Control Centre
What needs you today, the headline against the same weeks last year, and the six questions.
What "like for like" covers, and what it leaves out
No dashboards yet
No dashboards yet. Ask an analyst a question and choose Add to dashboard.
Your dashboards
Every answer you chose to keep. A dashboard is re-run each time it is opened, so a tile is never yesterday’s number — and a tile that can no longer be confirmed says so.
Explore
What do you want to see?
Revenue
The numbers
Why the comparison window is 364 days back, not a year
The comparison window is 364 days earlier, not a calendar year. The same date falls on a different weekday each year, and a produce week is shaped by its weekdays — shifting a full year moves a Monday onto a Sunday and the comparison reads as a collapse that never happened. 52 weeks keeps the weekdays lined up.
What needs doing
Working it out…
How this list is ranked, and what each action does
Ranked by what it is worth. This week's decisions first, then the season's — a grower contract and a pallet ageing on the floor are both real, but they are not the same kind of urgent, so they are never ranked against each other. Acknowledge clears it from your Control Centre; decide, do or snooze it from here, and every action is dated in your own record above.
Data health
Every known fault in the feed, whether last night’s load ran, and what the feed contains — so a stale or distorted number has a named cause.
Known faults in the feed
Published rather than hidden. Every one of these is accounted for in how the figures are built — none of them silently distorts a number on another screen.
How current this is
Working out when the data last landed…
What the feed contains
Each family Radfords sends, and the span actually loaded.
Recent loads
The nightly job. A failure here is why a number would be stale.
Stock recorded over time
Radfords only ever shows stock as it stands right now, so this history exists only because it is captured every night. It cannot be recovered later.
Forecast track record
Every forecast this platform has published, checked against what then happened. Tested on history is the backtest over past seasons; live is what the nightly run scored after the week closed.
Sales Centre
Accounts, chains, price position, margin, cover, orders and service level — every figure against the same weeks a year earlier.
Accounts moving against you
Same weeks last year. Largest falls first.
Accounts & movement
Every customer against the same financial weeks a year earlier. Falls first.
Stopped buying
Traded a year ago, nothing in the matched weeks this year.
Started buying
No trade in these weeks a year ago.
By chain
Any customer, in detail
Pick a chain, then an account. Intercompany and forward-dated lines are left out of every figure here, and both amounts are stated below.
Every account
Accounts over $10,000 in either year.
Why stopped and started are counted over a longer window
Stopped and started are counted over more matched financial weeks than the eight above — the exact weeks, and how many, are printed beside each list. A produce customer can skip a fortnight without having left; a shorter window calls people lapsed who are simply between seasons.
Price position
What each account paid for the same product code, in the same financial week, as everyone else who bought it. Grade, size and pack type sit inside the product code, so what is left after that is price and not mix.
By channel
Above market means the channel paid more than the same fruit earned elsewhere that week. A channel has the volume to be compared even where a single account inside it does not, so the account count on each row is every account in that channel — a wider set than the accounts counted in the tiles above, which are only those over the volume bar.
Every account
Furthest below market first. Select an account to see it product by product.
Product detail
Same product code, same week, against every other buyer of it.
| Pick an account in the table above to see every product it bought, and what everyone else paid for the same code in the same week. |
What each channel actually buys
Share of revenue by grade. This is the other half of the price story: the independent trade takes the whole crop — including most of the Class 2, Class 3 and grade-outs.
Where prices are against last year
The last four weeks’ average selling price per kilo beside the same four weeks a year earlier, by produce — a different cut from the cards above, which compare accounts with each other inside one week. No price is forecast: these prices swing too much week to week for a prediction to be honest, so what is published is the typical weekly move and a bad week, both measured.
How the benchmark works, and why these figures are not money left on the table
The benchmark leaves each account out of its own comparison, and a week only counts when at least three other buyers took that product, at least a tonne of it remains once this account is removed, and this account was not more than 60% of it. That last guard matters: without it the worst-looking lines were the ones where the account was the market — one held 93% of a product week and was judged against the 728 kg left over. “Covered” is how much of the account’s actual trade cleared those tests, and it is never all of it. The percentage holds steady however strictly the guards are set; the dollar figure only covers what could be compared.
These figures do not add up to money left on the table. The market being compared against is RRF’s own selling, so whatever one account sits below, another sits above, and the book nets to roughly nothing by construction. What this screen ranks is placement — which accounts and channels sit on which side of RRF’s own average, and by how much. Moving one account up moves the benchmark for everyone else down a little. Trading below market is not by itself a fault: a clearance channel taking the tail of a pack run should sit under it, because the alternative is dumping the fruit. The figure is here so the size of that discount is a decision rather than a habit.
Chain accounts
Woolworths, Coles and Aldi, over any window you choose. The comparison is always the same weeks a year earlier — never the period before, because produce is seasonal and that reads a mango season ending as a collapse.
What do you want to see?
Week by week
By distribution centre
What they buy
What is counted
Service level
How much of what customers ordered actually went out.
Where orders fall short
Ordered by how many orders fell short; the percentage is the share arriving complete. Where a group also has orders that never despatched at all, the share with those counted as failures is shown beside it.
By financial year
Most often short-supplied
What is counted
On-time delivery is not shown anywhere in this product. The order file's required date mirrors the order date on most rows, so any on-time percentage built on it would be fiction.
Margin
Revenue less matched purchase cost. Worst first — that is where the money is.
How margin is calculated, and what is excluded
Cover & buying
What runs out first, what is backing up, and what is already committed.
Days of cover
Kilos on hand ÷ how fast it sells at this time of year = days left. Ranked by value.
This is a rate, and it covers every produce because it needs no forecast. For a dated answer with an earliest and latest case, on the produce the forecast can stand behind, see Stock & Freshness → Run-out.
What to buy, and by when
Packs on open orders, counted cumulatively — the 10-day figure includes the 7-day figure. Only orders still open. Select a produce line to see it broken down by product and customer.
Orders
What came in each week, what is open now, and what stock it commits — in packs.
Orders by week
Each full week against the same week last year.
Open orders, by the date they must go out
The book runs about a week ahead — that is as far ahead as anyone orders, so it is as far as anyone can see.
Confirmed orders ahead, by chain
Real orders already booked for the days ahead, in packs.
How far ahead each customer orders
Coles books about a week out. Woolworths’ paperwork doesn’t record notice, but its orders still show in the book. The market floor orders on the day.
Stock the open orders commit
Free is what is left to sell after open orders are packed. A dash means no stock in store — it will be packed to order, which is normal for chain lines.
Product detail
Pick a produce, then a product. Every figure is this financial year to the last complete week, against the same weeks a year earlier.
Which product
Every financial year
Whole years, not the eleven-week window above — clipping the older years to match it would make this year look like a collapse. The current year is marked, because it is not finished.
Where it ships from
This financial year to the last complete week. The sites add to the revenue on the strip above; nothing is left out of this split.
Who buys it
Every account that bought this product this financial year, largest first, with the same weeks a year earlier beside it. These add to the revenue above.
The next six weeks
Same week last year × how this year is running = the week ahead. Only produce whose measured miss cleared the accuracy gate is forecast at all — the rest are named under the table, because silence about a bad number is honest and the number is not.
Next week against the same week last year
Solid is the forecast for the week ahead; the outline is what the same week sold a year ago. The gap between them is this year’s pace.
By produce
Click a produce for the week-by-week.
By chain
Only what tested reliable. Aldi orders steadily enough to forecast; Woolworths and Coles jump around too much week to week for an honest number.
Sites
Each RRF site this financial year to the last complete week, against the same weeks a year earlier. Our own sites buying from us is out of every figure here, and its size is stated at the foot of the page.
Site by site
Largest first. A margin is measured on costed lines only and travels with how much of the site’s revenue that is; a site with no costed line reads a dash, never 0%. Select a site to open Explore narrowed to it.
Share of the book
Revenue this financial year to the last complete week, by site. The slices add to the book total on the strip; third-party stores are one slice, never folded into a site.
What is counted
Week by week
Stock & Freshness Centre
What is physically in store, how fresh it goes out the door, who grew it, and where any pallet came from.
Cover
Kilos on hand ÷ how fast it sells at this time of year = days.
Stock on hand
What is physically in store right now, aged and valued. Stock is held at more than one site — pick one to see it on its own.
Needs attention
Over 30 days old, or best-before inside 14 days.
Where it is
Always every site, whatever is picked above — so the split is visible first.
How long it has been sitting
Days since pack date. Darker is older.
By produce
Click a line to see the boxes — by product, grower and grade. “Est. from category” is how much of that row’s value is a category estimate rather than the product’s own realised price; it is not uniform, so it is per row rather than one figure for the floor.
How the estimated value is worked out
Run-out
When today’s stock runs out, at the forecast rate of sale — with the earliest and latest date the forecast band allows, not one date pretending to be certain.
What empties first
Only produce whose measured miss cleared the accuracy gate carries a run-out date. The rest are listed here too, with their kilos and their value, marked no forecast — stock you hold and cannot project is the thing most worth seeing, not the thing to leave off the screen.
The other way of counting it
Days of cover on Sales → Cover & buying divides what is on hand by how fast the produce sold the same four weeks last year. It is a rate, not a date, and it covers every produce because it needs no forecast. Where the two disagree, the season is turning — the day-of-cover figure is in the last column above so the two can be read side by side.
Freshness at despatch
Days from pack line to truck, by channel
Twelve months of despatched pallets. This is the shelf-life question asked the only way the feed can answer it — the best-before field is present on a quarter of pallets and, where present, sometimes holds the pack date instead.
By channel and produce
Slowest out the door first.
Growers
Judged against what the same produce earns from everyone else — not on dollars per kilo, which only measures who sends the dearer fruit.
Against their own produce mix
How “vs category” is worked out
Against the same variety, in the same week
The sharper question, and its guards
The table beside this compares a grower with their own produce mix, and “Citrus” holds both cheap juice oranges and premium navels. This compares Kensington Pride with other people’s Kensington Pride, in the same financial week, with the grower removed from their own benchmark. Three growers of that variety that week, at least a tonne on each side, and nobody who is more than 60% of the market that week — without that last guard the worst-looking line was a grower being judged against themselves. Coverage is how much of what they sold could be compared at all; the dollars apply only to that part.
This is not a verdict on a grower. An importer landing counter-seasonal fruit should sit above local supply, and a line taking the tail of a pack run should sit below it. The number’s job is to size the gap so it can be a decision.
Season calendar
When each grower usually starts and finishes, from three years of deliveries. The dates repeat; the tonnages don’t, so none are promised.
Trace
Farm to store. A pallet number, batch, pack run, grower or product.
Type a pallet id, batch, pack run, grower or product code above and this shows its whole history — the grower and farm it came from, the pack run and date, the site it was held at, and every despatch that moved it, with the customer and the order. Nothing is stored here; each trace is read from the feed when you ask for it.
Logistics Centre
What the trade costs to physically move — counted events, never a modelled dollar: consignments raised, pallets moved, kilos shipped.
Between the sites
Packed at, against sold from. → marks Eastern Creek supplying the floor.
Despatch effort
A despatch is one customer on one date — the real consignment. Every one of them in the feed is exactly that, with no exceptions, and the count agrees with the sales book to within 2%.
What each channel costs to move
Ranked by the number of consignments raised. A channel that books its revenue under a parent chain shows its physical work but no per-dollar figure — a ratio against roughly zero revenue would be meaningless rather than merely large.
Pallet fill
How full a pallet leaves the floor. A part pallet is not waste — a customer buying five cartons gets one by definition — but it is the clearest measure of how differently the two halves of this business are shaped.
Average load per pallet, by channel
By channel and produce
Lightest pallets first.
The despatch clock
What time of day the business actually ships. Read as one number it describes neither half of the trade: the market floor goes out early, small and often, and the chains go out late, large and rarely.
Hour of the trading day
By channel — and why the average misleads
The last two columns are the same question asked of lines and of dollars. Where they disagree, the channel ships its value at a different time from its paperwork.
Between the sites
Where each site’s packed fruit is sold
Despatched pallets, counted once each. A pallet joined straight to the sales book matches every line of its despatch, which multiplies both pallets and kilos — the sale is collapsed to one row per despatch first.
What Eastern Creek sends where, by grade
The two columns are the same packing line sorted two ways. This is the shape of the business, and it is why the floor and the chains cannot be compared on price alone.
About this flow, and why Victoria is not in it yet
This flow covers the two New South Wales sites. Eastern Creek packs and grades, and sells to the supermarket chains. Sydney Markets is the wholesale floor. Eastern Creek also supplies the floor — with good fruit, and with the lower grades that come off the line after grading. Everything else that appears as a “site” in the Radfords feed is a third-party cold store or carrier, and all of them have stopped.
Victoria is not in this table yet. Radfords delivered its pallet and order feeds on 8 September 2026, but not its sales, and this flow is built by matching pallets to the sale that shipped them. Victorian stock is on Stock & Freshness → Stock on hand, where it is named as its own site. The flow fills in on its own once the sales feed lands.
Pallet control
Hired pallets leaving the site, by account. Around 98% of RRF's pallets are CHEP, which are charged for until they come back.
Hired pallets despatched
Read the limit before acting on this. These are pallets going out. Whether one came back cannot be answered from this feed — pallet status records only the current state and it changes over time, so there is no return history to count against. This is exposure, not loss, and it is not a dollar figure.
Account & people
Your own sign-in, and — if you are an administrator — who else can get in.
Your sign-in
Everyone signed in can change their own password here. Use at least 10 characters. Nobody else can see or set it, including an administrator.
Add someone
Creates the account straight away and shows a one-time password once. Pass it on and have them change it above. Administrators can add and remove people; viewers can see every screen but cannot manage accounts.
Who can sign in
Changing a role takes effect immediately. You cannot change or remove your own account here — that stops the last administrator locking everyone out.
You are signed in as a viewer. You can see every screen; adding and removing people is limited to administrators. Ask Matt or Opal IQ if you need someone added.