Tracking Utilization Across Offices and Remote Teams (Before You Hire Again)

When a firm runs out of one office, capacity is managed by line of sight. A principal walks the studio, sees who's heads-down and who's idling on Pinterest, and staffs accordingly. It's imprecise, but it's fast — and for decades it was good enough.

Then the firm opens a second office. Or acquires a small studio two states away. Or keeps the flexible-remote arrangement that started in 2020 because it helps recruiting. Nothing about the work changed — but the line of sight is gone, and with it the firm's only utilization instrument.

What replaces it is, too often, nothing. And the most expensive symptom of flying blind isn't chaos — it's unnecessary hiring.

How invisible capacity turns into headcount

The failure sequence is easy to reconstruct:

1. Each office sees only itself. The Denver studio lead knows Denver's workload cold. She has no real picture of Portland — just an impression from the all-hands that "they're slammed too."
2. Everyone reports busy. In every office, asked or unasked, the answer is the same, because in a distributed firm being visibly busy is how remote and satellite teams prove they exist. Self-reported capacity always reads high.
3. A project lands, and the local office staffs it locally. Cross-office staffing feels like friction — different managers, a time zone, "their" people versus "ours" — so nobody seriously checks whether Portland has 200 available hours next month.
4. The local office is "full," so the firm hires. A rational decision, made on local data. Meanwhile Portland's actual average utilization has been 68% for a quarter, hidden behind a wall of self-reported busyness.

Run that loop a few times and the firm has added two or three salaries it didn't need. That's real money: the BLS puts the median architect wage at $96,690 (May 2024), and benefits make up roughly 30% of total employer compensation costs — putting the all-in cost of a median architect near $140,000 a year, and two unnecessary hires at over a quarter-million dollars annually. The capacity existed. It was just in a different city, invisible.

The mirror image is quieter and just as costly: one office runs chronically hot while another coasts, because workload balancing requires exactly the cross-office picture nobody has. The hot office starts showing the over-allocation patterns that precede burnout and turnover; the cool office quietly worries about layoffs. Same firm, same month.

Why "just ask the office leads" doesn't work

Most firms try to solve this with a recurring call where each office lead reports status. It fails for predictable reasons:

• It's self-reported, with all the inflation that implies — nobody dials in to say "we're 60% booked."
• It's a snapshot of this week, when hiring decisions need a forward view: what does utilization look like across the next quarter given the projects we've actually won?
• It's not in hours. "We're pretty stretched" cannot be compared to "we have 340 unallocated hours in March." Capacity decisions need units.
• It doesn't survive disagreement. When two offices both want the new hire, status-call impressions become negotiating positions. Only shared numbers end that argument.

What a real cross-office utilization picture requires

Four ingredients, regardless of tooling:

1. One plan, one system, every office. Not Denver's spreadsheet and Portland's whiteboard photographed into Slack. If allocations don't live in one shared system, every "firm-wide view" is a manual aggregation that's stale before it's finished. (Spreadsheets break down at exactly this point — multi-office is the multiplayer problem they're worst at.)

2. A common definition of capacity. Utilization percentages are meaningless if offices compute them differently. Agree firm-wide: what's a standard workweek, what counts as billable allocation, how PTO and holidays (which differ by location!) subtract from capacity. Boring, foundational, and worth an hour of partner time to settle once.

3. Forward projections, not just timesheets. Timesheets tell you Portland was underutilized last month. Only projected allocations — who is planned on what, week by week, through the backlog — tell you whether the firm can absorb the new project without hiring. The hiring question is always a question about the future.

4. Views by office, role, and person. The firm-wide number hides everything interesting. The useful questions are one filter away: Show me utilization by office for the next eight weeks. Now just the technical staff. Now who in any office is under 70%. That last view is where unnecessary hires go to die.

The hiring test

With that picture in place, "do we need to hire?" stops being a debate and becomes a checklist:

1. Is projected utilization across all offices — not just the loudest one — above the target your firm has set for billable staff, and does it stay there for the next 8–12 weeks? (Set that target deliberately: the Deltek Clarity A&E study puts the firm-wide utilization median at about 61% — a figure that includes principals and non-billable staff — so role-level targets for billable staff necessarily sit well above it.)
2. Does it stay there even after rebalancing — i.e., is there no one in any office below your rebalancing threshold whose skills fit the need?
3. Is the demand durable (signed work, not pipeline optimism)?

Three yeses: hire, confidently — you'll get the new person productive while the need is still real. Any no: rebalance first. Remote collaboration has removed most of the historical excuse — a Portland designer can carry 15 hours a week of a Denver project's DD documentation without boarding a plane.

Applied consistently, the test tends to produce two outcomes: fewer hires than instinct suggests — instinct runs on the loudest office's worst month — and more durable ones, because each hire is backed by a quarter of red projections rather than a single office's bad stretch.

The remote-work wrinkle

Fully remote and hybrid staff make the visibility problem universal: with nobody in line of sight, every firm is now a distributed firm, even the single-office ones. The failure modes just get personal — the remote employee who's chronically over-allocated but invisible until they resign, or the one who's been at 50% for two months and is increasingly anxious nobody has noticed.

The fix is the same one, applied at individual grain: allocation and utilization visible per person, regardless of where they sit. Presence was always a bad proxy for workload; remote work just forced the issue.

Seeing the whole firm at once

This is, transparently, a problem Resource was built for: staffing projections put every person in every office on one color-coded timeline — filterable by office, department, or role — with PTO subtracting from capacity automatically, and weekly projections extending out through the backlog. The "who in any office is under 70%?" question is a glance, not a meeting. The Starter plan is free for up to 10 people if you want to see your own data this way.

But tool or no tool, the operating principle for a distributed firm is this: never make a hiring decision on one office's data. The whole point of having multiple offices is that capacity in one can serve demand in another — and that only works if someone can see both at once.