Farm Management Software: What to Look For Before You Buy

Most people looking for the best farm management software in the UK start with reviews. That’s the wrong order. A review tells you what one writer liked on one farm, and a five-star mapping package on a 600-hectare arable unit in Lincolnshire may be no use at all to a 180-cow dairy herd in Carmarthenshire with patchy signal and a Red Tractor audit in March.

Start with the checklist below and score every product against it. Then read the reviews, and you’ll be reading them to check specific claims, not just picking up a general impression.

There are ten points. They’re ordered roughly by how much grief each one causes when it goes wrong.

1. UK compliance outputs, by name

“Compliance features” on a sales page could mean anything. Ask which named records the software produces, in a format an inspector will accept without you retyping anything.

For most UK holdings that list looks like this:

Here’s a worked example. A mixed farm with 140 ha of cereals and 90 suckler cows runs a Red Tractor audit. If the software prints spray and medicine records as separate PDFs filtered to the audit period, the paperwork takes about 20 minutes. If it only gives you a raw activity log, you’ll spend a Sunday afternoon pulling it apart in Excel. Ask the vendor to send a sample audit export before the demo, not during it.

Location matters too. A product built for England may not handle ScotEID or EIDCymru, and a lot of New Zealand and Australian livestock apps sold here still don’t submit movements directly to UK databases.

2. Movement reporting that actually submits

This deserves its own line because it’s where livestock software most often looks better than it is. You have three days to report a cattle movement to BCMS. If the app “supports BCMS” by generating a file you then upload through the government portal yourself, you’ve saved maybe two minutes per movement.

Ask one question: when I record a movement on my phone, does it reach BCMS (or ARAMS, or ScotEID) without me touching a computer? Herdwatch and Breedr both advertise direct links for cattle in England. Check the exact scope for your nation and species before you sign.

3. Offline use on a phone, tested in the field

Every vendor says the app works offline. What that usually means is that it won’t crash. That’s not the same as letting you record a full day’s work and syncing it cleanly that evening.

You can test this yourself in ten minutes:

1. Put the phone in airplane mode.
2. Record a spray application, a calving and a field note with a photo.
3. Close the app completely, then reopen it.
4. Turn signal back on and see what syncs, what duplicates and what disappears.

If the field map won’t load offline, it fails. If you can’t look up a cow’s treatment history offline, it fails as well. That second one is the record you need at 2am in a shed with one bar of signal and a vet on the phone.

4. Data export, and what happens when you leave

This is the question vendors least like being asked, so ask it first: if I cancel, what do I get back and in what format?

A good answer is full exports in open formats: CSV for records, shapefile or GeoJSON for field boundaries, ISOXML for machinery task data. A bad answer is “you can download PDF reports”. You can’t import a PDF into the next system, so a PDF-only export locks you in.

Ask for a sample export of an animal register. A usable one looks something like this:

eartag,breed,sex,dob,dam_eartag,status,last_treatment,withdrawal_end
UK123456700123,LIMX,F,2023-03-14,UK123456500089,On farm,Footvax,2026-09-02
UK123456700130,LIMX,M,2023-04-02,UK123456300041,Sold,,
UK123456700147,BBX,F,2023-04-19,UK123456600112,On farm,Ivomec Classic,2026-10-21

It should have one row per animal, readable column names and ISO dates. If the export uses internal codes you can’t decode without the software, take that as a warning.

5. Integrations you’ll actually use

Integration lists get long because vendors like adding logos to them. Cut the list down to the systems you already pay for, and confirm each link works both ways.

• Arable: John Deere Operations Center, CLAAS telematics, Trimble, and RPA field boundaries via your SBI number. Pulling yield maps in and sending variable-rate prescriptions out is where the time saving comes from.
• Dairy: NMR or CIS milk recording, parlour software (Lely, DeLaval, GEA) and collar data from systems such as Afimilk or SCR/Allflex.
• Accounts: Xero, QuickBooks, Sage or Farmplan. This matters more than it did a year ago. Making Tax Digital for Income Tax started in April 2026 for sole traders and landlords with qualifying income over £50,000, and the threshold drops to £30,000 in April 2027. If your farm software can’t feed an MTD-compatible ledger, you’ll be keying everything twice.

“Integrates with Xero” can mean a live sync of every transaction or a CSV you upload once a month. Ask which one it is.

6. Support hours that match farming hours

Office-hours support, 9 to 5 Monday to Friday, is the default. That doesn’t fit a sector where problems turn up at lambing, at harvest and on bank holidays.

Before you buy, find out three things. When is phone support available, and does that include weekends from March to May and July to September? Is the support team in the UK, and do they understand the difference between BCMS and ARAMS? And how fast do they answer an email on a Saturday?

A practical test is to send a pre-sales support question at 6pm on a Friday and see when the reply comes. A vendor that’s slow while trying to win your business won’t get faster once you’re paying.

7. Total cost over three years, not the headline price

Pricing models vary a lot: per hectare, per head, per user, per module, or a flat annual licence. Some have free tiers (fieldmargin and Breedr both offer one) that are worth using for a season before you pay for anything. Others, such as Gatekeeper, are priced by quote.

Work it out over three years. The figures below are an illustrative example for a 250 ha arable and 120-ewe farm, not real quotes:

Option A looks cheaper. Then the farm takes on another 80 ha of contract land, and Option A goes up by £320 a year while Option B stays the same. Run the sum for the farm you expect to have in three years, not the one you have now.

8. Users and permissions

Your agronomist, vet, contractor and accountant may all need access, but not to the same things. You want the agronomist writing recommendations straight into the field record, the contractor marking jobs as done, and the accountant seeing invoices but not your gross margins.

Check whether extra users cost money (see the table above) and whether permissions can be set per person. Also check that the record shows who entered what. If an inspector queries a spray record, “a user” isn’t a good enough answer.

9. AI features you can check

Many products now advertise AI: disease risk alerts, satellite crop-health maps, yield forecasts, and photo recognition for weeds or lameness scoring. Some are useful. Omnia from Hutchinsons, for example, combines satellite imagery with soil data for variable-rate planning. Others are a weather feed with a new label.

There are three questions to ask about any AI feature:

• What data is it based on? UK trial data, or a model trained in the American Midwest?
• Can I see why it made a recommendation? A septoria risk alert should give its reasons, such as rainfall over the past 10 days, the variety’s resistance rating and drilling date. A score on its own isn’t enough to act on.
• Who owns the data it learns from? Read the data-sharing clause. Your yield maps are worth something commercially, so find out whether they’re being pooled and sold on.

If the AI recommendation saves one unnecessary T1 fungicide pass across 100 ha of wheat, at a rough £25/ha for product and application, that’s £2,500. That’s often enough to pay for the whole subscription. But you’ll only trust it enough to skip the pass if you can see the reasoning behind it.

10. A trial on your own farm, with your own data

Demo data always looks good: fields with tidy boundaries, a herd where every animal has a full pedigree, no missing tags.

Insist on a trial of at least 30 days, and spend it on real work. Import your RPA boundaries, load last year’s spray records and enter a week of calvings or lambings. That’s when you’ll find out whether boundaries come through cleanly, whether your old Excel herd register imports or has to be typed in by hand, and whether the person who’ll actually use the app every day (often not the one who chose it) can get on with it.

Putting the scorecard to work

Score each product 0, 1 or 2 on all ten points, giving 2 only where you’ve seen it working yourself, not just been told it works. Anything under 12 out of 20 shouldn’t make your shortlist, however good its reviews are. Treat a 0 on points 1, 3 or 4 as a fail even if the total is high.

Once you have three or four products left, read the detailed write-ups in our farm AI software reviews and cost comparisons and check each reviewer’s claims against your own scores. Where a reviewer praises offline use and your airplane-mode test failed, go with your test.

The trial is also the cheapest time to find out you’ve picked the wrong product. Book it for a quiet month if you can. November suits most arable farms; for spring lambers, try to finish by the end of January.