Outgrown Spreadsheets? 6 Signs a Freight Broker Needs a TMS

A spreadsheet is not a stage you outgrow by headcount alone. Plenty of two-person brokerages run one carrier tariff in one workbook for years without a single incident. What breaks a rate workbook is not volume — it is the number of independent things it has to keep consistent simultaneously.

Count them on your own file. A working rate workbook holds the carrier buy tariff, the customer sell price, the taxable weight rules (loading-meter floor, volumetric ratio, rounding), the fuel surcharge in force this month, the price of each option, and the margin that applies to each account. Six moving parts, each maintained by hand, each with its own update cycle. Two are manageable. Six drift.

Read this table row by row, not as a total score. One cell in the right-hand column is enough, because each row fails in a different way and the failures compound. Four carrier tariffs across three grid types is a harder spreadsheet than twelve tariffs that are all weight-zone.

The six signs below are the observable symptoms of those failures. Each one is a test you can run this week on files you already have, and each carries a worked example of what it costs. If several apply, the question stops being "should we look at software" and becomes the sequencing question covered in our TMS selection resources.

This is the most expensive sign and the hardest to see, because nothing looks wrong. The quote is well formatted, the number is plausible, the customer accepts. The margin simply is not what you think it is.

It happens because the sell side of a spreadsheet is almost always a copy. Someone built a "customer price list" tab by pasting the carrier's rates and adding a markup column. From that moment, the tab is a snapshot. The carrier's next increase lands in the cost file, and the snapshot keeps quoting last quarter's cost.

Worked example: one missed rate increase

Take a groupage shipment of 3 pallets, 620 kg gross, into zone 3, on a weight-zone tariff. The target margin is 18%, applied with the standard sell formula p/(1−m).

• Current cost sheet — bracket 500 to 999 kg, zone 3: €18.40 per 100 kg. 6.20 × 18.40 = €114.08, plus a 9% fuel surcharge = €124.35 buy.
• Correct sell price — 124.35 / 0.82 = €151.65.
• Stale customer price list — still holds the previous €16.90 per 100 kg. 6.20 × 16.90 = €104.78, plus 9% = €114.21, then 114.21 / 0.82 = €139.28 quoted.
• Realised margin — (139.28 − 124.35) / 139.28 = 10.7%, not 18%.
• Gap — €12.37 per shipment. Across 12 shipments a month on that lane, €148.44 of margin gone, with nothing in the file to flag it.

The structural fix is not discipline, it is architecture: one engine that computes the buy price and then derives the sell price from it, so a sell price that does not match the current cost cannot exist. That separation between the stored table and the calculation is explained in rate grid vs rate engine, and the margin arithmetic itself in markup vs margin. In Cargavo, a TMS for freight brokers, the same engine produces both prices — the buy price without margins, the sell price with them.

Run the test literally. Give a colleague who does not normally quote a real request — 4 pallets, 1.2 × 0.8 m, 210 kg each, non-stackable, Lyon to Barcelona — and the workbook. Say nothing else. Then watch which questions they ask.

They will ask most of these, and each one is a rule that exists nowhere in the file:

• Which weight do I look up — the gross weight, or the chargeable weight after the loading-meter floor?
• This falls at 995 kg; is the 1,000 kg bracket minimum cheaper, and am I allowed to charge it?
• Is the fuel surcharge in these cells, or added after?
• Does this customer get 15% or 22%, and on which base?
• Is this tariff still valid, or did it expire on 31 March?

A spreadsheet stores numbers. It does not store the procedure that turns those numbers into a price. That procedure lives with whoever built the file, which is why brokerages of five people routinely have a pricing capacity of one, and why holidays and sick days show up as unanswered quote requests rather than as absences.

The specific rules that go missing are well known: taxable weight (loading-meter floor and volumetric ratio), the paying-for rule on weight brackets, rounding, the minimum charge, and which options are already in the rate. Each one is a stated parameter in a structured rate grid and an unwritten habit in a workbook.

Moving them into configuration changes what onboarding a colleague means. Instead of months of apprenticeship on which customer gets which markup, a new operator quotes correctly on the first day, because the rules are applied by the engine rather than remembered. In Cargavo, that is also a permissions boundary: an operator role can quote and book without being able to edit grids or margins at all. The rules that get lost in a spreadsheet are listed carrier by carrier in the carrier onboarding checklist.

Rate files multiply for ordinary reasons. The carrier sends a new PDF, someone retypes it. A colleague makes a local copy to test a scenario. A version travels by email to a laptop and comes back edited. Within a year the same tariff exists as TarifXYZ_2026.xlsx, TarifXYZ_2026_v2.xlsx, TarifXYZ_2026_MJ.xlsx and TarifXYZ_FINAL.xlsx, and the differences between them are three cells nobody can find.

The damage is not the duplication. It is that a spreadsheet has no concept of validity. A rate table is only meaningful with two dates attached — the day it takes effect and the day it stops — and a file has neither. So the file cannot answer the two questions that matter when a customer disputes an invoice: which rates were in force on 14 April, and which of them did we actually quote from?

The test

1. Search your shared drive for one carrier's name.
2. Count the files that contain rates.
3. Open the two most recent and compare the same cell — a mid-range bracket in your busiest zone.

If the two cells differ, every quote sent in the past month was priced from a coin toss. If they are identical, you have redundancy rather than divergence, which is survivable but still means every increase has to be applied in four places.

A structured grid replaces the filename convention with effective and expiry dates on the rates themselves, so an expired tariff stops producing prices instead of quietly producing wrong ones. The rebuild mechanics — merged cells, free-text zones, overlapping brackets — are covered in rebuilding an Excel rate sheet into a TMS. Cargavo, a TMS for freight brokers and transport commissionaires, keeps one active grid per carrier tariff, so "which version" is not a question the team can ask.

The dangerous errors are not typos. A typo produces €4,782 where €478 was meant and gets caught immediately. The errors that survive are the ones that produce a believable number.

Wrong weight looked up. The tariff is applied to gross weight when it should be applied to chargeable weight, or the reverse.

Wrong unit read. A rate expressed per 100 kg is multiplied as if it were per kg, or a €0.184 per kg cell is read as €18.40 per 100 kg in a workbook where both conventions appear on different tabs.

Approximate-match lookup. A bracket table read with an approximate match returns the row below the shipment weight when the boundary is set at the upper limit — one bracket too cheap, every time, on every shipment near a boundary.

Worked example: gross weight instead of chargeable weight

Four non-stackable pallets of 1.2 × 0.8 m, 120 kg each — 480 kg gross.

• Floor space: 4 × 0.96 m² = 3.84 m². Divided by 2.4 = 1.6 loading meters.
• With a carrier loading-meter floor of 1,850 kg per LDM: 1.6 × 1,850 = 2,960 kg chargeable.
• Priced on 480 kg — bracket 400 to 499 kg at €26.00 per 100 kg = €124.80.
• Priced on 2,960 kg — bracket 2,500 to 2,999 kg at €13.20 per 100 kg = €390.72.
• Understatement: €265.92 on one quote, on a shipment that looked entirely routine.

Note that the correct price is higher despite a much lower rate per 100 kg — which is exactly why the error is invisible to the person quoting. Check your own exposure with the loading meter calculator, and read linear meter (LDM) pricing explained for the floor and volumetric rules. In an engine, chargeable weight is derived from the dimensions before the lookup runs, so the wrong weight cannot be looked up.

Every brokerage knows this one and most treat it as a fact of life. It is worth measuring instead, because the measurement is easy and the result is usually uncomfortable.

Take last month's sent items, filter the quotes, and sort by the timestamp of the incoming request rather than your reply. Count how many arrived after 18:00, before 08:00, or at a weekend. Then count how many of those were answered within the hour. In most small brokerages the second number is zero, because the file was on a desktop and the person was not.

The delay is rarely the pricing itself. Timing the steps separately — reading the request, chasing missing dimensions, opening the file, doing the lookup, writing the email — shows that the calculation is a small fraction of the elapsed time, as broken down in the quote turnaround time benchmark. The rest is queueing: waiting for a person to be at a desk, with a file open.

There are two independent fixes and they solve different halves of the problem:

• Move the calculation to a server. Any authorised colleague can then price from a phone between appointments, because there is no file to have.
• Let the customer price it themselves. A branded portal answers at 21:00 with your rates and your margin for that account, and returns a structured request instead of "roughly 2 pallets, maybe 800 kg".

Cargavo, a TMS for freight brokers, includes the customer portal on every plan, from €49 per month; full white-label branding is on the Scale plan (€399 per month) — see plans and pricing. The portal is not a load board and it is not a tendering marketplace: it exposes your prices to your own customers only. What it changes for a small team is set out in customer quote portals for freight brokers.

Ask your team a question with a definite answer: how many quotes are open right now, for how much, and who owns each one? If answering requires several people to search their own mailboxes, the pipeline does not exist as data. It exists as recollection.

This costs money in three distinct places, and they are worth separating because only the third is obvious.

Nothing gets followed up

A quote sent on Tuesday and never answered is indistinguishable from one that was accepted, unless someone remembers it. The follow-up that wins the load is the one nobody makes.

Disputes are unwinnable

Three months later a customer says the agreed price was lower. Without a dated record linking the quote to the tariff version and the options selected, the discussion is one memory against another, and the broker generally concedes.

You cannot see which lanes make money

Margin per shipment is knowable only when buy price, sell price, customer and carrier are recorded together on the same object. In a mailbox they are four unrelated facts, so questions like "which customer looks busy but earns nothing" stay unanswerable.

The mechanism that fixes this is unglamorous: a reference. Every quote gets an identifier such as Q-2026-0311 with a status, an author, a date and the priced lines attached. When it is accepted it becomes a shipment with its own reference, S-2026-0184, carrying the same financial record forward. The message thread about the shipment hangs off the same object instead of living in one person's inbox.

Cargavo, a TMS for freight brokers and transport commissionaires, stores buy price, sell price and margin in a private area of each quote and shipment that customers can never read, while showing them their own sell price. That is a design requirement rather than a display setting — see how to show a price without revealing your buy rate.

Be precise about the boundary, because a vendor that promises to fix everything is describing marketing rather than software.

What stays your problem

A TMS prices what you negotiated. If your zone 4 buy rate is 12% above the market, a grid engine will apply that rate faster and more consistently than a spreadsheet ever did. Software also cannot invent a rate for a lane no carrier has quoted you — though it should still accept the request rather than dead-end it, which is the point of quoting without a rate grid.

Be clear on scope too. Cargavo, a TMS for freight brokers, does not connect to carrier APIs or EDI, is not a load board or a tendering marketplace, and is not a fleet, dispatch or accounting system. It prices the tariffs you already hold, in EUR, USD, GBP, CHF, CAD, MAD or PLN, in English, French, Spanish or German.

The migration itself is data entry, not an IT project: roughly one to two hours per tariff, with the timings broken down in the TMS implementation timeline. Plans are €49, €149 and €399 per month with a 14-day free trial and no credit card — see what each includes.