Which channels produced booked orders and consumables?

A capital placement worth $40,000 can generate $30,000 a year in consumables for eight years. Judging the campaign that opened the account on the capital order alone understates it by an order of magnitude.

Matching on the account means every consumable order credits the original channel, which is the only ranking that reflects how the business is actually financed.

Clinical congresses and society meetings absorb enormous budgets and are evaluated on booth traffic and impressions.

The badge-scan list goes into the lead file, tagged with the congress. Ranked on booked and consumable revenue, a congress can be compared directly against digital spend — a comparison device marketers have rarely been able to make.

In most device markets the purchasing decision is preceded by an eligibility decision. A group purchasing agreement, a national framework or a hospital's own approved-supplier list determines which vendors a clinician may even specify, and that determination is made by a procurement function which never sees a campaign. A device business can run flawless demand generation into an account where it is not listed and produce nothing at all. The inquiries still arrive and still get counted, which is how a channel earns a reputation it does not deserve.

Splitting the revenue file by contract status makes the effect visible. A column naming the framework or agreement an order sits under, mapped as Product / service, shows listed accounts and open ones as separate lines, each with its own revenue and how much of it traces to a channel. It does not rank channels within each line, but if most listed revenue traces to a channel while open revenue does not, or the reverse, that is already a reason to fund the two jobs differently. Persuading a new account and reminding a listed department that the option exists need different creative, and they are almost always funded as though they were one, which hides the cheaper of them behind the more visible one.

The listing activity itself is measurable the same way. Tender submissions, framework applications and the accounts that invited you to respond are all lists of organizations with a contact detail attached, which is what a lead file is made of; tagged with their source, they rank beside everything else. Ranked on booked and consumable revenue, a framework that quietly delivered orders for four years stops being an administrative fact and may turn out to be one of the highest-return acquisitions the business has made. It also puts a number on the perennial question of whether the bid team is a cost center or a sales channel, which is otherwise answered by assertion.

Reps are the most expensive thing a device business buys and the least measured. A territory manager's calls, evaluations, in-service sessions and operating-room visits are logged somewhere — a CRM, a visit report, an evaluation tracker — and each of those logs is a list of clinicians with contact details. Nothing joins them to what the account subsequently ordered, so field effectiveness gets discussed in terms of activity levels rather than revenue. That is the same mistake marketing makes with inquiry counts, made with a much larger budget and much harder to raise without evidence.

Added to the lead file with visits as the source, the visit log ranks against the paid channels on the revenue that followed. The comparison is not a verdict on individual reps and should not be presented as one — territories differ and account mixes differ. Because each order is credited to the earliest record of that contact, it shows which activity first reached the accounts that went on to order, and a territory column mapped as Location shows where the attributed share is thin. Where the accounts a campaign opens are not the ones the field team visits, there is usually a routing problem rather than a performance one, and nobody fixes that without seeing it.

There is a caveat worth stating plainly. A visit log records where effort went, not where influence came from, and a rep calling on an account that was going to order anyway will appear to have produced the order. The report does not resolve that and does not claim to. What it does is put field activity on the same axis as everything else, so the question can at least be asked with a number attached to it. Comparing territories where field coverage differs sharply is the nearest thing to a controlled reading available, and it beats an activity count comfortably.

The first surprise is usually the size of the consumable tail relative to everything else. Once later orders credit the channel that opened the account, a campaign that looked marginal on capital placements alone can turn out to be the best-performing line in the budget by a wide margin. The effect is largest in accounts opened three or more years ago, which is why the chosen range matters more here than in almost any other category. Running a single quarter produces the answer the business already believes; running four years often produces a different one.

The second is how much sits in Direct / Unknown. Institutional purchasing generates orders that arrive through a portal with no marketing trace of any kind, and reorders placed by a department that has bought the same item for a decade were not produced by any campaign. Seeing that share at its real size is the point of printing it. A vendor reporting a high attributed share for a device business is redistributing that volume, and the ranking underneath cannot be checked. Nothing here is redistributed, which is what lets the attributed portion survive a follow-up question.

The third finding is about columns rather than channels. It is common to be able to split by product line but not by territory, or the reverse, because the two live in different systems and nobody has ever needed them in one file. Bringing both into the export — product line mapped as Product / service, territory as Location — is usually modest work in a reporting tool, and it changes the report from interesting to usable. It is the highest-value change to make before the second month's upload. Everything else stays the same: two exports, a mapping confirmed once, and nothing installed anywhere.