How to remove up to 60% of routine on a freight desk: audit manual operations, cost them in euros, and follow a practical 90-day plan.
A freight desk rarely grinds to a halt because of one big problem. It grinds to a halt because of two hundred small actions, each taking about ninety seconds: copy an address out of an email, read a weight out loud over the phone, dig a trailer plate out of a chat thread, retype VAT details into an invoice, call the driver again to ask where he is. None of these looks like a problem on its own. Together they consume the working day and leave the forwarder with no time for the part of the job that actually earns money - negotiating the rate and controlling risk.
Logistics dispatcher works with a digital dispatch system at a desk overlooking a container port
What follows is a working method that lets a haulier or freight forwarder remove between 30 and 60 percent of manual operations without buying a heavyweight system and without freezing operations for a month. It rests on one principle: measure first, remove what is unnecessary, and only then automate whatever is left.
Step zero: a five-day inventory of manual work
Automation without measurement is money spent badly. The company buys licences, and three months later the team is still keeping a parallel spreadsheet because "it's faster this way". The reason is always the same - the wrong thing was automated.
Take one week and ask every person on the desk to log operations. Not "what I did", but "which action I repeated and how many times". Three columns are enough: the action, how many times per day, roughly how long one repetition takes.
A typical picture for a desk of three operators looks like this:
| Operation | Repeats per day | Time each | Daily total |
|---|---|---|---|
| Scanning freight exchanges for a load for an empty truck | 40 | 2 min | 80 min |
| Copying data from email or chat into a spreadsheet | 25 | 3 min | 75 min |
| Calling the driver to ask for his current position | 18 | 2 min | 36 min |
| Retyping counterparty details into contracts and invoices | 12 | 6 min | 72 min |
| Hunting for a document in the mailbox or shared drive | 15 | 3 min | 45 min |
| Explaining the same trip conditions to different people | 10 | 4 min | 40 min |
| Assembling the payment document pack | 6 | 12 min | 72 min |
That is 420 minutes - seven hours of pure routine per desk, every day. Those seven hours are the actual subject of the work, not some abstract need for "digitalisation".
Two notes on measuring. First, do not collect the data from memory: people consistently underestimate short repetitive tasks and overestimate long ones. Ask for the log to be filled in immediately after the action. Second, mark separately every operation performed again because someone made a mistake - a call-back because the address was written down wrong, a reissued invoice because the VAT number belonged to a similarly named company. This is the most expensive category, and it is the one to remove first.
Put a euro figure on the routine
Management approves budgets when it sees a number. The formula for the annual cost of a single repeated operation:
Cost = (minutes per operation ÷ 60) × repetitions per year × fully loaded hourly cost
The hourly cost is not the net salary. It is the full cost of the workplace: gross pay, employer contributions, office space, hardware, connectivity. Across Central and Eastern Europe a freight operator's fully loaded hour typically lands between €18 and €30; in Western Europe the same figure is usually €35 to €55.
Take one operation - assembling the payment pack. Six times a day, 12 minutes each, 250 working days, at €24 per hour:
(12 ÷ 60) × (6 × 250) × 24 = €7,200 per year
That is one operation on one desk. Run the same calculation across all seven rows above and the total for a three-person desk usually lands between €35,000 and €70,000 a year. Against that number, the question of whether it is worth spending two weeks configuring templates answers itself.
Cost the errors separately. A wrong weight on the consignment note means a fine risk and a delay at a weighbridge. Wrong payer details mean a rejected invoice and payment pushed back by a week. A lost original consignment note means payment deferred by another 30 to 60 days, because most European shippers pay against a complete document set and not against a completed trip. In most companies the cost of errors exceeds the cost of the routine itself.
A priority matrix: what to automate first
Not all operations are equal. Score each on three parameters from 1 to 5: frequency, time to perform, cost of an error. Multiply them. Anything scoring 60 or above goes first.
Four things end up at the top of the list almost every time:
Manual data entry - the same information typed twice or three times: once in the email, once in the spreadsheet, once in the document.
Searching for information - time spent not on the action itself but on finding where the action was recorded.
Status communication - calls and messages whose only purpose is to find out the current state of play.
Producing standard documents - contracts, order confirmations, invoices, completion certificates, where only the details and the figures change.
And almost every time, the item management wants to automate first sits at the bottom: sophisticated analytics and "smart" vehicle matching. Useful, but it does not free up hours. Clearing small routine frees up hours.
Block one: finding freight without manual monitoring
The classic setup has an operator with several tabs open, refreshing lists every few minutes. That is 60 to 90 minutes a day of pure scrolling, of which perhaps five minutes produce a result.
Instead of refreshing pages, set a standing filter built around the criteria you actually work with: direction, body type, weight and volume, payment terms, minimum rate or rate per kilometre. The system then tells you when something matching appears.
In the Cargo Radar and Transport Radar the filter is stored in full, including loading and unloading zones and the radius around a point. Notifications arrive on the site and in Telegram, so monitoring stops being a separate job. Straightforward cargo search and transport search remain for one-off tasks when you need to read the market at a given moment.
Getting the configuration right is what keeps the tool from becoming noise:
Build several narrow filters rather than one wide one. "Poland to Benelux, curtainsider, 22 t, bank transfer" outperforms "anywhere in Europe, any body type".
Set a minimum rate threshold. Offers below your cost base should never reach you.
Keep the radius realistic. A 150 km empty run against a rate that does not cover it is a loss, not a load.
Every fortnight, review which alerts you keep ignoring. A category you ignore systematically should come out of the filter.
A well-configured radar removes most of those 80 daily monitoring minutes and, more importantly, eliminates the "missed it while I was at lunch" effect.
Block two: from the customer's email to the trip record
The second largest source of routine is moving data around. The customer sends the booking in free form: part of the terms in the body of an email, part as a voice note, the delivery address on its own line in a chat. The operator collects it and retypes it.
Three techniques remove this.
A single point of entry. The booking is created in one place - the order creation form - not across five messengers. A customer used to writing on WhatsApp needs to be told once: a booking counts as accepted when it is in the system. This is not bureaucracy, it is protection against "but we agreed on a different date".
Templates. In most companies the majority of trips repeat: the same lanes, the same commodities, the same terms. A saved template turns twenty minutes of form filling into two minutes of adjusting dates and weight.
Automatic repeat and bump. For regular lanes a booking can repeat on a schedule, and an already published one can be pushed back up the list without being retyped. Nobody has to remember each morning to repost a standing lane.
One point on data quality. Make the critical fields mandatory and do not allow a booking to be saved without them: exact loading and unloading addresses, the contact person at each point, payment method and terms, equipment requirements. Every field left empty here comes back later as a phone call, and sometimes as demurrage at the loading bay.
Block three: documents as a production line, not as craftwork
Documents are the most expensive manual operation in money terms and the easiest to automate. The logic is simple: a document consists of a fixed part (the template) and a variable part (party details, trip data, amounts). The variable part already exists in the system the moment the trip is created, so there is no reason to type it again.
In the documents section the pack is generated from booking and trip data: the transport order, the invoice, the completion certificate. Counterparty details are pulled from the company record rather than retyped, which removes the single most common error - the VAT number of a different company with a similar trading name.
What delivers the biggest effect in document handling:
One counterparty register as the only source of details. A single record per company, validated against the VIES database for intra-EU VAT numbers. Every document draws from it. Changed details are updated once.
Numbering without manual control. Sequential invoice and certificate numbering removes the duplicates that later have to be explained to accounting.
The payment pack as one complete set. The customer pays not for the trip but for a correct set of documents. When the invoice, the certificate and the consignment note are produced together and immediately, the payment term shortens by days and sometimes by weeks.
A recorded handover moment. The date and fact of sending documents belong in the system, not in the manager's memory. This settles most disputes about late payment before they start.
On transport documentation. For cross-border road freight within Europe, the CMR consignment note remains the document that evidences the contract of carriage and the condition of the goods at handover. Automation contributes two things here: the data for the note is not retyped, because it already exists on the trip, and the note itself does not get lost, because it is stored with the trip rather than in a folder in the cab. The electronic consignment note, e-CMR, takes this further - a digitally signed note reaches the accounts department the moment it is signed at delivery, which is precisely where the classic two-to-three-week wait for paper originals disappears from the payment cycle.
There is a compliance angle too. Under Mobility Package rules, operators have to evidence matters such as driver return cycles, posting declarations and cabotage sequences. When trip records, driver assignments and dates are captured once as part of normal work, producing that evidence during an inspection is a matter of retrieval rather than reconstruction.
Warehouse worker scans pallets with a handheld terminal before dispatch
Block four: communication in one place
The most expensive form of manual work is the one nobody counts as work: messaging. Trip terms are agreed on WhatsApp, an address change arrives on Telegram, confirmation goes by email, and the payment query happens on a call that leaves no trace.
The problem is not inconvenience. The problem is that the information is not attached to the trip. When a dispute surfaces a month later, nobody can reconstruct who agreed to what.
The workable approach ties correspondence to a specific booking or trip. In ChatLink the conversation with a counterparty is stored alongside the deal, so a colleague picking up the trip sees the context without needing a verbal handover. That removes the 40 minutes of daily re-explaining from the table at the top of this article.
The second half is status communication. The "where are you now" call should be replaced by a status the driver or the system updates. Each of those calls costs you two minutes and interrupts the driver - and interrupting a driver during his driving period is not a neutral act. Twenty calls a day is close to an hour, multiplied across both sides.
Block five: money and analytics without manual consolidation
The classic end-of-day ritual has the operator consolidating a spreadsheet: how many trips closed, what the margin was, who has not paid. It takes 30 to 60 minutes and is built entirely from data the system already holds.
When rates, costs and payments are recorded on the trip as the work happens, analytics and the finance section assemble themselves. Manual consolidation disappears, and with it the central weakness of hand-built reports - they are always slightly different depending on who prepared them.
The minimum set of figures a desk manager should see without asking anyone: margin per trip, margin per lane, average time from booking to confirmation, receivables split by ageing bucket, and trips per operator.
For load planning and for reading where demand is currently concentrating, the market map shows booking density by zone and helps avoid sending a truck into a region with no return load. For working out how the cargo actually fits in the trailer without a tape measure and a sketch, there is the 3D load planner.
What not to automate
Automation has a boundary beyond which it starts doing damage.
Do not automate rate negotiation. Price belongs to the person who can see the context: urgency, the customer's payment behaviour, the current balance of the lane.
Do not automate the decision on whether a new counterparty is trustworthy. A system can gather data and surface warning signals, but the decision to work with an unfamiliar company is made by a human being.
Do not automate a process that has not been defined. If nobody in the company has agreed who checks documents before loading, automation will simply lock in the existing disorder and make it run faster.
And do not try to automate everything at once. Companies that launch ten changes in a month typically roll back nine of them.
A ninety-day plan
Weeks 1-2. Measure. Log the operations, calculate the cost of the routine, build the priority matrix. The output is a list of five to seven operations that consume the most time.
Weeks 3-4. Master data. Clean up the registers: one list of counterparties with correct and validated details, one list of vehicles and drivers. This is the dull stage that determines whether anything else works. Automation on dirty data produces dirty output faster.
Weeks 5-6. Documents. Configure templates against real trips from the previous month. The completion criterion: the payment pack is produced without a single detail typed by hand.
Weeks 7-8. Search and bookings. Configure radar filters, build templates for repeating lanes, make the critical booking fields mandatory.
Weeks 9-10. Communication. Move trip correspondence into one channel and replace status calls with status updates.
Weeks 11-12. Verify. Repeat the measurement using the same method as at the start. Compare the numbers and adjust whatever failed to take hold.
The second measurement is a mandatory stage rather than a formality. It shows which changes the team accepted and which they are quietly working around. An operation people keep doing by hand despite a configured tool almost always means the tool is awkward, not that the people are wrong.
Why automation fails to stick
Parallel systems. Some data in the system, some in the old file. Within a month nobody knows which one is right. Either move the process completely or do not move it at all.
No named owner. If template configuration is not somebody's explicit responsibility, the templates will stay half-finished.
Ignoring resistance. The experienced operator who has worked his own way for twenty years resists not out of stubbornness but because his method produces predictable results. Show the effect using his own numbers: how long documents took him this week and how long they will take next week.
A start that is too ambitious. The larger the first release, the higher the chance of a rollback. One operation carried through to completion is worth more than five started.
Training on abstractions. A general presentation does not work. Walking through three of the company's own typical trips, from booking to payment, does.
Six metrics that show the result
Time from receiving a booking to confirming it. Measures processing speed; halving it is a realistic target.
Number of times the same data is entered manually. The target is one entry per piece of information.
Time to assemble the payment document pack. The fastest-reacting indicator of all.
Share of trips with a complete document set delivered on time. Feeds straight into receivables.
Number of status calls per day. Shows whether communication genuinely moved into the system.
Trips per operator per month. The summary productivity figure for the desk.
Record these before you change anything. Without a baseline, any improvement reads as a subjective impression, and the first difficult week will call the whole effort into question.
A checklist for auditing your desk
Work through the list and mark what is already true:
Counterparty details are entered once and flow into every document
Bookings are created in one place, not in messengers
Templates exist for repeating lanes
Freight and vehicle search runs on saved filters with alerts
The payment document pack is produced without manual typing
Trip correspondence is stored with the trip
Trip status updates without a phone call
Margin per trip is visible without building a spreadsheet
Receivables are visible with ageing buckets
Somebody is responsible for maintaining templates
Every unmarked line is a specific number of hours per month the desk spends instead of earning.
Automation in freight is not about technology and not about digital transformation. It is about letting the person who can negotiate and spot risk do exactly that, instead of moving addresses from an email into a spreadsheet. Start with the measurement, pick the single most expensive operation and carry it through - the desk will take it from there, because people notice a change in their own working day far faster than they notice it in a report. The scope available on each plan is set out on the pricing page.