Which channels produced signed engagements?

Consultancies win work by being visible and credible — a conference talk, a well-read report, a webinar. The pipeline that produces is real and is almost never connected to a signed engagement.

Attendee lists, download lists and webinar registrations can all go into the lead file, each tagged with its source. Included there, they rank on engagement fees alongside paid channels, which can show that the cheapest activity produces the largest work.

Consulting clients return. An initial diagnostic that becomes an implementation and then a retainer is worth many times its first invoice, and the campaign that produced it is credited for none of that.

Matching on the client means later engagements keep crediting the original channel, which is the only ranking that reflects how consultancies actually make money.

A meaningful share of consulting revenue arrives through other firms: a larger consultancy subcontracting a specialist workstream, a technology vendor bringing you into an implementation, a boutique passing on work that falls outside its capability. The client is somebody else's client, the introduction was a phone call between two partners, and the revenue lands in exactly the same accounts as everything the marketing budget produced, where it is counted without being distinguished. On a channel report it looks like unattributed demand, which is one of the reasons that bucket is so large in this segment.

Left unseparated it makes every channel figure on the page difficult to read. A list of the clients each firm or individual introduced, with the introducer as the source, can go into the lead file, and each named partner ranks on collected fees beside paid channels, content and events. The ranking may show concentration: one or two relationships producing most of the referred work, with a long tail that costs roughly as much to maintain. Partner-sourced work and marketing-sourced work are different businesses with different costs, and both deserve a number.

It also reframes a decision that partnerships teams tend to make on instinct. A reciprocal relationship that has produced nothing for two years appears as a row with no revenue against it. That is not automatically a reason to end it, because these relationships run on long horizons and reputations are not quarterly, but it is a reason to have the conversation with a figure in front of you rather than with an impression of how things are going. A quiet row is information; it just is not the kind of information that should be acted on within a quarter.

A consultancy can describe the same engagement three different ways: a contracted fee, a forecast number of days, or the amount actually invoiced once the scope moved. Different people inside the firm habitually use different ones, and an attribution report built on whichever was easiest to export ends up ranking channels against a quantity that shifts depending on who happened to run the export and in which week they ran it. The channel ranking then changes between exports, and nobody can say whether the market moved or the definition did.

Pick collected fees by client and month, and then hold to it. It is the figure that survives scope changes, unbilled time, discounts negotiated at the end of a project and engagements that simply stopped halfway through. It also means a retainer credits the acquiring channel every month that it runs rather than once at signature, which is much closer to how retainer-led firms actually earn their living across a year. It is also the figure a finance director will recognize, which matters when the report leaves the marketing team.

The consistency matters more than the particular choice. A report that used contracted value last year and collected fees this year will show channel movements that are entirely artefacts of the export, and those are precisely the movements people act on because they look like news. Fixing the definition once, at the point the file is first mapped, removes a whole category of argument from every subsequent year of reporting. Deciding it once is cheaper than relitigating it every time somebody new runs the report.

Many consultancies close between ten and fifty engagements a year. At that volume a monthly channel ranking is noise: one good win moves a channel to the top of the table, and its absence the following month moves the same channel to the bottom, and neither movement carries any information about the channel itself that anybody could act on with confidence. Reading a monthly report at that volume is how a firm talks itself into cutting a channel that was working perfectly well, on the evidence of one quiet month.

The right unit here is several years. Because there is no attribution window and the records exist for billing rather than for tracking, a firm can upload three or four years at once and read a ranking with enough engagements behind it to mean something. Deal count is shown next to revenue for the same reason: a channel with one large win and a channel with eleven modest ones should never be read the same way. Ranking a handful of large engagements without their count beside them is close to reading an average of two numbers.

The same caution applies to the unattributed share. In a practice built on reputation it will be large, and it will move between years for reasons that have nothing to do with marketing at all. Its level across several years is informative and worth watching. Its movement between two quarters usually is not, and treating the latter as a signal is how a firm ends up cutting the spending that was quietly working. Reputation moves slowly, and a report on it should be read on the same timescale as the thing it measures.