4:12 PM on a Tuesday, and the furniture wasn't coming
I run rapid deployment for a commercial interiors firm — mostly tenant buildouts where somebody signed a lease, somebody else promised a move-in date, and the two dates don't quite line up. I've handled 200-plus rush installs over six years, including a handful where the client called on a Monday and wanted desks by Friday.
This one was the closest I've cut it.
The client was a 90-person software company that had taken a floor in a downtown tower. Their board was flying in Friday morning. The lease started Wednesday. And their furniture order — placed three weeks earlier with a national supplier — had gone from "processing" to "scheduled for delivery" to, when I finally got a human on the phone at 4:12 PM, "November 28."
Friday was November 15.
So we had 48 hours. Realistically, 41 working hours once you subtract the time nobody's awake.
Normal lead time for 60 workstations is three to four weeks. Ten business days if everything's in stock and you get lucky with freight. We didn't have ten business days. We had one very bad Tuesday afternoon and a client who was, understandably, not interested in hearing about lead times.
The first thing I got wrong
Marcus, our installation lead, asked me to pull the electrical drawings before I ordered anything. I told him we didn't have time to wait on the building engineer.
He said, "You're going to regret that."
He was right. It cost us nine hours on Thursday. I'll come back to it.
Wednesday: buying desks against a clock
I had three criteria, in order. In stock in a US warehouse. Shippable as flat-packed frames instead of palletized freight. And assembles with hand tools and one impact driver.
We went with Autonomous. Not because they're the only game in town — there are several direct-to-consumer desk brands that'll get you a frame inside a week — but because they sell the frame and the top as separate DIY kits. That sounds like a minor detail. It isn't. A 60-desk order in complete units arrives as freight: one dock appointment, one crew, one narrow delivery window, and a building manager who will absolutely charge you for the freight elevator. Frame kits arrive as boxes. Boxes you can carry up a stairwell at 7 AM.
Final order: 52 Autonomous SmartDesk 5 frames with 55-inch tops, six L-shaped units for the dev pods, two Autonomous Surface standing desk configurations for the conference rooms, 60 task chairs, 12 monitor arms, and four filing cabinets. Just under $38,000 with expedited freight.
The client's alternative was rental furniture at $2,400 a week for a setup that looked like a WeWork overflow closet. They approved the PO in eleven minutes.
Thursday: the floor plan catches up with us
The building engineer's drawings showed two power and data raceways running down the center of the floor, plus a column grid that didn't match anything the leasing agent had sent us in October. Which meant roughly half the desks had to swing 90 degrees from the plan.
Fine. Desks rotate.
The problem was the six L-shaped units in the dev pods. In the planned orientation they wrapped neatly around the columns. Rotated, they didn't fit. Not "fit awkwardly." Didn't fit.
We salvaged four of them by pulling the return sections and running standard 55-inch tops on the L-frames — same motor, same legs, one fewer section of desk. The other two went into the only spots on the floor where the columns actually allowed them. Total damage: three hours of unbolting and re-bolting, a $600 expedite on two replacement tops from a local shop, and one extremely patient electrician who had to re-pull four floor drops while we stood around pretending to help.
Then the second thing happened.
Out of 60 frames, two arrived with motors that wouldn't index. One ran to full height and refused to come back down; one just clicked and did nothing. A third box had clearly been shipped to somebody else and returned — scratched top, and a packing slip with a name that wasn't ours still taped inside the carton.
Autonomous shipped replacements overnight. They landed at 10:30 AM Friday. That's the difference between a rough week and a catastrophe, by the way — not whether something fails, but whether the vendor picks up the phone at 7 PM and moves.
At 11:40 PM Thursday the office manager asked if I could "just plug in the printer" so she could print the board packets. Two Macs, one networked laser printer, nobody on-site who'd set one up before. I spent 40 minutes on what amounted to adding a printer to a Mac, most of it fighting a driver conflict left over from the previous tenant's print server. Forty minutes. On a printer. In the middle of a furniture install.
Meanwhile one of the college kids we'd hired for the day shift spent her first hour refreshing an AP score calculator on her phone — her daughter's scores were posting that morning. She still out-assembled everyone on the floor, including me.
Friday, 8:12 AM
Sixty desks up. Chairs unboxed. Presets programmed.
We'd collected height preferences in a spreadsheet Wednesday night — the client sent a form to all 90 employees, got 61 responses, and we set sitting and standing heights for each responding desk before we left. Took about 90 seconds per desk. It was the single highest-return thing we did all week.
The board walked through at 9:00. Nobody looked under a desk. The CEO said it "felt like the company had been here for years," which is the nicest thing you can hear in this job and also completely meaningless, because that's what everyone says.
So what do I actually think of the desks?
Here's the honest version. What worked:
- Frame rigidity at standing height. I'm 210 pounds and I leaned on one at full extension. There's some flex up past 46 inches, but nothing that made me nervous. I've installed frames that felt like a wet noodle at 40 inches.
- Memory presets. This is the feature people actually use. Set it once, press a button, done. It's also the reason the desks stop being "furniture" and start being equipment — the difference between a desk people adjust and a desk people leave in one position forever.
- The DIY kits. This is the entire reason we could move at all. Two people, one impact driver, 22 minutes per desk once we had a rhythm. The first one took 50 minutes and a lot of swearing.
- The assembly documentation. Adequate. Not great. Adequate is fine.
Where it's less rosy:
- Two dead motors out of 60 is a 3.3% failure rate. On a normal install that's an annoyance. On a 48-hour install it's a five-alarm fire, because there's no slack in the schedule to absorb it.
- There's no local service network. If a motor dies in month eight, you're pulling the frame apart and shipping parts, not calling a tech. For a 60-desk floor that's a real line item. We now carry four spare frames in our warehouse specifically because of this job.
- The cable tray is a separate purchase. It's not hidden from you, but it's not front-and-center either. And 60 desks without trays look like a server room had a yard sale, which is not the impression you want on board day.
- The instructions are vague on control box cable routing. A first-timer will lose an hour to that. We did. The printed sheet covers the frame well and the electronics poorly.
One caveat on all of that: this was accurate as of January 2025, and Autonomous rotates SKUs and pricing more often than most furniture brands. If you're building a timeline around a specific model, verify the current spec sheet and lead time yourself before you promise a client anything.
Second caveat: my read is based on roughly 200 units across 11 buildouts, almost all mid-market commercial offices in the 40-to-120 desk range. If you're furnishing 500 workstations, or you're on a union site with labor rules I don't deal with, your constraints are different from mine and this whole write-up may not transfer.
Three things I changed after this job
- Electrical drawings before purchase orders. No exceptions, not even for a 48-hour turnaround — especially not for a 48-hour turnaround. Waiting two hours for a PDF would have saved us nine.
- Five percent spares on anything with a motor. Two dead units out of sixty isn't a quality crisis. It's a scheduling crisis, and it's cheap to insure against.
- Test bench on day one. Assemble two desks, run them up and down for an hour, find the problems before you're staring at 58 half-built frames at 1 AM.
What this actually taught me about efficiency
We cut assembly time from 50 minutes to 22 by building a sequence and then refusing to deviate from it. Two people per desk. One on the frame, one on the top. Torque in the same order every time. Presets programmed last, after the desk has run its full range once. That's the whole trick. No new software, no fancy tooling — just a checklist and a rule against improvising when you're tired.
Here's the thing, though: it's tempting to think a rush job is just a normal job compressed. It isn't. A rush job is a job where you decide, in the first hour, what you're willing to do badly. The cable routing could be ugly on day one; we cleaned it up the following Tuesday. The executive office kept its existing chairs. The conference rooms got the two Surface desks and nothing else. Every one of those was a trade, and none of them were on the original scope sheet. That's the actual skill — not speed, but triage.
The other piece: efficiency only buys you time where the demand is standardized. Half our floor was 55-inch tops on identical frames, and that half flew. The six L-shaped pods — the only custom piece of the job — ate roughly 40% of our labor. That's not an argument against custom desks. It's a reminder that "just standardize everything" isn't advice. It's a decision about what you're willing to lose.
How it ended
We finished at 8:12 AM Friday, 48 minutes before the board walked in. The client's alternative — six weeks of rental furniture — would have cost more than the desks did, so the expedite fees were, in the end, almost a rounding error. We paid about $2,300 in rush freight and replacement parts to protect a project that would have collapsed without them. That math works out fine.
If you're staring down a similar timeline: the desk brand isn't your bottleneck. Your bottleneck is drawings, spares, and the two hours you're going to lose to something nobody planned for. Ours was a printer. Yours will be something else.
And if you take one thing from this whole mess — call the installation lead first, then order. Not the other way around. I've watched people get that backwards for six years, and I finally did it myself in November.
Need help applying this idea?