ERP for Import Export SMEs: Why Small Is Not Just Scaled Down

The binding constraint is not features, it is bandwidth

A large company can absorb a bad ERP decision. It has an IT department, a budget line, and enough slack in the organisation that a difficult quarter is survivable. A small importer or exporter has none of that. The same wrong decision that costs a large firm a quarter can cost an SME a year, because there is no one to carry the extra load while the business keeps shipping.

That difference is not a matter of degree, it changes what the selection criteria should be. The constraints an SME actually operates under look like this:

• No dedicated IT. The project owner is usually the owner or the finance lead, who still has a day job.
• No written processes. The process lives in the owner's head, which means the implementation consultant has nothing to configure against.
• No redundancy in headcount. If the rollout interferes with shipping, there is no second attempt.
• No budget for a failed project. For a large firm, replacing a system is a reinvestment. For an SME, it is a wound.
• A single decision maker. Fast, but there is nobody to stop a wrong call.

Notice what is absent from that list: none of the constraints mention features. That is the whole point. An SME selection that starts from a feature checklist has already chosen the wrong lens.

Decision order: cut scope first, compare products second

Most SME selections begin with the question of which products exist. The correct order is the reverse. Decide first which two modules you will run in wave one, then look at who does those two modules well. The reason is that the primary failure mode is not choosing a weak product, it is choosing too much of a good one.

Practitioner accounts of this market converge on the same lesson. An owner who tried to bring up procurement, order follow-up, customs and finance as one project found the organisation simply would not move; splitting it into a first wave of order management and follow-up, a second wave adding customs two months later, and finance three months after that met far less resistance. The consultant's summary is worth remembering: the refusal is rarely about the software, it is about how much change the team can absorb at once.

There is also a volume threshold worth calculating before any of this. One commonly cited line of guidance holds that below roughly two hundred orders a year, forcing a full ERP in can actively reduce efficiency, and a lightweight inventory or even disciplined spreadsheets are the better answer. Whatever the exact number for your business, the exercise matters: count your documents per year before you count your requirements.

What belongs in the first wave

Order and document flow

The enquiry-to-quotation-to-order-to-purchase-to-shipment line, plus reusable document templates. The value here is not measured in efficiency percentages. It is that information stops living in one person's head and phone, which is the same reason a departing salesperson stops taking the customer list with them.

The money side: rebate, settlement, multi-currency

For a Chinese exporter this is not optional, because the export tax rebate and exchange handling are where the second half of the margin lives. If the first wave excludes the money side, you have bought a better filing cabinet. Our export tax rebate guide explains why the rebate base has to be the FOB value and what that implies for how orders are recorded.

What can wait

Complex approval chains, BI dashboards, custom mobile apps and deep CRM. None of these pay for themselves in the first year of an SME rollout, and each of them adds configuration surface that has to be maintained by the same person who is already short of time.

The point of sequencing is not to save money on licences. It is to raise the probability of a successful first wave. A first wave that works becomes the argument for the second; a project that fails because it tried everything at once usually ends with even the first wave rolled back.

The number that actually sets the budget

Budget is not an absolute figure, it is a function of revenue. One published reference point for Chinese trading companies holds that below roughly thirty million RMB of annual revenue, total spend should sit in the thirty to eighty thousand RMB per year range. The other common way to size it is by annual import and export value, with different product tiers matching different bands.

More important than the headline price is whether you have counted everything. Implementation fees, training, and the second year's renewal or maintenance charge are the three items most often missing from a comparison. Annual subscription models reduce the initial outlay but carry a renewal; perpetual licences tend to look cheaper in year one and more expensive later. The full set of line items is broken out in our ERP implementation cost breakdown, and where low-cost or free options genuinely work is covered in free and low-cost ERP options.

Seven ways SMEs lose money in this market

These failure patterns recur often enough in vendor and industry commentary to be treated as a checklist rather than a warning.

• Buying on price and finding out the low tier excludes customs, rebate and settlement — filling the gaps afterwards costs more than buying the complete set up front.
• Treating data security as an afterthought. Order and customer data are the core asset, and a system without layered backups and role-based isolation is exposed.
• Underestimating operating difficulty. With no dedicated IT, a complex interface becomes something staff route around rather than use.
• Being unable to adapt documents. If the system forces you to change your business process to fit the software, efficiency goes down rather than up.
• Weak post-sale support, which turns any problem into a stall in the shipping process.
• Being bundled into a full suite when you needed two modules.
• Believing a case study from a company that does not resemble you.

That last one is the most insidious. A reference customer with a full IT department proves nothing about what will happen in a company where the finance manager is also the network administrator. Comparability of scale is the only thing that makes a reference meaningful, which is the argument we make at length in why best-ERP lists are useless.

Judge the implementation team, not the brand

One piece of advice from practitioners deserves to be copied verbatim into your evaluation notes: not every foreign-trade ERP implementation team understands foreign trade. Some are pure technology backgrounds, and the difference between FOB and CIF is something they will look up after the meeting. The suggested test is to have the implementation team spend a day at your office and see whether they can hold a conversation with your salespeople.

This is why implementation capability carries heavy weight in our evaluation criteria: when two products score similarly on function, the industry experience of the team that will actually do the work is the deciding factor, because it determines how much of the analysis you have to do for them.

How to test it in a demo

For an SME the demo should be short and concrete. Bring your own documents and make the vendor do the work.

• Take a document you actually use — a customs declaration or a proforma invoice — and have them enter it live and generate the downstream documents.
• Ask for the go-live timeline for the two modules you named, and how much of your people's time it will consume during that period.
• Ask for a reference customer of your size, then ask that customer what the hardest part of the first month was.
• Have them demonstrate a rejected document being corrected, and show what else in the system is affected.
• Ask whether the rebate calculation produces the FOB base automatically rather than requiring the freight and insurance to be stripped out by hand.
• Get second-year numbers: renewal, additional users, additional modules.

When this becomes urgent

Small trading businesses usually decide to act when two of three signals appear at the same time: month-end reconciliation takes more than three days, customer history exists only inside individual salespeople's phones, and the owner has become the bottleneck in their own process. When two of the three are true, start selection. By the time all three are true, a customer has probably already been lost. Where your company sits on the wider growth curve is a related question, and our guide to when a growing company should upgrade sets out the usual thresholds.