Crowding & Low Visits
Farm Walk Checklist

A structured walk-through to find why cows bunch up in the morning and afternoon and why voluntary robot visits are below target in the renovated barn.

Farm

Date of visit

Cows / robots

Why the pattern matters: crowding at two set times with a quiet robot in between means the herd is being synchronized by something on a schedule — feed delivery, push-up, or the fetch routine. The goal is a steady flow of visits around the clock, not two rushes and a dead robot.

Today — 2 peaks, robot idle between

Target — steady visits, 24 h

0 of 0 checked

1 · Pull the data first

Software

Before walking the barn, let the robot software tell us when the problem happens and who it happens to.

Print the hourly milking / visit distribution for the last 7 days and mark the two peak windows Visits per cow per day target 2.6–3.0+ and refusals target 1–2 per milking Robot free time >12–15% — and check whether cleaning cycles land inside the peak windows Fetch list for 7 days: how many cows, and are they the same cows repeating? repeat cows = feet or fresh-cow issue Note exact times of feed delivery, push-ups, and fetching — overlay them on the visit graph

2 · Feeding & push-up routine

Most likely cause

If the peaks sit 30–90 minutes after PMR delivery, the feeding schedule is synchronizing the herd. This is the most common driver and the cheapest to fix.

Record PMR delivery times — do they line up with the two crowding windows? Push-up frequency 6–10× / day, spread over 24 h — not just around deliveries Bunk never empty: check feed available at night and mid-day, not just after delivery Feed refusals 2–4% and PMR balanced to the intended base milk level (M-minus) — if PMR covers too much milk, the pellet stops motivating visits Pellet check at the robot: fines in the bin, dispenser calibration, leftover pellet in the manger

3 · Human routine

Habit carryover

After a renovation, cows keep the old parlour rhythm for weeks — and every fetch or barn walk at the same clock time retrains them to wait instead of volunteering.

What time does the producer fetch? if it matches the peaks, he is training the pattern How long since start-up, and is the pattern improving, flat, or getting worse? Are fetched cows brought quietly in small groups (not a big drive that fills the commitment area at once)?

4 · Robot area & cow traffic

Renovation layout

Retrofits put robots where they fit, not always where they work. Walk the route a timid cow has to take.

Open floor in front of robot entry ≈ 4–4.5 m clear — can one boss cow block the doorway? Robot exit path: no dead end, tight turn, or waterer parked right at the exit backing cows up Crossovers: number and width 3.6–4.3 m; note any dead-end alleys the new walls created Traffic type (free-flow vs guided) and selection-gate settings if guided Watch 15 minutes at a peak: are waiting cows queued at the robot, or is the "crowd" actually bunched elsewhere in the pen? bunching elsewhere → heat / air, see §7

5 · Stalls & lying time

Chronic driver

Cows that can't lie down stand in the alleys — which looks like crowding — and once they finally get a stall they stay put instead of visiting. Count during a quiet window, about 2 h after feed delivery.

Stall use index (lying ÷ touching a stall) > 85% < 75% = stall problem Perching (front feet in, rear feet out) < 20% of cows at stalls Lying time from collars, if available 11–13 h/day, 8–12 bouts Stalls per cow — did cow numbers rise with the renovation without adding stalls? ≤ 1.0 cow/stall Measure: width 122–132 cm c-c, neck rail from curb 173–178 cm, neck rail height 122–127 cm, brisket locator ≤ 168 cm, ≤ 10 cm high Lunge space: posts, walls, or water lines blocking the forward lunge in the retrofit rows Bedding depth and condition (especially if mattresses replaced deep bedding)

6 · Feet & new concrete

Follows §5

New floors are either slick or abrasive; both cut walking. Lameness shows up first as low visits and a growing fetch list.

Locomotion score a sample of 20–30 cows score ≥ 3 cows vs the fetch list Hock scores (bald / swollen hocks confirm a stall-surface problem) Floor finish on new concrete: too smooth (slipping, hesitant walking) or too aggressive (sole wear)? Footbath frequency and last hoof-trimmer visit / trim records

7 · Heat, air & flies

Environment

If the "crowd" is a bunch in one corner or under fans — mainly in the afternoon — it's the environment, not the robot. Renovations often leave dead-air zones nobody commissioned.

Where exactly do the cows bunch? Map it against fan throw and inlets Air speed at cow level in the resting area ≥ 2 m/s in heat, including over the stalls, holding area, and robot Respiration rate on 10 cows mid-afternoon > 60 breaths/min = heat stress Fly pressure and control program — bunching is a classic stable-fly response

8 · Robot settings

Software

Milk access / minimum milking interval — is permission too restrictive for mid-lactation cows? Feed table: minimum pellet generous enough that low-visit cows still get most of their allocation? if yes, there's little reason to come more often Box time per milking and attach failures (long box times shrink capacity exactly at the peaks)

Recommendations

Ordered by cost and speed. Start at the top — the schedule changes are free and usually break the two-peak pattern within 2–3 weeks.

Break the schedule synchrony Start now

Split or shift PMR deliveries so the bunk is refreshed more often and less dramatically; consider moving the largest delivery to late evening to pull night visits.
Push up 6–10× daily, spread around the clock — an automatic feed pusher pays for itself here.
Move fetching off the peak times, keep fetch groups small and quiet, and vary the timing day to day.

Restore pellet motivation Start now

Confirm the PMR is balanced to the intended base milk level and not creeping up toward actual production — the robot pellet must stay the reason to visit.
Fix pellet quality at the robot: no fines, correct dispenser calibration, palatable pellet.
Tighten minimum pellet at the low end so a two-visit cow does not already receive nearly her full allocation.

Fix what the walk-through finds 2–6 weeks

Stalls: correct neck rail and brisket locator positions, add bedding, and clear lunge obstructions before considering anything structural.
Feet: book the trimmer, run the footbath 3–5×/week during the catch-up period, and groove or scarify slick new concrete.
Traffic: remove bottlenecks at robot entry/exit; relocate waterers that block the exit path.
Air: add or re-aim fans to kill dead zones over stalls and the robot area; verify ≥ 2 m/s at cow level.

Re-measure Week 3–4

Re-pull the hourly visit distribution and compare to the baseline printed on day one.
Track: visits/cow/day, refusals, fetch-list size, robot free time.
Success = flattening peaks, refusals rising toward 1–2 per milking, fetch list shrinking week over week.

Prepared by Basim · New Life Mills — Ruminant Nutrition. Targets are general industry guidelines for mature Holstein free-stall robotic herds; adjust for breed and barn.