- Who is this for?
- Freight brokers, freight forwarders, operations managers running a brokerage desk and owners sizing a team
- Question addressed
- Calculate a broker’s capacity in loads per day and build the risk taken on each carrier into the result.
Short answer
A freight broker’s or freight forwarder’s capacity is net available time divided by the active time one load requires, from qualifying the request to invoicing the shipper. That time splits into decision steps (quoting, selecting the carrier, accepting a variance) and administrative steps (data entry, carrier documents, rate confirmation, follow-ups, CMR, invoice matching), and the split should be measured rather than assumed. Risk then enters as an expected loss per load: a fraudulent carrier, a claim or an unchecked carrier invoice is expressed as the number of loads’ margin needed to absorb it. The result is only meaningful when measured on your own flow, never from an industry average.
Key takeaways
- Capacity per desk = net available time ÷ active time per load, with timing boundaries fixed once and applied to every load.
- Separating decision steps from administrative steps shows where capacity is capped without touching the quality of the carrier choice.
- Risk is counted as an expected loss per load, then as equivalent loads of margin, using frequencies observed in your own history.
- A lever raises capacity if it cuts administrative time or expected loss; answering faster without vetting better moves time into claims.
- Measure a sample before any change, then the same scope afterwards, counting the review work that remains.
Break the load file into steps you can time
A freight broker, or a freight forwarder acting as intermediary, receives a shipper’s request, buys the transport from a subcontracted carrier and remains responsible for the load until it is paid. Revenue on each load is a gross margin: the rate sold to the shipper minus the rate bought from the carrier. Capacity is the number of loads the broker can carry in a day without error. To calculate it, split the load file into steps with a defined start and end, then time the active work in each one, meaning the time someone is actually working on that load.
The breakdown separates two kinds of steps. Decision steps commit margin or risk: setting a price, selecting the carrier, accepting or refusing a variance. Administrative steps produce or move information without deciding anything: entering data, collecting carrier documents, sending the rate confirmation (the document that fixes the agreed rate, locations, dates and conditions with the carrier), following up, retrieving the CMR consignment note (the international document signed at loading and at delivery that serves as proof of carriage), matching an invoice. Some steps mix both; the grid should then isolate each part.
| Step | Nature | Timing starts and stops | What consumes the time |
|---|---|---|---|
| Qualify the request | Decision, with some data entry | Opened → accepted, declined or completed | Reading, questions to the shipper, creating the file |
| Quote | Decision | Request qualified → rate sent | Buy rate estimated, target margin, reply |
| Find and vet the carrier | Mixed | Rate accepted → carrier selected | Calls, load board, documents requested and checked, choice |
| Confirm | Administrative | Carrier selected → rate confirmation accepted | Drafting, sending, acknowledgement, amendments |
| Track and retrieve the CMR | Administrative | Loading → CMR attached to the file | Status follow-ups, requesting the CMR or POD, checking its entries |
| Check the carrier invoice and invoice the shipper | Administrative, then a decision on the variance | Carrier invoice received → shipper invoice issued | Matching against confirmation and CMR, variance, attachments, issuing |
Calculate the theoretical capacity of one desk
The calculation takes three formulas. Net available time = length of the working day × share of time actually spent on loads, once unrelated calls, meetings and interruptions are removed; that share is measured, not guessed. Theoretical capacity = net available time ÷ average active time per load. Daily gross margin = capacity × average margin per load. If a loaded hourly cost enters the calculation, it comes from the company’s own management accounts, not from an industry average.
Fictional example, used only to show the calculation: a desk has 480 minutes a day, 65% of which go to loads, giving 312 net minutes. The grid measures 52 active minutes per load, split as below, with a fictional average margin of €90 per load. None of these values describes a real brokerage or stands as an industry reference; they exist only to make the formulas concrete.
| Fictional example line | Formula | Result |
|---|---|---|
| Decision time per load | Qualify 6 + quote 8 + select 5 | 19 minutes |
| Administrative time per load | Carrier documents 7 + confirm 5 + track and CMR 11 + invoice check 6 + invoice the shipper 4 | 33 minutes |
| Active time per load | 19 + 33 | 52 minutes |
| Theoretical capacity | 312 ÷ 52 | 6 loads per day |
| Daily gross margin | 6 × €90 | €540 |
| Administrative share of time | 33 ÷ 52 | 63% |
Build risk into the calculation
Theoretical capacity assumes every load ends as planned. A broker, however, takes a risk on every carrier it engages: a carrier that is not who it claims to be (carrier identity theft), freight that disappears after unauthorised double brokering, a delivery claim, a carrier invoice paid without matching it to the rate confirmation. These patterns are known across the profession; their frequency is only known from your own history. They cost money and time, and that time is taken from the next loads. The calculation therefore adds an expected loss per load and an incident time per load.
Expected loss per load = the sum, for each event type, of its observed frequency × its average cost. Frequency comes from the broker’s own history: number of events over number of loads in a period. The cost of an event adds lost margin, amounts paid or not recovered, and handling time valued at the internal hourly cost. To make the result readable, each event is converted into equivalent loads: cost of the event ÷ average margin per load, that is, the number of incident-free loads needed to absorb it.
| Event | Cost to count | Fictional example (€90 margin, €32 hourly cost) | Equivalent loads |
|---|---|---|---|
| Load lost to fraud (identity theft, unauthorised double brokering) | Freight paid out or deductible, lost margin, handling time, commercial time | €3,800 + €90 + 12 h × €32 = €4,274 | 4,274 ÷ 90 ≈ 47 loads |
| Delivery claim (delay, damage) | Credit or compensation, handling time | €400 + 6 h × €32 = €592 | 592 ÷ 90 ≈ 6.6 loads |
| Unchecked carrier invoice | Variance paid between the invoice and the rate confirmation | €120 | 120 ÷ 90 ≈ 1.3 loads |
Reading the fictional example: a single load lost to fraud consumes the margin of almost eight days of capacity at 6 loads per day. If the broker’s history shows one such event per 500 loads, the expected loss is 4,274 ÷ 500 ≈ €8.5 per load, close to a tenth of the fictional €90 margin. A check that adds two minutes to the load but lowers that frequency can therefore raise net margin while lowering gross capacity; the calculation has to keep both columns.
What genuinely raises capacity, and what does not
A lever raises capacity if it cuts active time per load without raising expected loss, or if it cuts expected loss without lengthening the load beyond what that saves. The step grid lets you test each lever against both columns. Levers that reduce administrative time are the safest: they leave the quality of the decision untouched. Levers that compress a decision step deserve a harder look.
Answering a quote request faster without vetting the carrier better is the typical false gain. The time saved at the selection step comes back, with interest, in the incident column. Likewise, confirming before the carrier’s documents have arrived, or paying a carrier invoice without matching it to the rate confirmation, shortens the load on paper and lengthens it in practice.
| Lever | Step affected | Effect on time per load | Effect on expected loss |
|---|---|---|---|
| Remove rekeying between email, TMS and rate confirmation | Qualify, confirm, invoice | Cuts administrative time | Neutral, or favourable if copy errors fall |
| Retrieve carrier documents without asking again | Find and vet | Cuts search time | Favourable if validity is checked on every load |
| Follow up on status, CMR and invoice without manual action | Track, CMR, invoice check | Cuts manual follow-ups | Neutral |
| Keep a memory of each carrier’s behaviour | Select the carrier | Cuts hesitation time | Favourable: the decision rests on what was observed |
| Answer faster without vetting more | Quote, select | Cuts decision time | Unfavourable: incident frequency rises |
Measure before changing anything
The sequence below produces a baseline capacity and a baseline expected loss. It takes a few days of observation, not a project. The result is then used to compare the same scope after any change of organisation or tooling, counting the review work that remains. Without that baseline, any discussion of a desk’s productivity rests on impressions, and a claimed gain can be neither confirmed nor challenged.
- 01Set the scope
Choose the desks, load types and shippers observed, then write down the exclusions: recurring lanes, batches, urgent loads.
- 02Fix the boundaries
Define the start event and end event of each step, and the rule for waiting time with no active work.
- 03Time a sample
Record active time per step over several days and time slots, separating the administrative part of mixed steps.
- 04Measure net available time
Over the same period, count the time actually spent on loads, interruptions deducted.
- 05Record incidents
Over a longer period, count fraud cases, claims, carrier failures and invoice variances, with the cost and time of each.
- 06Calculate and keep
Produce theoretical capacity, expected loss and equivalent loads, then keep the grid, readings and assumptions for the later comparison.
What Vigilo takes on in the load file, and what stays with the broker
Vigilo, the AI operator for the transport file developed by Axonovia, works on the administrative steps of the grid, within a scope defined with the team and with the tools already in place. It enters the shipper’s request into the TMS from the email or portal, retrieves and attaches the subcontracted carrier’s documents (operating licence, insurance certificates, the French URSSAF vigilance certificate that a principal must obtain from its subcontractor, company registration extract), flags what is missing or about to expire, follows up on status, CMR and invoice, then matches the carrier invoice against the rate confirmation and the CMR.
Decision steps stay with the broker: the rate, the choice of carrier, accepting or refusing an invoice variance. At selection time, Vigilo presents the carrier’s score, built load after load from measured signals (document completeness, respect of confirmations, CMR return, invoice variances, reported incidents). The score informs the decision; it does not make it. Suspicious emails (a look-alike domain, bank details that differ from the file) are flagged before a confirmation is sent, with no guarantee of detection.
No capacity gain is promised here. The method in this article exists precisely to measure it: the same grid, on the same scope, before and after, counting supervision time and the cases Vigilo hands back to the broker with the context already assembled. The result belongs to the flow observed; it does not transfer from one team to another.
Frequently asked questions
Direct answers
How many loads per day does a freight broker handle on average?
No figure is usable without measurement: the number depends on load types, shippers, the share of time genuinely spent on loads and how rigorously carriers are vetted. Calculate your own capacity: net available time divided by average active time per load, measured on a sample of your own flow.
How do you calculate whether a broker desk is profitable?
Start from daily gross margin (capacity × average margin per load), subtract expected loss per load multiplied by capacity, then compare the result with the loaded cost of the desk using your management accounts’ method. The result only means something for a defined scope and period.
Does finding a carrier count as administrative time?
Split it into two readings. Searching for candidates, collecting and checking their documents are administrative; the final choice, which commits margin and risk, is a decision. Mixing the two hides the part of the step that can be moved without touching the quality of the choice.
What does Vigilo do for a freight broker’s capacity?
Vigilo takes on the administrative steps of the load file: entering the request into the TMS, collecting and checking carrier documents, following up on status, CMR and invoice, and matching the carrier invoice against the rate confirmation and the CMR. It presents the carrier’s score at selection time without deciding in the broker’s place. Any gain is measured on each flow with the grid in this article; it is not promised.