Which channels produced signed security contracts?

Tender portals and bid services carry a real subscription cost and are justified by the number of opportunities they surface, not by the contracts actually won through them.

Added to the lead file and tagged with its portal, a bid list ranks on contracted billing against paid media — which may show that a portal producing hundreds of opportunities has produced almost no revenue.

Two contracts signed on the same day can differ tenfold in hours. A count of wins tells you almost nothing about which channel produced the growth.

Ranking on billed revenue rather than contract count corrects that, and it keeps correcting as coverage expands on an existing account.

A guarding contract is a baseline, and the revenue around it is not. Vacation and sick-leave cover, extra officers for a shutdown, event staffing, a temporary uplift after an incident, short-notice cover at a sister site — all billed at premium rates against an existing account, and none of it thought of as marketing revenue. It was produced by whatever won the account, because none of it would have been requested from a supplier the client did not already have. The billing system records it correctly and the marketing report has never seen it.

Exporting billed revenue rather than contract wins captures all of it automatically. Each invoice period becomes a row against the same account contact, and every hour of cover credits the channel that produced the client. The effect is largest on accounts with irregular demand — retail with seasonal peaks, venues, construction sites, anything with an event calendar — where ad hoc work can rival the scheduled hours across a year. A column separating scheduled from ad hoc, mapped as Product / service, shows the two as separate lines, each with how much of it traces to a channel.

The same logic covers coverage changes on a standing contract. A client adding a second site, extending patrol frequency, or moving from unmanned monitoring to a physical presence is expanding revenue on a relationship a campaign or a tender originally produced. None of those expansions looks like a sale in a CRM and most are never recorded as one. Billed revenue does not care how the growth was categorized internally; it simply shows more money arriving against the same account. That is often the difference between a channel that appears to have broken even and one that clearly paid for itself several times over.

The three main lines a security company sells reach different buyers by different routes. Manned guarding is a procurement decision at site or estate level, frequently tendered. Mobile patrol is bought by smaller premises on price and availability, often over the phone in an afternoon. Monitoring is a technical sale with an installation attached and a long recurring tail behind it. Their inquiry volumes, contract values and conversion rates have almost nothing in common, so a single cost-per-inquiry target across all three pushes budget toward the smallest accounts, quietly and every year.

A service column mapped as Product / service splits them, and the split is where the first real decision can come from. The report shows guarding, patrol and monitoring revenue as separate lines, each with how much of it traces to a channel, so a line whose revenue sits mostly in Direct / Unknown is visibly one the marketing is not reaching. It does not rank channels within each line, but it stops one blended figure from describing three different businesses at once. Each line is a different budget conversation, and none of them is visible in a blended report, which changes monthly for reasons nobody can explain.

There is an interaction worth measuring as well. Monitoring installations frequently lead to patrol contracts and guarding tenders at the same client, because the relationship exists and the account is already open. A channel that only ever produces small monitoring sales may be the entry point for the larger lines, which makes its apparent cost per dollar of revenue misleading in the useful direction. Because later contracts at the same client credit the channel that won it, a long range puts that patrol and guarding revenue in the entry channel's total. Read over a short range instead, the entry point looks marginal, gets funded accordingly, and the effect appears two years later in the tender pipeline.

Security contracts do not start when they are signed. Officers have to be recruited, licensed, vetted and screened; uniforms, equipment and site inductions have to be arranged; a mobilization plan has to be agreed with a client working to their own timetable. Three to six months between award and first invoice is ordinary, and on a large multi-site contract it is longer. Every attribution window in common use has closed well before the first dollar is billed, and the campaign behind the inquiry was assessed months earlier on the assumption that nothing came of it.

Because the match is on the account contact and there is no window, the timing makes no difference to the result. A contract mobilized in October credits whatever produced the inquiry in February, and the revenue continuing for the next four years credits it too. For a sector where the path from first contact to first invoice routinely exceeds a financial year, that is the whole difference between a measurable marketing function and an unmeasurable one. It also means the right unit of analysis is a year at minimum, because a quarterly read measures which contracts happened to mobilize.

One practical consequence: decide early whether the revenue file is dated to the award or to the billing. Both are defensible and they answer different questions — the first tells you when marketing did its job, the second tells you when the money arrived and at what rate. What is not defensible is mixing them across periods, because the resulting trend is a product of the export rather than of the business. Most providers use billing dates, since that file already exists and is reconciled. Whichever is chosen, write it down beside the export so the next person makes the same choice.