Review farm performance

What this workflow helps you decide

Use this workflow for a repeatable weekly or monthly management review. It connects four different kinds of evidence that should not be mistaken for one another:

• Farm Dashboard — today’s work, live exceptions and a small set of operational indicators;
• Farm performance — a fixed 30-day milk, tracked-cost, margin and estimated-time-saving summary;
• module analytics — detailed milk, breeding, health, feed, beef or inventory analysis; and
• reports — filterable, exportable records used to verify the animals, dates and transactions behind a conclusion.

The person presenting a result should be able to name its farm, period, unit, denominator and missing-cost limitations. Changing a dashboard filter or opening a chart never records or corrects farm activity.

1. Confirm context before reading a number

1. Check Current Farm in the Farm Portal header.
2. Confirm the date shown by Farmlync agrees with that farm’s local day. A user in another timezone still reviews the farm’s date.
3. Confirm the organisation’s milk unit and currency. A currency setting formats values; it does not convert historic amounts.
4. Read loading and error states. A failed card is not a measured zero, and a dash is not proof that nothing happened.
5. State the period. The top Farm Dashboard cards and the dedicated Farm performance page use different windows.

When comparing two farms, keep the unit, date window, herd definition and record completeness the same. A larger farm total is not automatically better performance per cow.

2. Start with today’s operational dashboard

Open Today → Dashboard. Use it to decide what needs attention before interpreting long-term performance.

Headline indicators

Card	Current period and meaning
Conception rate	Resolved breeding outcomes over the last 90 days. Open Breeding analytics before explaining a change.
Average daily yield	Per-cow milk summary over the last 30 days. Open Milk for recording coverage and the animals behind it.
Herd SCC	Volume-weighted cell count over the last 30 days. Inspect milk-quality and udder-health evidence before acting.
Calving interval	Rolling 24-month breeding measure. It needs consecutive calving history; a dash can reflect incomplete evidence.

Work, exceptions and farm pulse

• Start Today’s Work opens Execution mode; its counts are unresolved jobs, not completed animal records.
• Today’s Work previews urgent tasks. Use Open all jobs for the complete list.
• Compliance & exceptions surfaces active milk/meat withdrawals, low supplies and registry submissions. Open the source before dispatching milk or animals.
• Next actions is a calendar preview in farm-local dates.
• Routine, progress & financial context contains daily recording progress, short-period economics and 12-month health spend. It is separate from the dedicated 30-day Farm performance calculation.
• Farm pulse summarises herd composition and links to the filtered animals behind its counts.

Do not report a dashboard preview as a closed management figure. Use Farm dashboard for each card and shortcut, then verify unusual counts in Work, Alerts, Calendar or the owning module.

3. Open the dedicated Farm performance page

Open Farm Portal → Manage → Farm performance dashboard. Access requires the current farm permission to view dashboard and analytics. Editing labour costs requires Owner or Manager authority.

The four headline cards cover the 30 completed farm-local days before today and compare percentage changes with the immediately preceding equal-length window. When the prior value is zero, no percentage delta is shown.

Headline	How Farmlync currently calculates it
Milk yield	For each recorded day, total active milk quantity divided by distinct recorded animals; Farmlync then averages those daily values across days that have records. The value is displayed in litres per cow per day.
Milk revenue	Sum of computed payouts on bulk milk collection deliveries in the period. A milk record by itself creates no payout revenue.
Net margin	Milk payout revenue minus tracked feed-event, treatment, veterinary-visit and weekly labour costs, divided by distinct cows with active milk records in the period and by the 30-day window. It is unavailable when no such cows exist.
Time saved	Estimated minutes from the last seven completed days of recorded operations, divided by seven. It is not clocked labour time.

Tracked margin is not a farm profit-and-loss statement

Farmlync excludes utilities, capital purchases, depreciation, financing, land, machinery and other overheads. Feed cost comes from recorded feeding-event quantity multiplied by that event’s unit cost. A feeding event with no unit cost currently contributes zero, so verify cost coverage before presenting margin.

Current missing milk observations can appear as zero in the headline/trend rather than unavailable. Check the milk-recording source whenever zero would be operationally surprising.

4. Read the performance charts correctly

Yield trend

The chart spans 90 completed farm-local days and shows a daily per-cow series plus a seven-day rolling average. The optional baseline is the average of available daily values during the first 30 days after farm creation and appears only when enough history exists.

Use the rolling line to distinguish a sustained movement from one unusual day. The current series inserts zero on a day without a milk record, and the rolling average includes it. Confirm recording completeness before treating a drop as lost production.

Milk revenue

The chart shows computed milk-collection payouts for the current month and preceding five calendar months. A blank or low month can mean missing delivery, price or payout data rather than low physical production. Compare it with bulk-tank collections and milk-price configuration.

Cost mix

The stacked chart uses the same six calendar months:

• Feed — feeding-event quantity × event unit cost;
• Drugs — treatment cost;
• Vet visits — veterinary-visit total cost; and
• Labour — weekly labour entries grouped by their Monday start month.

Null costs contribute zero. A week whose Monday falls inside a 30-day headline window contributes its whole amount even when the week crosses the window boundary. The footnote states whether any labour entry exists, but it does not prove that every week or other source cost is complete.

Estimated time saved

Select How is time saved estimated? to inspect the calculation. The last-seven-day operation counts use these assumptions:

Recorded operation	Assumed minutes saved per record
Milk recording	0.48
Manual alert resolution	2
PD-check scheduling	2
Treatment logging	4
Lifecycle event	2

Farmlync rounds category estimates, totals them and divides by seven for the daily average. Cumulative since signup extrapolates that current seven-day average over all days since the farm was created; it does not reconstruct historic time saved. Label both numbers as estimates.

5. Correct the owning record

If a result is wrong, correct its source rather than compensating on the dashboard:

1. Milk yield: check milk date, session, animal, quantity, active status and recording gaps.
2. Milk revenue: check buyer, milk price, route, collection quantity/quality and computed payout.
3. Feed cost: check actual feeding events and their quantity and unit cost. A stock receipt does not prove consumption.
4. Medicine cost: check the treatment and selected inventory product/batch. Do not add a second treatment to repair a cost.
5. Veterinary cost: check the veterinary visit’s total cost.
6. Labour cost: edit the existing week rather than creating a duplicate.

After an accepted correction, return to Farm performance and choose Refresh. Results may be cached briefly, while source invalidation normally prompts a current read.

6. Record and reconcile weekly labour cost

Use Quick-log labour on Farm performance or open Manage → Labour costs.

1. Choose one of the last 12 completed Week starting (Monday) values.
2. Enter a non-negative Amount in the organisation currency.
3. Add optional Notes of up to 500 characters, such as payroll, contractor invoice or owner estimate.
4. Select Log labour cost. Farmlync permits one live entry per farm/week; edit the existing row if that week already exists.

On the full Labour costs page choose Last 12 weeks or Last 52 weeks. Inline Edit changes Amount and Notes; Delete asks for confirmation and removes the row from performance calculations. Check the source evidence first because both actions can change margin and the cost chart. If the list unexpectedly looks empty, refresh before adding a row: the current page can display its empty message after a failed list request.

7. Investigate seasonal heat effect

From Farm performance choose Seasonal effect. The farm must have saved latitude and longitude.

1. Choose 14, 30, 60, 90 or 180 days.
2. Farmlync analyses the window ending two days before the farm’s current day because the historical weather source is delayed.
3. Read Heat-stress days at THI ≥ 72 and Severe-stress days at THI ≥ 78.
4. Compare average recorded yield on stress days with normal-weather days.
5. Use the time-series chart to inspect timing and the scatter plot to inspect association.
6. Read Pearson correlation only when at least 10 days have both THI and yield.

Correlation does not establish that heat caused a yield change. Feed, lactation stage, recording coverage, disease and management can move at the same time. The current page’s strong-correlation label uses causal wording; treat it as an association requiring investigation. If weather service evidence is unavailable, do not interpret a displayed zero heat-stress count as confirmed absence.

8. Use the right analytical page

Open Operations → Reports → Browse all reports and choose the page that matches the decision:

Question	Best evidence
Are current milk yield and quality changing?	Milk production and quality reports, Milk overview and Milk cohort analysis.
Are particular calving groups performing differently?	Milk Cohort analysis; compare DIM, cohort size, parity/breed/pen filters and record completeness.
Is fertility changing?	Breeding analytics; compare 3/6/12-month KPI samples, trend, cohorts, returns and cycle timing.
Which animals need action?	Work board, module worklist or an animal-level report; a herd KPI alone cannot identify the case.
Are withdrawals, treatments or medicine costs controlled?	Withdrawal calendar, treatment summary, medicines record and antimicrobial usage.
Is feed use economical?	Feed consumption, Income over feed cost and feeding-event source records.
Is inventory likely to run short?	Inventory burn-rate forecast plus current batches and pending deliveries.
Is a beef cohort commercially on track?	Beef performance economics, drafting lists, carcass and settlement evidence.
Which cows contribute margin?	Per-cow profitability, with its included milk income and assigned feed/health cost definitions.

Keep the displayed sample count and denominator with any percentage. In Breeding analytics, for example, heat detection, conception and pregnancy rates do not share one numerator, and some panels use the current eligibility roster rather than a frozen historical cohort. In Milk cohorts, filtered and unfiltered 305-day summaries can use different evidence. Use each page’s guide before comparing values.

9. Complete the review

Before sharing a figure, record:

• selected organisation and farm;
• farm-local period and observation time;
• currency, milk/weight unit and whether conversion applies;
• numerator, denominator and sample size;
• whether zero means measured zero or potentially missing input;
• which costs are included and which are absent;
• whether the value is a recorded fact, calculated projection, correlation or estimate; and
• the source page used to verify it.

The review is complete when an unusual result has been traced to its source, any correction was made in the owning workflow, and the explanation carries the period and limitations needed for another person to reproduce it. Continue with Farm performance and labour costs, Report catalogue, Breeding analytics or Milk Cohort analysis for control-level detail.