Scalatris
ENGINEERED TO ELEVATE
ISSUE #14
THE FACTORY FORWARD · 1 SEPTEMBER 2026 · 6 MIN READ, DESIGNED TO RUN IN 12
Build the Case Before You Hire.
▸ FIELD NOTE. The overtime bill that made a weak headcount ask look strong.
▸ TOOL TO TRY. A prompt that builds a headcount case from your overtime log, not a guess.
▸ THIS WEEK'S NUMBER. How long a new hire actually takes to pay for themselves.
WE'RE LIVE ON PRODUCT HUNT TODAY
The Factory Forward launched on Product Hunt today. If this newsletter has been useful, a vote helps more manufacturing leaders find it.
The Factory Forward
One practical AI move for your plant, every Tuesday
Four Tuesdays ago the cycle time at the inspection station moved from 40 seconds a piece to 24. You've been running the 30-day update since. The rejection number is still catching up, but the honest paragraph you send every month has kept the room believing you.
Now the machine has bought you something you didn't expect. More parts are clearing inspection than your finishing team can pack and deburr in a normal shift. Every extra piece is waiting in a tray at the end of the line, and someone is covering it in overtime.
You want to hire. You're about to make the same mistake that got your budget ask rejected two months ago.
This week: turn "we're behind, hire someone" into a headcount case, built the same way you built the machine's payback case.
THE GAP
What hiring costs. What not hiring costs.
COST OF HIRING
₹26,000/month
One finishing operator, fully loaded. No overtime once trained.
COST OF NOT HIRING
₹27,500/month
Overtime to clear the finishing backlog, most days.
The overtime line barely clears the case. What it protects doesn't.
Full arithmetic in the field note below. Illustrative figures, not a client's.
ONE FIELD NOTE
The number that made the ask land wasn't the overtime line
DISCLOSURE, READ THIS BEFORE THE NUMBERS.
This is a constructed scenario, not a client. We don't have a named engagement to cite yet. What follows extends the same composite pattern from Issue #13, the press-parts unit near Nashik that added vision inspection to a stamping line, because the shape of what happens next is one we see repeat. Read it as a shape, not a record.
What happened. Cycle time held at 24 seconds, same as the 30-day update reported. That's roughly 40 percent more parts clearing inspection per shift. The finishing station downstream, still manual, couldn't absorb the extra volume. Within three weeks the plant was paying for close to two hours of overtime a day, most days.
Why the first version of the ask failed. "We're behind on finishing, I need to hire" is the same shape as "We need ₹8 lakh for vibration monitoring" from Issue #12. Cost stated, no payback date, no downside case. It reads as a request to spend more just to keep up.
"A payback case proves the machine was right. A headcount case has to prove the person pays for themselves too."
What the second version did instead. It named the overtime cost, named the new hire's fully loaded cost, then went one step further. Without a finishing operator, the plant can't sustain the pace that made the ₹8 lakh inspection case work in the first place. The overtime line alone gives a ten-month payback on the ₹15,000 hiring and training cost. Counted against the ₹73,000 a month the earlier case depends on, the same investment pays back in under a month.
What it costs you to skip this. Say nothing and let the backlog sit, and the machine's own payback case quietly stops being true. Nobody tells the promoter that directly. He just watches next quarter's number come in soft, with no explanation attached.
ONE TOOL TO TRY
The headcount case, built from your overtime log
Run this on what you already have: your 30-day update notes, and whatever you use to track overtime. A punch-card export, a WhatsApp tally, a foreman's notebook.
Before you paste anything: use a company-approved AI tool only. Strip the employee's name, write "the finishing operator," not a person. Strip any salary figure that isn't already an internal HR band.
You are helping me turn an operations request for one additional hire into a case for internal review. You are producing draft analysis a human must verify, not a decision.
INPUTS
Role and one-time hiring/training cost: [PASTE]
Fully loaded monthly cost of the role: [PASTE]
What is currently costing money or capacity without this person, however messy: [PASTE, overtime log, backlog, rework, delay]
Any other capital or budget decision this role protects or completes: [PASTE, or NONE]
My maximum acceptable payback: [MONTHS, or blank]
STEPS
1. Split my inputs into VERIFIED (from records) and ASSUMED (inferred). Show both. Never move an item from ASSUMED to VERIFIED.
2. Restate the request in one sentence a promoter would use, with no HR or ops jargon.
3. Name the monthly cost of NOT hiring, in rupees per month. Show the arithmetic line by line. If a number is missing, name it and stop. Do not estimate it.
4. Net monthly benefit equals monthly cost avoided minus the role's fully loaded monthly cost. Print it on its own line, even if it is small or negative.
5. If net monthly benefit is small or negative on its own, do not stop there. Ask me: does this role protect or complete a separate investment already approved? If yes, name that investment's monthly value and show the combined payback. If no, say plainly that the case does not clear on its own.
6. Payback equals one-time hiring and training cost divided by net monthly benefit, or combined benefit from step 5. Round up.
7. Name the single biggest reason this gets rejected, and write the one-sentence answer to it.
8. End with one line, under 20 words, that the finance lead can repeat to their CA when I am not there.
Close with: "Draft analysis. Verify all inputs and assumptions before internal presentation, and before this goes to HR."
Do not use the words leverage, synergize, holistic, cutting-edge, optimise, streamline, or efficiency.The line doing the heavy lifting: "If net monthly benefit is small or negative on its own, do not stop there." That's the difference between a case that ends at the first rejection and one that goes looking for the real number.
YOUR ONE MOVE THIS WEEK
Get the fill-in template
The prompt above, a worked example filled in, and a blank version for your own numbers. Takes 5 minutes. Nothing to install.
Was this issue worth 5 minutes?
THIS WEEK'S NUMBER
3 to 8 months
That's how long research on new-hire ramp time says it takes most new employees to reach full productivity. If the finishing operator you're about to hire isn't paying for themselves in month one, that's normal. It isn't a reason to walk the ask back.
Source: Click Boarding, "How Long Does It Take for a New Employee to Be Productive?" Read the article. US and international workforce research, not an India-specific MSME benchmark, flagged so you can weigh it accordingly.
What's one headcount ask you gave up on after it got rejected, and did anyone go back later to show what not hiring was actually costing?
Hit reply and tell me. I read every one, and the sharpest answers become the spine of a future issue.
COMING NEXT TUESDAY
Budget was the first of three asks. Headcount is the second. Issue #11 promised a third: mandate, the point where you stop asking permission for individual line items and start owning a number.
P.S. Keep every version of the headcount case, rejected or approved. Six months from now, when someone asks why your team grew and another plant's didn't, the paper trail is the answer.
Until next Tuesday,
Sachin
Founder, Scalatris LLP · scalatris.com
REFER A COLLEAGUE
Know the ops lead who just got told "we can't add headcount right now" and hasn't been handed a number to argue back with? Forward this. That's the whole ask.