken stays pay-neutral on the timesheet. This sits alongside the office-hours/after-hours billing split (next section), which is a separate question: one governs what the client is charged, the other governs what the engineer is compensated, and the two don't have to move together.

Office-hours vs after-hours billing

Charging more for after-hours work is standard practice across field services, but it only works smoothly when the cut-over is unambiguous and applied the same way every time, not decided case by case.

Define office hours once, and define the exceptions too. "9-to-5, Monday to Friday" isn't enough on its own. Public and national holidays need the same after-hours treatment as an evening, or a bank-holiday callout quietly bills at the daytime rate by accident.

Decide the cut-over rule in advance, including the edge case. A job that starts in hours and finishes after them needs a stated rule: does the whole job bill at the entry rate, or does the work split at the boundary? Deciding this once, as policy, avoids a dispute (with a client or with the engineer) every time it happens to come up.

Let the engineer log the work, not decide the rate. The billing classification should be a system calculation from the clock and the policy, not a judgement call made in the moment by whoever's doing the job.

Make the split visible on the invoice. A client charged a blended or unexplained rate for a job that ran into the evening is a client who calls to ask why. Showing the office-hours and after-hours portions separately, with their respective rates, heads that question off before it's asked.

In MSP Pulse: tenants can opt into automatic office-hours vs after-hours rate differentiation, derived from defined business hours plus a national holiday list, rather than left to engineer judgement. A tenant-configured cut-over policy decides how a boundary-crossing entry is handled: bill the whole entry at its start-time rate, or split it into separate office-hours and after-hours segments priced independently. It's opt-in and only applies where a labour role has both an office and an after-hours rate configured; otherwise entries fall back to a single line rather than guessing. Quoted work keeps its quoted rate and isn't split. It is entirely separate from the overtime question above: the clock decides what the client is billed; a human decides what the engineer is compensated.

A small set of metrics to watch

More dashboards don't produce better decisions; a handful of numbers reviewed on a fixed cadence does. The industry-standard target range for billable utilisation is roughly 70-85%. Planning above that leaves no slack for the unexpected, and a figure that's healthy for a junior engineer looks low for an owner who's meant to be split across delivery, sales and management.

Metric What it answers Watch for
Billable utilisation (billable hours ÷ available hours) How much of paid capacity earns revenue Below ~60% = demand or logging gap; above ~90% = burnout risk, no room for the unexpected
Logged vs scheduled hours Whether the rota and reality agree A persistent gap means the rota isn't reflecting real coverage
Same-day logging rate Whether time is captured while it's accurate A rising share of late entries predicts revenue leakage before it shows up on an invoice
Overtime/TOIL volume by person Whether the load is spread fairly Concentration on one or two people is a burnout and single-point-of-failure risk
Unbilled time aged > 30 days Work done but not yet invoiced A growing balance is deferred revenue, or revenue quietly being written off

In MSP Pulse: the role-aware Dashboard gives an engineer their own queue and logged time, and an owner or admin a business-wide view. It reads the same underlying time-entry and schedule data at two different levels, rather than being a separate reporting exercise bolted on afterward.

Pre-flight checklist for time management, scheduling and timesheets