Guide · Freight
Booking freight is the visible job. The invisible one — confirmations, documents, invoices, and chasing money — is where a small brokerage's margin quietly goes, and it is the part that automates best.
Every step below is small. There are just a great many of them, each with a deadline, each attached to an email thread, and each capable of costing real money when it slips. Durations are typical for a small brokerage running email plus a spreadsheet; they are not benchmarks.
| Step | Owner | Typical duration | Failure mode when it slips |
|---|---|---|---|
| Rate confirmation out, signed copy back | Broker or ops | 10–30 min, same day | Unsigned rate con means the accessorial terms and free time are unagreed when you need them |
| Carrier vetting: authority, insurance, W-9, agreement | Ops or compliance | 12 min first time, 90 sec per load after | Dispatching against expired coverage; a claim with no cargo policy behind it |
| Dispatch and check calls | Ops | 5–10 min per call, several per load | No timestamps, so detention cannot be evidenced and is never billed |
| POD collection | Ops or billing | Minutes if it arrives; days to weeks if it does not | The invoice cannot be submitted at all. This is the single largest source of AR delay |
| Invoice generation and submission | Billing | 5–15 min per load, more for portal customers | Wrong format or missing reference, so the customer's AP system rejects it silently |
| Carrier settlement, direct or via factor | Accounting | Batch, weekly or per terms | Paying the wrong party after a factoring change; double payment on a re-brokered load |
| AR tracking, dunning, dispute resolution | Owner or AR clerk | Ongoing to 30/45/60/90 days | Nobody chased at 30 days, so cash arrives at 70 days instead of 45 |
| Reconciliation and close | Owner or bookkeeper | Half a day to two days monthly | Unbilled accessorials are discovered too late to bill, or never discovered |
Most back-office pain is a document that does not exist yet, in the hands of somebody who has no reason to hurry. Knowing who produces each one, and when it becomes blocking, is what turns chasing into a schedule instead of a panic.
| Document | Who produces it | When it is needed | What breaks without it |
|---|---|---|---|
| Broker-carrier agreement | You, signed by the carrier | Once, before the first load | No anti-re-brokering clause, no insurance minimums, no indemnity |
| Carrier authority record | FMCSA, pulled by you | At onboarding and re-checked per load | No evidence the carrier was authorised on the pickup date |
| Certificate of insurance | The carrier's insurance agent | At onboarding, monitored to expiry | No proof cargo coverage existed on the day of loss |
| W-9 | The carrier | Before first payment | Incorrect 1099 reporting and an unverified legal entity |
| Notice of assignment (factoring) | The carrier's factor | At onboarding and on any change | Paying the carrier directly when the factor owns the receivable, and paying twice |
| Rate confirmation, signed | You, signed by the carrier | Every load, before dispatch | The rate, free time, and accessorial terms are all arguable later |
| Bill of lading, signed at pickup | Shipper and driver | At pickup | No record of condition or count at origin, which is the whole fight in a damage claim |
| Proof of delivery, signed | Consignee and driver | At delivery | The invoice usually cannot be submitted at all |
| Lumper receipt | The facility | At the dock, in the moment | An unrecoverable pass-through cost, because reimbursement requires the receipt |
| Check-call log with timestamps | You, continuously | Throughout the load | Detention is unbillable because arrival and departure cannot be evidenced |
| Customer invoice with attachments | You | Per customer terms after delivery | Nothing to age, nothing to chase, no receivable |
| Accessorial backup | You, from timestamps and receipts | With the invoice | Accessorial lines get short-paid because they are unsupported |
| Leak | Mechanism | Cost |
|---|---|---|
| Missing POD | Invoice cannot be sent, or is sent and rejected | Days to weeks of delay on every affected load |
| Wrong invoice format | Customer's AP system rejects or ignores it | Re-work plus a full payment-cycle delay |
| Unbilled accessorials | Detention and layover never make it onto the invoice | Pure lost margin, invisible in the P and L |
| Late dunning | Nobody chased at 30 days | Cash tied up; occasionally never collected |
| Expired carrier insurance | Not caught before dispatch | Potentially catastrophic liability |
| Missing lumper receipt | Pass-through cost cannot be substantiated | The full lumper amount, absorbed |
| Duplicate carrier payment | Factoring change not tracked | The entire carrier pay, twice |
None of these are strategy failures. They are all attention failures — the predictable result of a small team handling a high volume of small deadlines through an inbox. The margin arithmetic that these leaks eat into is worked through in the margin and rate-per-mile guide.
Accessorials are the clearest case of money that is already earned and simply not collected. Detention is the most common, and it is lost for one reason: nobody recorded the two timestamps that prove it. Here is the arithmetic on a single illustrative load.
ILLUSTRATIVE LOAD - detention at delivery Appointment time 14:00 Driver arrival (check call) 13:48 Checked in at gate 14:05 Departed loaded/empty 19:40 Clock starts at the later of appointment and arrival = 14:00 Time on site = 19:40 - 14:00 = 5h 40m = 340 minutes Free time per rate confirmation = 2h 00m Detention time = 340 - 120 = 220 minutes Billed in completed 15-minute increments: 220 / 15 = 14.67 → 14 increments = 3.50 hours Customer detention rate $60/hr → 3.50 x 60 = $210.00 Carrier detention rate $50/hr → 3.50 x 50 = $175.00 Margin on the accessorial = $35.00 Daily cap per rate con $360 → not reachedNow the part that matters. Illustrative monthly arithmetic on a brokerage running 40 loads a month: suppose 6 of them incur detention, and 4 of the 6 go unbilled because the timestamps were never captured. That is 4 × $210 = $840 of customer billing gone, and 4 × $35 = $140 of pure margin gone, every month. Annualised, the margin alone is $140 × 12 = $1,680, and the billing is $10,080 that should have been collected and passed through.
The whole detention question is decided by two timestamps recorded at the time, not reconstructed afterwards. A check-call log is not administrative overhead — it is the evidence file for every accessorial you will ever argue about. The full mechanics of detention, demurrage, and the rest sit in the accessorials guide.
Which lines are billable, how free time is normally defined, and where demurrage differs from detention are covered in the detention and accessorials guide, and the calculators are free on the freight tools page.
Dunning fails for a stylistic reason more often than a commercial one. Early messages that sound like collections damage the relationship; late messages that sound apologetic get ignored. The escalation should be visible in the language, and the language should be written down once so it does not depend on whoever is having a bad week.
| Bucket | Status | Channel | Opening line | Escalation |
|---|---|---|---|---|
| 0–15 days | Current | None | No contact needed | Confirm receipt only if the customer uses a portal that can silently reject |
| 16–30 days | Approaching terms | Email, automated | Quick confirmation that invoice 10482 is in your system and scheduled | None. This is a receipt check, not a chase |
| 31–45 days | Past terms | Email, named sender | Invoice 10482 is now past the agreed 30-day terms. POD and rate confirmation are attached again for convenience | Attach the backup pre-emptively. Half of these are a missing-document problem |
| 46–60 days | Aged | Email plus phone call | Following up on invoice 10482, now 52 days out. Can you confirm the approval status and who it is sitting with | Ask for a name and a date, not a promise. Get the AP contact |
| 61–90 days | Seriously aged | Phone first, email to confirm | We need to resolve invoice 10482, 74 days outstanding. Is there a dispute we have not been told about | Owner-to-owner contact. New loads move to prepay or hold |
| 90+ days | At risk | Formal written notice | Formal notice regarding invoice 10482, outstanding 96 days | Credit hold, then third-party collections or legal. Decide the threshold in advance, in writing |
Two rules. Never let the first contact be a chase — a confirmation email at day 20 catches the silent-rejection cases while everyone is still friendly. And decide the credit-hold threshold before you have a customer sitting at 80 days, because in the moment the pressure is always to take one more load from someone who owes you money.
DSO is the average number of days between invoicing and getting paid. Every day of it is cash you have already earned, funded out of your own pocket, while you still have to pay carriers on much shorter terms. Here is the arithmetic, on illustrative numbers.
ILLUSTRATIVE BROKERAGE Customer billings $250,000 / month = $3,000,000 / year Revenue per day 3,000,000 / 365 = $8,219 / day Cash tied up in AR = revenue per day x DSO At DSO 52 days 8,219 x 52 = $427,400 At DSO 38 days 8,219 x 38 = $312,300 ------------------------------------------ Cash released by 14 days = $115,100 What those 14 days cost while you carry them: Financed on a line of credit at 12% APR 115,100 x 0.12 = $13,812 / year Or, if the same gap is bridged by factoring your own receivables at 2.5% 3,000,000 x 0.025 = $75,000 / year Fourteen days of DSO is not an accounting statistic. It is between $13.8k and $75k a year of real cost.This is why POD chasing is the highest-value automation in a brokerage, and it is not close. It is not that the task is hard. It is that every day the document is missing is a day of DSO, and DSO has a price per day that most small operators have never calculated.
The rule of thumb: automate what is repetitive, structured, and deadline-driven; keep humans on judgement and relationships. That selection filter, generalised beyond freight, is the subject of the small-business automation guide.
What should stay human: negotiating rates, deciding whether to keep a difficult customer, handling a claim, and any conversation where the relationship is the asset.
| Option | Upfront | Ongoing | Best when | Fails when |
|---|---|---|---|---|
| Do nothing, absorb the hours | $0 | The hours, forever | Under roughly 20 loads a month, one person, no growth intent | Volume grows, or the person who holds it in their head leaves |
| Hire an AR or billing clerk | Recruiting time | A salary, plus management | Volume is high and irregular, and exceptions dominate | Volume is high but repetitive; you have hired a person to be a script |
| Buy a full TMS | Setup, migration, training | Per-user or per-load fees | You want one system of record and can standardise your process to fit it | Your customers each demand a different invoice format the TMS does not produce |
| Buy point solutions and glue them | Low | Several subscriptions | One specific step dominates the pain and a good tool exists for it | You end up with five logins and no place that answers cross-cutting questions |
| Automate the specific workflows | Build hours | Near-zero per unit, plus maintenance | The work is repetitive, structured, deadline-driven, and customer-specific | The task is genuinely judgement-heavy, or the volume does not justify a build |
For most small brokerages the honest answer is a combination: a lightweight system of record, plus automation on the two or three workflows that actually consume the week. Buying a large platform to solve POD chasing is expensive, and doing POD chasing by hand at volume is also expensive; the middle is usually right.
Do these in order. Each step pays for the next, and the order is chosen so that the first thing you fix is the thing with a price per day attached.
Total build time in that plan is on the order of 45 to 70 hours spread over two months, alongside running the business. The measurement step is non-negotiable, because without it you will automate the workflow that annoys you most rather than the one that costs the most, and those are rarely the same.
Most small brokerages run on email plus a spreadsheet. It works until volume rises, and then it fails in a specific way: information exists but is not queryable. Nobody can answer "which loads are missing a POD right now" without a person reading threads. Every leak in the tables above is a version of that same problem.
The fix is not necessarily new software with a login. It is turning the email stream into structured records — load, document, status, deadline — and letting something watch that structure instead of a human watching an inbox. The moment a question like "which loads are past 45 days with no dispute logged" can be answered by a query rather than by a person, the whole class of attention failures stops being possible.
That is what FreightDesk AI does: it reads the billing inbox, drafts the invoice, chases the missing paperwork on a schedule, and hands a human the exceptions. The workflow-by-workflow case study — 18 AR and billing workflows — is in the HwyHaul writeup, and unfamiliar terms are defined in the glossary.
FAQ
Issue and collect a signed rate confirmation, verify carrier authority and insurance, run check calls through pickup and delivery, collect the proof of delivery, generate an invoice in the customer's required format, settle with the carrier or their factoring company, and then track the receivable to payment with dunning at 30, 45, and 60 days.
Most commonly a missing proof of delivery, an invoice submitted in a format the customer's accounts-payable system rejects, or nobody chasing at 30 days. Each of these delays a full payment cycle, and none of them are visible until cash is already tight.
The proof of delivery. Without it the invoice usually cannot be submitted at all, which is why POD collection is the step most worth automating in a small brokerage.
Multiply revenue per day by DSO to get the cash tied up in receivables. On an illustrative $3 million of annual billings, revenue per day is about $8,219, so cutting DSO from 52 days to 38 releases roughly $115,000 of cash. Financed at 12 percent that gap costs about $13,800 a year, and bridged by factoring at 2.5 percent it costs far more.
Because detention is proved by two timestamps recorded at the time, and small operations reconstruct them afterwards or not at all. On an illustrative load with a 14:00 appointment, a 19:40 departure, and two hours of free time, 3.5 billable hours at $60 is $210 of customer billing that simply never appears on the invoice.
Reading billing email and extracting load references and amounts, matching documents to loads, drafting invoices in customer-specific formats, chasing missing documents on a schedule, running accounts-receivable ageing and dunning, and flagging expiring carrier insurance or authority before dispatch.
Rate negotiation, decisions about which customers and carriers to keep, claims handling, and any exchange where the relationship itself is the asset. Automation should handle repetition and deadlines, not judgement.
Because information in an inbox is not queryable. Nobody can answer which loads are missing a POD right now without a person reading threads, so every deadline depends on someone remembering it, which is exactly where margin leaks.