Every task in Barnyard Boss, in the order you meet them — plus the questions that come up, answered honestly. Where the app deliberately refuses to show you a number, this says so and explains why.
The app answers one question: what is each animal actually making me? Everything else exists to feed that. If you understand one thing before opening it, make it this:
Margin is what an animal is worth today, less everything it has cost you. Nothing is profit until it sells. The app never books an unrealised gain as income. That is the difference between this and a livestock app that reports your inventory value as though it were money in the bank.
That last step is the one worth doing before you rely on it in a yard. READY means the whole app is stored on the phone and will open with no bars. If it says NOT YET, open it once more on a connection.
While you are on that screen: Settings → Help → Field guide opens this page. It is the one thing in the app that needs signal, because it lives on the website rather than in the app — everything else works with no bars.
The sample farm is a real working farm of 40-odd head. Nothing you do to it touches your own records — you can re-run setup from Settings at any time and start empty.
If you filtered the register before tapping + — say to pigs, in Lot A — the form arrives already set to pigs, in Lot A. Change either if you need to.
A mob stops at 500 head, because past that it is almost always a typing slip — and 18,000 in the head box would build eighteen thousand separate animal records and several megabytes of them. If you really do hold that many, they belong in a group. The screen says so rather than just refusing.
The fields start empty. The greyed-out text in each one is an example, not a value — it disappears when you type. Nothing is entered on your behalf, including the price: leave it blank and the animal goes in with no cost basis, which is honest for something born on the farm or given to you.
Rows are skipped for two reasons and the count separates them, because the fixes differ. No tag — that row needs one. A tag already in the register — that row needs a different one, because a tag is the animal's identity and two animals sharing one are a single animal to the rest of the app: sell either and both leave. The animal you already have keeps its history.
No paper record on earth carries a cost basis — what each animal has cost you to date. Without it there is no margin. So the app asks rather than guessing: use a species average (each animal is marked estimated), one flat figure, or leave it blank and see no margin until the animal sells. If your file has a cost column, it uses that and the question only covers the rows where it is empty.
Opening basis is not a cost of this month. Stock you already own is not an August expense. The basis loads onto the animals and adds nothing to the month's net margin. Getting that wrong is what makes a farm's first month look catastrophic.
Prefer to type? Type it instead gives you a keypad with the same categories and allocations. The written path takes any phrasing and runs the same parser.
Where you say it lands is where it lands. Name an ear tag and the cost goes on that animal. Name a pen and it spreads across what is in it. Say nothing and it is farm-wide, split across every head by your allocation rule in Settings — by live weight, or evenly. That split is what makes a per-head figure mean anything.
“Twenty-one forty” is $2,140, not twenty-one and forty. “Four ton” is a quantity and is never mistaken for the amount. The parser knows the farm's own vocabulary rather than general English, which is why it can read a sentence you would actually say.
These are the costs you actually logged, allocated as you allocated them. An earlier build split the total by fixed percentages and labelled the slices feed, vet and labour — so "feed to date" was 26% of whatever the animal cost, and never feed. If a figure here looks wrong, the entry behind it is wrong, and you can find it in the ledger.
Sunlight? Big text scales the whole screen 1.18× and thickens the type. Everything stays — nothing is hidden to make room.
If the lot was treated recently, the sale screen warns you with the date it clears. The figures still calculate so you can see what waiting costs — but the warning is on the screen where the decision happens, not buried in a health tab.
What is due appears at the top of the screen, soonest first, marked OVERDUE, DUE or SCHEDULED. Nothing is due until you have recorded a treatment that repeats — the app does not invent a herd health programme for you.
Two treatments on one pen: the one that clears last is the one that counts. A 14-day drench and a 21-day booster on the same lot means that lot is held for 21 days, not 14, and the sale screen will name the booster. That is worth knowing because it is the direction that matters — a withdrawal shown shorter than it is puts meat where it should not be.
Delete a treatment and its withdrawal goes with it, so an animal held back by mistake can be sold again — and if another treatment is still running on that pen, its date takes over rather than the pen going clear. The cost stays on the month; take that off from Home if it was never spent.
The young go into the register. Each one becomes its own row, taking species and pen from the dam, with a provisional tag like G‑01/1 that is plainly not an ear tag — retag them at the chute when you are handling them anyway.
They arrive with no cost basis, and that is deliberate. Nothing born on your farm was bought, so there is nothing to subtract and no margin until it sells. The app will not put a birth value on a calf to fill the column.
A death is a cost line, not a hole in the average. Writing it off properly is what keeps the surviving animals' figures true.
Removing is not selling, and not a death. A sale books revenue. A death writes the cost off as mortality and lands in your loss figures for the year. A removal books nothing at all — it is for something that never should have been entered, and it leaves no trace. Reach for the right one: a mistake recorded as a death makes your mortality rate a lie.
Keeping an animal books the review and lets the real costs accrue, so the same comparison next month runs on actuals rather than a forecast.
The screen tells you what you have already spent, and then tells you to ignore it: that money sits in both columns, so it cancels out. What you have already spent cannot change which choice is better. It only changes how much you make or lose either way. That is a hard thing to accept about an animal you have fed all winter, and it is the most common way good stockmen talk themselves into keeping the wrong one.
On a farm with stock, Fill it from your own stock puts all eight numbers in from any animal you own. Unlike Cull or carry, this works on anything — an animal that is making money, a hypothetical, or stock you have not entered yet.
Records are held in this phone's browser storage. Clearing the browser's site data erases them — both copies. Losing the phone loses them. The backup file is the only thing that survives either. The app nags after 14 days, which should tell you how seriously to take it.
The ones that come up, and a few where the app is doing something deliberate that can look like a fault.
Yes. Once installed, the whole app lives on the phone and opens with no bars at all. Entering records never needs a connection.
Check it before you rely on it: Settings → This phone → Opens without signal reports the real state on that device — READY, INSTALLING, NOT YET or UNSUPPORTED. It can be refused by private browsing or some embedded browsers, and the app tells you rather than pretending.
On one phone, yes — properly. Set who is holding it under People & access and every entry is stamped with that person: the ledger shows “by Marta”, the CSV export carries an Entered by column, and People & access shows entries per person. A wrong number always has an author.
On two phones, not yet. There is no account and no server behind it today, so a second phone is a second, separate farm. Records do not merge.
That is what the Outfit tier is for — up to five people on one outfit, with a real conflict shown instead of a silent winner, backup off the phone, and control over who sees the money. The pricing page lists it at $149 a year and marks it not yet available, which is accurate: it is designed, the screens are built and you can walk through them, but there is no server behind them.
So the honest position is that the app already knows who did what. What it cannot yet do is carry that between two phones.
None of them. The app has no payment, no account and no gating today — everything in it is available to you, including the cost basis and margin work the pricing page puts on the paid tiers.
Tally (free) is the record book: every animal, no signal needed, the chute, voice entry, weights and treatments. Boss ($79 a year) is the arithmetic — cost basis building on every head, margin, cull or carry, profit by pen and species. Outfit ($149 a year, not yet available) adds more than one person.
Those tiers describe the intended shape, not what is switched on. If you are testing this now, you have all of it.
Because nobody has told it what that animal cost. Margin is value less cost — with no cost to subtract there is no margin, only a valuation.
Rather than show you the full market value and call it profit, the app shows a dash and says why. Put an opening cost on the animal and the margin starts computing. Until then it counts as nothing in your totals rather than as money you have not made.
They are their own category, Working animals, chosen when you add one. It covers anything kept for work rather than grown for sale: a heading dog, a cat earning its keep, a pack horse, a donkey running with the sheep.
They carry a margin like everything else — what they cost to buy and keep, against what they were bought for — and they appear in Profit beside the stock.
This changed. The app used to refuse them a margin on the grounds that a working animal is never sold. That is wrong often enough to matter: breeders sell pups, and a broken-in dog changes hands. So they can now be sold, priced whole rather than by the pound — nobody prices a collie at $2.40 a pound, so the sale screen drops the weight field and takes the price of the animal.
They stay out of Cull or carry on purpose. That screen weighs another month's growth against selling now, which is not what a working animal is for.
Yes, and the app is careful about which kind of undo you get, because they mean different things to your figures.
Correcting a ledger line keeps it and marks it amended, so the change can be traced. Deleting it removes it entirely — and if that line put animals in the register, it tells you how many and takes them out too. A sale can be deleted, which returns the animal at its old basis. An animal can be removed outright.
What the app will not do is quietly turn a mistake into an event. Removing an animal is not a sale and not a death: neither is booked, nothing lands in your mortality figures, and nothing is left behind saying it happened.
Into the register, one row each, taking species and pen from the dam. They get a provisional tag like G‑01/1 — visibly not an ear tag, so you know which ones still need doing at the chute.
They carry no cost basis, so no margin until they sell. Nothing born on your farm was bought, so there is nothing to subtract. An earlier build printed a birth value beside them and put nothing in the register at all; both were wrong and both are gone.
Yes — that is the app working. A cost logged against an animal, or against the pen it stands in, is money that animal has cost you, so it comes off its margin the moment you post it.
It did not always. Until recently a cost moved the farm's bottom line and the pen table and never touched the animal, so a margin was fixed on the day the animal was entered and stayed there for life. The add screen promised margin would move as feed and vet landed on it; now it does.
Its feed is showing against it. A ration logged to a house lands on the birds in that house, and for a flock on bought-in feed that is most of what it costs.
On the sample farm the layer ration alone is more than the layer flock's whole margin. That was always true and was simply not visible: the cost sat in the month's total and never against the flock. A negative number here is not a fault, it is the answer — and it is the reason to look at feed per bird rather than at the flock as a lump.
It will — the net margin on the home screen is exactly that. What it will not do is blend a per-head figure across species.
A broiler and a steer are not the same unit. An average across both is arithmetically true and operationally useless: it tells you nothing you can act on. So species and pen figures stay separate.
They are seeded figures or ones you enter — not a live market feed. Set your own under Settings, or per group on the batch screen.
Each species has its own rate, and the same table drives the sale screen, the importer and the cull decision.
Some animals are not priced by the pound at all. Working animals and equines are sold whole: the sale screen drops the weight field and asks what the animal fetches, because nobody prices a collie or a gelding per pound. The opening figure offered is what the animal cost you, which is the only thing the app actually knows about its worth, and you type over it. Cull or carry drops its growth assumption for them too — carrying a horse is not a weight bet.
Anywhere a price drives a decision, the screen says where the price came from.
Today's. Costs, sales, treatments and births are stamped with the day you enter them, the open month follows the calendar, and the Sunday close covers the week that just ended.
Withdrawal periods and breeding due dates count from the day of the treatment or the service — so a booster given this morning clears the right number of days from this morning.
Voice uses the browser's own speech recognition, which exists on Chrome for Android and Safari on iOS 14.5 and later. If the microphone is blocked or unsupported, the app says so rather than failing quietly.
Use Type it instead. It takes any phrasing and runs the identical parser, so you lose nothing but the talking.
No, and that is a decision rather than a gap. Handwriting recognition on a weather-beaten book produces plausible-looking wrong numbers, which is worse than no numbers.
The paper-book path sends you to mob entry instead: one row per lot with a head count and an average weight. Individuals get tagged properly at the chute, when you are handling them anyway.
Two reasons, counted separately. No tag: a row without one cannot become an animal, because weights, costs and treatments all hang off the tag.
A tag you already use: importing it would put a second animal on the same identity, and the rest of the app would treat the pair as one — sell either and both leave the register, with the second one's cost basis going too. The animal already in the register keeps its history, and the incoming row is left out.
Either way: fix the tags in your spreadsheet and import again, or enter those few by hand.
Correct, and the screen says so. The write-off is real — the money spent on those birds comes out of the flock's cost — but the app has no way to adjust a group's head count from that screen.
Correct the count in the register if it should change. The app tells you what it did and did not do rather than leaving you to find out later.
Nothing leaves the phone, so nobody can see it — there is no server to send it to and no account to sign into. But it is not encrypted either. Treat it as a notebook in a pocket: safe from the internet, not safe from someone holding the phone.
They are not gone, but they are on the old address. Records are tied to the exact web address they were entered on — that is how phone browsers work, and there is no way around it.
Open the old address, Back up to a file, then open the new one and Restore. Then stay on the new address.
Yes — plain CSV, from Reports → Export. Ledger, register, margin by pen, margin by type, or a period summary. It opens in any spreadsheet.
The figures in the export are exactly what the app shows. Nothing is rounded or re-derived on the way out.
You edited it after posting. The app marks it rather than changing it silently, so a figure can always be traced.
Superseded weights stay in the animal's history for the same reason, and deletions state their effect before they happen.
Onto every head, by your allocation rule in Settings. A cost you cannot pin to one pen — fuel, insurance, your own labour — still has to land somewhere or your per-head figures are wrong.
Name a pen or a house as you enter it (“vet, lot B”) and it lands there instead.
In the app: More → Settings, then Field guide at the bottom under Help. On any browser: barnyardboss.com/guide, which is the same page you are reading.
It opens your browser rather than staying in the app, so it is the only part of this that needs a signal. That is deliberate — the guide changes when the app does, and a copy carried inside the app would be the copy that went stale.
Read this before putting it in front of anyone else. These are not oversights to be quietly worked around — they are what the app genuinely is today.
No account, no cloud, no merging between two people's phones. A second phone is a second farm. The People & access screen and the two-person conflict flow are built and you can walk through them, but there is no server behind them.
That is the Outfit tier, and the pricing page says so: listed at $149 a year and marked not yet available. Everything else on this page works now.
Clearing the browser's site data erases everything on the phone, both copies. So does losing the phone. Take a backup weekly and send it somewhere else.
Prices are what you set. Withdrawal periods are tracked from what you record, not from a regulatory database — check the label. Nothing here is veterinary or financial advice.
The honest use for this build is one farm, one phone, one operator, with a CSV exported at month end. That is genuinely useful — and it is a real test of whether the ideas hold up in a yard.
A shot list, in the order that makes sense to a farmer who has never seen the app. Each one is short on purpose — people look up a single task, they do not watch a tour.
| Clip | What to show | The one line to say | Len |
|---|---|---|---|
| Install | Add to Home Screen, then Settings showing READY | “It lives on the phone. No signal needed.” | 45s |
| First animal | Stock → + → tag, pen, weight, price → added | “What it cost is the start of what it is worth.” | 60s |
| Import | Pick a CSV, remap one column, the cost-basis question | “A farm with sixty head is not going to type them in.” | 90s |
| Voice | Speak a real cost, show the parsed fields, post it | “Say it the way you would in the yard.” | 45s |
| Chute | A run of three, weight in, treatment, next animal | “Gloves on, arm's length, no signal.” | 75s |
| Cull or carry | Pick an animal, both columns, change feed, watch it flip | “What you already spent cannot change the answer.” | 2m |
| Sale | Sell an animal, settlement, realised against projected | “Now it is profit. Not before.” | 60s |
| Backup | Back up to a file, mail it, erase, restore | “A backup that lives on the phone dies with the phone.” | 60s |
Film on a phone, in daylight, with dirt on your hands. A polished studio recording of a field app is its own kind of lie — and the audience will know.