What a customs entry writer actually costs you

There is a reliable way to tell whether a business problem is real: check whether somebody has been hired specifically to solve it.

For customs entry data, they have. The job title exists. Search any US jobs board for "Customs Entry Writer", "Customs Brokerage Clerk" or "Import/Export Clerk" and you will find live postings, at real companies, with published pay bands. Recent listings advertise somewhere in the region of 24 to 36 dollars an hour.

That is not a market that needs to be convinced the problem exists. It is a market that has already priced it and put it on the payroll.

Turning a salary into a per-entry number

The hourly rate on its own is not decision-useful. What you want is cost per entry, because that is the unit everything else in your business is measured in.

Start with the time. Industry figures for a US import entry put the hands-on keying at roughly 45 to 60 minutes per shipment, working from a packet of commercial invoice, packing list, bill of lading or air waybill, and certificate of origin. Take the middle of that, call it 50 minutes.

Then the fully-loaded rate. The advertised hourly figure is not what the person costs you. Add employer taxes, benefits, paid time off, equipment and the share of management attention the role consumes. A common rule of thumb is 1.25 to 1.4 times base. Take 30 dollars an hour advertised and you are realistically at somewhere near 39.

Fifty minutes at roughly 39 dollars an hour is a little over 32 dollars of labour per entry, before anyone has checked the work.

The costs that do not show up in that figure

The per-entry labour number is the floor, not the total. Four other costs sit on top of it and none of them appear in a salary line.

Review time. Somebody senior checks the work, at least on higher-value or unfamiliar shipments. That is more expensive time than the keying it is checking.

Correction and re-keying. When a field is wrong, the cost is not one field. It is finding it, fixing it, and in some cases a post-entry amendment with its own handling time.

The risk tail. A miskeyed HTS code or declared value is not an administrative annoyance. It can mean a hold, a penalty, or a CF-28 that arrives long after everyone has forgotten the shipment and has to be reconstructed from whatever documents still exist.

Hiring and turnover. The role is repetitive, which is why it turns over, and each cycle costs you a recruitment process plus weeks of someone at partial productivity while they learn your lanes.

Doing the arithmetic on your own business

Three numbers, fifteen minutes:

1. Entries per month. You already know this.
2. Fully-loaded hourly cost of whoever does the keying, at roughly 1.3 times their advertised rate.
3. Honest minutes per entry. Not the best case. Ask the person who does it, and ask specifically about the shipments where the packing list and the invoice disagree.

Multiply them out and you have annual keying cost. Then put that figure next to your software spend, because most brokerages have never seen those two numbers side by side and the comparison is usually the wrong way round from what they assumed.

Why it has stayed this expensive

Not because nobody has tried to automate it. Because the part everyone automates first is the wrong part.

Reading text off a PDF stopped being hard some time ago. Any competent model does it. What has stayed hard is everything on either side:

• Reconciliation. Four documents describing one shipment, and they disagree about weight, piece count or value. Which one does your entry believe?
• Mapping. The extracted value has to land in the field your entry system expects, under the name it expects, in the format it expects.
• Validation. Do the line totals sum to the invoice total? Do the packages on the packing list match the bill of lading? Does the HTS code fit the goods description? Is the PGA data complete?
• Provenance. Eighteen months later, which page of which document did this number come from?

An extraction tool that hands you a spreadsheet has done the easy 20 per cent and left the expensive 80 per cent on the desk of the person you were trying to help. That is why "we tried an OCR tool" and "we still have an entry writer" are so often true of the same brokerage.

The number worth tracking instead

Cost per entry is the right unit, but straight-through rate is the better leading indicator: the percentage of shipments that need no human correction at all.

Get that measured honestly, even roughly, and every automation decision becomes easier to judge. You stop asking whether a tool can read an invoice, which they all can, and start asking how many entries it lets you file without anyone touching a keyboard.