Which channels produced orders, not inquiries?

Industrial marketing is usually measured on quote requests, because that is the number the website can produce. Most quotes never convert, conversion rates differ enormously by channel, and order values differ more.

Matching shipped orders back to the original inquiry ranks channels on the only figure that matters. It is common for a channel producing a quarter of the RFQs to produce most of the order value — and for the cheapest RFQ source to produce almost none.

The badge-scan list from a show and the lead list from a distributor are both exports. They only look untrackable because nothing joins them to what shipped.

Add either to the lead file, each row tagged with its source, and it becomes a channel like any other. For manufacturers this can reorder the picture entirely, because show spend is large, slow to convert and traditionally justified on anecdote.

Much of an industrial order is decided long before a purchase order exists. A design engineer downloads a model, drops the part into an assembly, and the specification is written around it. By the time procurement issues the RFQ, the competitive question has largely been settled and the inquiry that arrives looks like a formality. Marketing sees that inquiry and credits whatever produced it, which is usually a branded search. The activity that got the part designed in — the model library, the technical page, the catalog listing a distributor maintains — is nowhere in that report at all.

Those earlier touches are not untracked so much as unjoined. A CAD portal keeps a registration list. A gated datasheet keeps an email address. A configurator logs who used it. Each of those is a list of people with contact details, which is all a lead source has to be. Combined with the advertising exports into one lead file, each row tagged with its source, they rank on shipped order value like any paid campaign. The resulting order can be unflattering to the channel that was funded generously and generous to the one that was merely tolerated for years.

Counting inquiries keeps that ranking wrong. An engineer who downloads a model may raise nothing for a year, and when the order finally lands it carries a procurement contact rather than the engineer's details. A sales row carries one phone number and one email, so where the order record holds the engineer's email beside the buyer's phone, export both and the match takes whichever it finds. Where you hold only the buyer, the revenue sits in Direct / Unknown instead of being handed to whichever campaign ran that week. An honest gap is more useful here than a confident guess about who influenced a specification.

A great deal of industrial revenue does not arrive as orders in any ordinary sense. It arrives as call-offs against a blanket purchase order, or as scheduled releases under a supply agreement signed two years earlier. The commercial win was the agreement; everything after it is fulfilment against a schedule the customer controls. A report built on orders received will either treat each release as a separate sale attributed to nothing, or drop them entirely, depending on what the extract happened to include.

Neither treatment helps. Export the releases with the account's contact detail and they credit the channel that produced the original agreement, which is the only reading that matches how the business was actually won. A campaign that produced a single framework in March looks trivial against that month's figures and substantial across the three years the agreement runs. The gap between those two views is the difference between cutting a channel and funding it properly. Nothing in the underlying data changes when you do this. What changes is whether the release rows are present in the file at all, which is a decision made by whoever writes the extract.

It also changes what a quiet quarter means. Release volumes follow the customer's own production schedule, not your advertising, so a drop under a large agreement is a customer event rather than a marketing one. A column naming the order type, mapped as Product / service, keeps the two questions apart in the breakdown: whether marketing is producing new agreements, and whether the existing ones are being drawn down at the rate the forecast assumed. Blending them produces a trend line that responds to neither lever anybody controls.

Three findings are likely. The unattributed share will probably be large, because established accounts reorder and long-standing distributor relationships produce business no campaign touched. Aftermarket revenue — spares, wear parts, service kits — may turn out to be a meaningful fraction of the total, credited to whichever channel produced the account, and only an order-type breakdown shows how much of it there is. And the channel with the lowest cost per inquiry can sit near the bottom of the table once shipped orders rather than RFQs are what is being counted.

That last one is what moves budgets. Cheap inquiries are cheap for a reason: they come from people pricing a category across four suppliers, and an estimating team can lose days to them. An expensive inquiry from a narrow technical term may arrive four times in a quarter and carry the year. Ranking on order value shows both effects at once, and it usually ends an argument that had been running on conviction for several planning cycles. It rarely ends with a channel being switched off. More often the two are separated, given their own budgets, and judged against the outcome each was always producing.

None of this asks the business to work differently. Shipped-order records exist because they were invoiced. Inquiry records exist because somebody was expected to follow them up. Both were kept for reasons unconnected to measurement, which is precisely why a completed financial year can be analyzed on the first afternoon. A manufacturer can therefore find out whether the exercise says anything before committing to change a single process around it. That is also the cheapest available way to test whether the answer is interesting enough to defend in front of a board that has already heard several attribution pitches.