DreamFieldsA plan for the Associate Vice President, Manufacturing Operations, Security & Facilities
seven sites.one standard.
Where I’d be every day, what I’d do every hour, how each plant gets followed up between visits, and how seven sites get to one audited scorecard, then to live daily production and cost numbers.
The seven plants come from public job postings and press. Nearby locations, like the Madison Heights warehouse and postings in Banning, CA, ride along with the closest plant’s visit. I’d confirm the full list in week one.
the whole plan, in two and a half minutes.
Every part of this page as a short film. It plays without sound, so it works with the volume off. Tap a chapter to jump to it.
six doubts, answered.
If I were hiring for this role, these are the questions I’d want answered. Each one links to the part of the plan that answers it.
four weeks. every floor.
Home base is Tempe. Phoenix Sky Harbor flies nonstop to every plant’s airport, so each trip is one flight out and one flight home. Two weeks a month on the road, two at home running the remote cadence. Press play to watch a full rotation, or pick any day.
Why Tempe: Phoenix Sky Harbor flies nonstop to Palm Springs, Detroit, Newark and Boston. From Palm Springs, Detroit and Boston take a connection and Newark is seasonal. Flights and drive times are typical, not booked.
or use ← → keys
Tempe’s first-shift huddles, Tier 3 from the gate, a nonstop to Detroit, then Hazel Park’s second shift and the Madison Heights warehouse.
- 5:00 AMAZTempe line boards
- 7:00 AMTier 3 from the PHX gate
- 7:45 AMFly PHX → DTW
- 2:30 PMETHazel Park second shift
- 4:00 PMMadison Heights warehouse: security walk
- Actions due in the tracker: Desert Hot Springs
- Desert Hot Springs actions in the tracker: owners and dates within 48 hours
the days, by the hour.
Four kinds of day, including the first one. Times are local to where I am. Hover or tap any block to see what it’s for.
Day one is for seeing the floor before anyone briefs me on it, meeting the people who run the plants, security, maintenance and money, and leaving with every plant’s first visit booked.
The plant manager meets me at the door, and I’m escorted and logged until my badge is issued. Watch start-up: attendance, line clearance, materials staged.
Watch two line boards. Day one is for listening, not running.
Watch how the site runs its daily meeting today: who owns which number, and what happens when one goes red.
Walk one pre-roll SKU from material issue to finished goods with the plant manager. Count stations and people; time ten cycles at the slowest station.
Coffee and one question: what gets in your way most days?
The site list, priorities, decision rights, and what good looks like at day 90. Book the weekly Tier 4.
Thirty minutes with the regional directors, then ten with each plant manager while it’s still their working day: their biggest problem, and the best day to walk in with their first shift. First visits booked.
With the security lead: perimeter, doors, the camera map against real coverage, vault, transport staging, officer post orders and licenses, the last twelve months of incidents.
Sit with the crew. Listen more than I talk.
With the maintenance lead: the PM backlog, the top causes of downtime, critical spares, open capex and construction.
See the handoff between shifts.
How cost per unit is built today and what’s reported weekly. Agree on one cost bridge from the floor to the P&L.
Request read access to line data, timeclock and labor-cost exports, METRC, the maintenance system and security logs.
One page to the COO: what I saw, what I asked for, what happens the rest of the week.
A visit day starts when the crew does. Everything I check, I check where the work happens. This one ends on Lowell’s second shift, fifteen minutes away.
Same door and badge check as the crew. The first look at access control is the one every employee gets.
Ten minutes each at the line board: yesterday’s count, today’s plan, one problem.
Time ten cycles at three manual stations. Check line clearance and label verification on the automated lines.
Supervisors, maintenance, QA and security with the plant manager. Every red leaves with an owner.
Yesterday’s shift log for two lines and three random batch records, traced to METRC and the timeclock.
Camera coverage against the site map, access exceptions, officer licenses, vault and transport steps, the maintenance backlog.
Fifteen minutes on one screen, run from this site’s board. Plant managers see each other’s reds and offer help.
Scorecard, open countermeasures, people and capex asks. Ends with dated actions.
Taught on the floor in 45 minutes, with Lowell’s supervisors in the room. This visit: find the bottleneck by timing a line.
Listen. Morale shows up on the floor long before it shows up on a scorecard.
The A3 on this site’s biggest loss, reviewed with its owner on the floor. Is the fix working?
AI drafts it from my notes and the day’s data. I edit it and send it to the plant manager before I leave the building.
Fifteen minutes.
Tier 1 with Lowell’s second shift, then the first hour on their floor. Both shifts run the same standard.
A different supervisor each visit, this time from Lowell’s second shift. One coaching topic, and one question: what’s in your way?
AI drafts them from my notes and the day’s data. I edit them and send them before I leave. Lowell’s first shift is tomorrow.
Home weeks run the remote cadence: every site reviewed by video, one site audited every day, and time on the Tempe floor. Times are for winter; Arizona skips daylight saving, so from March to November anything set in Eastern time lands an hour earlier.
What moved overnight, what’s red, and who owns it. By month twelve, the AI brief writes this.
The home site runs the same standard. Walk two line boards at Tier 1.
The plant manager runs it. I listen.
A different area each day: production, processing, warehouse, maintenance.
9:00 Eastern, fifteen minutes, every plant manager on one screen.
One site a day: five random records traced from shift log to METRC to timeclock to scorecard.
Capex, hiring and purchase approvals, so nothing waits on me.
Two sites a day, 45 minutes each, with the plant manager and their regional director. Every site gets a review or a visit every week.
Tuesdays, by video. The week in dollars: what the gains are worth and what the misses cost. Once a month it’s the business review, in person at headquarters.
Same rule as on the road.
Standards, investigations with HR and Legal, officer licensing, continuity drills, contractor and capex reviews.
Regional directors, security and facilities leads. Same slot every week; by video when I’m on the road.
Action tracker review. Anything past its date gets a call today.
Every flight is a nonstop out of Phoenix. Travel days start at home and end on a plant floor, or start on a floor and end at home.
At home. What moved overnight, what’s red, and who owns it.
Tier 1 on two lines at the home plant.
The plant manager runs it. I listen, then head to the airport.
Ten minutes from the plant. Security, then Hazel Park’s packet at the gate.
9:00 Eastern, fifteen minutes, every plant manager on one screen. Then board.
About four hours, nonstop. On the plane: Hazel Park’s last four scorecards, open actions from the last visit, this week’s audit picks.
1:40 PM Eastern. About thirty minutes to the plant.
2:30 PM Eastern. Walk second shift with its supervisors: the board, the handoff from first shift, the slowest station.
4:00 PM Eastern. Security walk: cameras, access, the vault and transport staging.
5:00 PM Eastern. First shift starts at 5:00, and tomorrow is the full visit day.
Tier 3 runs at 9:00 Eastern every weekday: 7:00 in Arizona in winter, 6:00 in summer, and 6:00 Pacific. Where I don’t know a site’s shift times yet, I’ve assumed first shift at 5:00 AM and second at about 1:30 PM. I’d confirm each site’s in week one.
no plant goes a day without contact.
A visit is two days in twenty-eight. The other twenty-six are where the follow-up happens. Here’s one plant’s month, using Hazel Park.
Weekends shaded. The action burn-down is an example of how the tracker reads, not real data.
one scorecard. one cadence.
Every site reports the same numbers, in the same order, owned by name, and every number ladders up to contribution margin per unit. Click any number for its formula, source, owner and rule.
Cost
Throughput
Quality
Safety, security & facilities
People
Targets come from each site’s own 30-day baseline. No number on this page is a DreamFields target.
What each unit leaves after the costs that move with volume.
- Source
- Finance
- Owner
- COO and Finance. I own the cost side.
- Cadence
- Weekly estimate, monthly close
the huddle cadence.
Four tiers, fifteen minutes or less, the same at every plant. Problems climb until someone can fix them, and fixes travel to every site with the same machine.
Yesterday’s count, today’s plan, one problem. At the line board.
The site board. Every red leaves with an owner.
One screen. Reds over two days, overdue actions, help needed between sites.
The week in dollars: what the gains are worth and what the misses cost.
The operator calls maintenance and logs a reason code on the line tablet.
Running again by 6:10. The cause is logged; the fix is temporary.
The maintenance lead owns the permanent fix.
Another site runs the same labeler and shares its setting.
Verified on the floor. The setting goes into standard work at every site with that machine.
supervisors run the floor.
Nine modules, each taught on the floor in 45 minutes and finished with an assignment on the supervisor’s own line. It covers what the posting asks for: root cause, Lean, cycle time, bottlenecks and corrective action.
OEE and units per labor-hour from your own board.
Explain yesterday’s numbers at Tier 1 without notes.
Cycle time, takt time, and the ten-cycle study.
Time ten cycles at two stations and post the result.
Where the line actually waits, and why adding people anywhere else does nothing.
Find your line’s constraint and prove it with timings.
Write it, train it, check it.
Write standard work for one station and train two operators on it.
Five whys and fishbone on a real stop.
Run a root cause on a stop from this week.
A3s and corrective actions that stay fixed.
Close one corrective action and show it held for two weeks.
Tier 1 in ten minutes, Tier 2 in fifteen.
Run Tier 1 every day for a week, observed twice.
Feedback, recognition, attendance, and the labor agreements, New York’s included. Taught with HR.
One recognition and one coaching conversation, logged.
Use AI tools to analyze your line’s data, see what changed and why, and make a better call. Then check it on the floor.
Analyze a week of your line’s data with an AI tool and bring one decision to Tier 2.
The two pilot sites. I teach it.
Every supervisor through modules one to seven.
Which is how it scales past me. Successors named for every plant manager.
try the math.
Two tools I’d use with every plant manager in the first month. The first starts from my own pre-roll line. Change anything.
A line only runs as fast as its slowest station. Add or move people, change cycle times, and watch where the units pile up.
Eight people, 24,000 per 8-hour shift, $0.06 direct labor per unit. The station breakdown is illustrative.
The posting asks for OEE gains translated into financial results. Here’s the arithmetic, both ways: more units if demand takes them, fewer line-hours if it doesn’t.
Placeholder numbers · put in yoursOEE today
The lines
If demand doesn’t take the units
trust the numbers first.
A scorecard is only as good as its worst number. Every week, random records get traced back to source, and a number that fails doesn’t count until it’s fixed.
Line 2 · second shift
Pre-roll package
8 people on the line
Units per labor-hour
the audit cadence.
Automatic checks every day, a human trace every week, a trust score every month, and on-site audits every quarter.
Every line and station reports within 30 minutes of shift end. Out-of-range numbers are flagged, never deleted. Reported units are checked against METRC packages.
Five random records per site, traced end to end: shift log → METRC package → timeclock hours → scorecard. All four have to agree.
The share of audited records that matched source. It sits next to every site’s scorecard, and a site’s numbers count once its trust is high.
On-site audits of each manufacturing partner, and a security audit at every owned site: camera map, access logs, officer licenses, transport steps.
accountability, in writing.
The same rules at every site, for every number.
- 01Every number has one owner, by name.
- 02Same scorecard, same order, at every site. Posted on the floor.
- 03Red two days running: the owner brings a countermeasure to Tier 2.
- 04Red five days: an A3 at Tier 4 with the COO.
- 05A follow-up past its date stays on the Tier 3 board until it closes.
- 06A number that fails an audit doesn’t count until it’s fixed.
- 07Recognition follows the measured numbers. Same rules for everyone.
A red number never sits. It either gets a countermeasure, gets help, or gets the COO’s attention.
protect people, product and facilities.
Two-thirds of this job’s title, and the part my resume says least about. Here’s the program, one standard across five state rulebooks, the audit I’d run at every site in the first 30 days, and what happens on a bad day.
In-house officers under regional security managers, plus outside providers where we use them. Every license tracked with its expiry date. Written post orders for every post, including use-of-force and weapons rules by state.
Badge tiers by role, reviewed quarterly. Visitors logged and escorted. Cameras mapped to every door, vault and limited-access room, with uptime as a daily number. Alarms and intrusion detection tested on a schedule. Video only, no audio.
Counts checked against METRC every day. Vault access by two people, logged. Cash under the same controls, and printed packaging counted like product. Transfers on manifests with chain of custody from dock to truck. Waste destroyed and documented to each state’s rule.
One maintenance system at every site, response times for line-down calls, critical spares on the shelf, fire and life safety inspections on one calendar, contractors scored and insured.
A playbook for each bad day, an owner for each, drills on a calendar, and one call tree for the company.
Where state rules differ, the standard follows the strictest one, with a one-page appendix per state. Camera audio stays off everywhere, which keeps audio-consent laws out of it; Legal approves camera placement, signage and retention, and there are no cameras in restrooms or locker rooms. Partner sites in Missouri, Ohio, Illinois and Ontario are audited against the same standard as each partner agreement allows, and any medical-licensed site gets a Schedule III readiness review with Compliance and Legal.
the first-30-days audit.
The same checklist at every owned site. Tick through it.
pick a bad day.
Every scenario has a playbook, an owner and a drill on the calendar. Here’s the first 15 minutes, the first hour and the first day.
People first: comply, nobody resists. Panic alarm only if it’s safe. Call 911 once they’ve left, then lock down and account for everyone. The security lead calls me.
Preserve all video. Start the incident log. Compliance starts the regulator and police notices. Once police release the scene, two people count the vault against METRC.
Regulator notice confirmed within the state’s deadline. Insurance claim. The point of entry fixed before the next shift.
facilities standards.
Maintenance, construction and capital, run the same way at every site.
- Line-down calls: a technician at the line within a response time set from the baseline, every shift.
- Preventive maintenance done on time, tracked in one maintenance system.
- Critical spares for every bottleneck machine, counted monthly.
- Contractors: no current insurance certificate, no badge. Every contractor scored on safety, schedule, budget and quality.
- Capex: every ask on one sheet, ranked by payback, effect on OEE, and risk.
- Fire and life safety inspections for every site on one calendar.
day one is already planned.
The first week day by day, the first 90 days in three cycles that match three laps of the rotation, and the year that follows. Drag through the year to see where everything stands.
First shift at 5:00, then the COO, security, maintenance, Finance and every plant manager.
A second first shift, in processing and packaging. QA and Compliance leads, and my state agent registrations started so I can walk every floor. HR: org charts, open roles, 90-day turnover. Pull 90 days of what each site reports today.
Every plant manager on one call: three questions, fifteen minutes. Home to Tempe: first shift on my home floor, and a video site review with Hazel Park. Start the officer license list: every officer, every license, every expiry date.
Video site reviews with Middletown, Peekskill, Tyngsboro and Lowell. Downtime by cause from the maintenance system. The one-page baseline template, the same for every site.
Meet whoever owns the partner relationships, then the Missouri, Ohio, Illinois and Ontario partners. A week-one note to the COO. Week two booked: Desert Hot Springs and Hazel Park.
three laps, three cycles.
Each 28-day cycle is one lap of the rotation, and ends with something you can hold me to.
- First shift at every owned site. Time the stations, read the boards, meet every supervisor.
- Baseline OEE on every automated line and units per labor-hour at every manual station.
- Security audit at every site: camera map, access tiers, officer licensing, alarms, the incident log.
- Maintenance backlog, downtime by cause, and open construction, site by site.
- One scorecard and one KPI tree, in the same order at every site.
- Tier 1, 2 and 3 huddles running at all seven sites.
- Weekly data audits start. Every number gets one owner, by name.
- Supervisor academy, cohort one, at the two pilot sites.
- Pilot results, before and after, in units and dollars.
- Pilot standard work copied to every site with the same equipment.
- Security standard version one and the continuity playbook, drilled at two sites.
- A capex plan ranked by return: throughput, cost, and risk.
the first year.
Months four to twelve make the standard live: machines first, then manual stations, then daily cost, then the AI layer.
First shift at 5:00.
every working day of year one.
Thirteen laps of the same rotation. The pattern doesn’t change, so every plant can plan around it.
Every working day of year one, before holidays and PTO. Tap a day to open it in the rotation.
live, lean, AI-assisted.
By month twelve, every number arrives on its own, every number traces to source, and an AI layer reads it all first thing every morning. People still decide and own every number.
- Line PLCs and packaging machines
- Tablets at manual stations
- Timeclock and payroll
- METRC
- ERP costs
- Lab results
- Maintenance system
- Access and camera logs
- Daily brief
- Anomaly flags
- Shift forecasts
- Drafted countermeasures
- Questions in plain English
- Tier 1–4 huddles
- Plant managers
- COO and Finance
AI drafts and flags. People decide and own the number.
the control tower.
What the COO, Finance and every plant manager would open each morning. This demo runs a sample morning at 60× speed.
| Plant | Units today | OEE | Schedule | Labor $ / unit | Overtime | Access exceptions | Camera uptime | Status |
|---|---|---|---|---|---|---|---|---|
| Desert Hot SpringsDHS · 6 lines | 0 | 74% | 97% | $0.068 | 4.2% | 0 | 99.8% | On track |
| TempeTMP · 4 lines | 0 | 71% | 95% | $0.071 | 5.0% | 1 | 99.4% | On track |
| Hazel ParkHZP · 4 lines | 0 | 66% | 92% | $0.074 | 6.8% | 0 | 99.7% | Watch |
| MiddletownMID · 3 lines | 0 | 58% | 86% | $0.088 | 9.4% | 2 | 98.6% | Act |
| PeekskillPKS · 2 lines | 0 | 67% | 92% | $0.078 | 6.3% | 0 | 99.9% | On track |
| TyngsboroTYN · 3 lines | 0 | 69% | 94% | $0.076 | 5.5% | 1 | 99.5% | On track |
| LowellLOW · 2 lines | 0 | 65% | 93% | $0.079 | 7.1% | 0 | 99.2% | Watch |
the receipts.
Thirteen years building regulated cannabis manufacturing in Colorado, California and Arizona. On the executive team at Vertis today, running 80–100 people through managers and supervisors.
Capacity for 24,000 per 8-hour shift from an eight-person team, at $0.06 direct labor per unit. $0.07–$0.09 Dutch-crowned and QC’d.
Moved the core team in-house from contract pay onto a performance-based pay band. Ten trimmers walked out; I won them back and promoted their leader. Still our top crew.
Grew Vertis manufacturing revenue, mostly by turning bulk flower into packaged products for Bright Lights, 3 Bros and Stereo.
Grew Nectar Bee from a 2,500 sq ft converted kitchen to a purpose-built 22,000 sq ft infused products plant. Revenue from $850K to $16.6M.
I built Vertis’s eQMS platform, including automated live financial reporting. It saves the company about $30K a year. I’m on the executive team and report to the President and the Board.
Developed and implemented the program across the whole Vertis facility, cultivation included. Six Sigma Black Belt with GMP auditor certification.
About four hours of my time, research included. AI tools did the drafting, coding and charting; I set the direction, and I own every number on it.
Bigger builds take real work. The eQMS I built at Vertis was a sustained build, and it saves the company about $30K a year, year after year.
It’s also what I’d teach: empowering supervisors and plant managers to use AI tools to analyze their own data and make better decisions, with a person checking the result on the floor.
let’s
talk.
I’d like to walk you through this plan on a call, or on the floor at any of your plants.
the first asks.
- 01The confirmed site list, headcounts and shift times.
- 02Read access to line data, timeclock and labor-cost exports, METRC, the maintenance system and security logs.
- 03An hour with each plant manager and each security lead in the first two weeks.
- 04The last 90 days of whatever each site reports today, so the baseline starts from your numbers.