Every marketing team that runs paid against named accounts eventually has to answer the question: what did this produce? And the honest answer — that a paid programme is one input among several into a multi-person, multi-month decision — does not fit in a board slide.
So we reach for influenced pipeline. It is a reasonable idea: find the opportunities that opened at accounts your programme touched, and sum them. The problem is entirely in the word "touched", and that word is almost never defined in the deck where the number appears.
The definition problem
Consider an account that saw one impression eleven months ago, on an awareness video, delivered to a junior analyst who is not on the buying committee. Nine months later a different team at that company opens an opportunity after a customer referral.
Under the loosest common definition — any impression, any time — that opportunity is influenced, and its full value lands in your marketing number. Nobody believes this. Yet it is the default in most reporting, because it is the definition that requires no decisions and produces the largest number.
A metric that nobody can restate the definition of is not a measurement. It is a negotiating position.
Four definitions, one quarter
Here is a worked example. Same company, same quarter, same 61 opportunities. The only thing that changes is what counts as a touch.
| Definition | Opps | Pipeline | Share |
|---|---|---|---|
| Any impression, any time | 61 | £4.20m | 100% |
| Any impression within 90 days of the open date | 44 | £3.05m | 73% |
| An engagement (click, video view, document open) within 90 days | 29 | £2.11m | 50% |
| Two or more job functions engaged before the open date | 17 | £1.34m | 32% |
Illustrative figures from a composite programme.
Nothing about the underlying reality differs between those four rows. The programme did what it did. The spread from £4.2m to £1.3m is entirely definitional, and if you are only ever shown one row, you cannot tell which one you are looking at.
Choosing yours
Three principles get you to a defensible choice.
Require engagement, not just delivery
An impression means an ad occupied space on a screen. It does not mean a person processed it. There is real evidence that unclicked impressions build recognition — that is the basis of most brand advertising — but recognition is not the claim being made when you attribute a specific opportunity. Requiring a click, a video view past a meaningful threshold, or a document open sets a bar that survives a sceptical CFO.
Set the window to your sales cycle, not to a round number
Ninety days is conventional and often wrong. If your median time from first serious contact to opportunity creation is five months, a 90-day window will systematically exclude touches that genuinely mattered. If your cycle is six weeks, 90 days will sweep in noise. Pull the actual distribution from your CRM and pick a window at roughly the 75th percentile.
Prefer committee breadth over touch count
Twelve impressions on one person is a weaker signal than three impressions each across finance, operations and engineering. A definition that requires two or more distinct job functions is harder to game with frequency and correlates better with a real internal conversation. It is also the definition sales finds most credible, which matters more than statistical elegance.
The metric that moves weekly
Influenced pipeline has a structural problem beyond definition: it arrives late and in lumps. A quarter's programme changes show up in pipeline one or two quarters later, by which point you have made three more rounds of decisions. It is a scorecard, not a steering wheel.
What you steer on is cost per account moved — spend divided by the number of accounts that crossed from one engagement stage into a higher one during the period. It has three properties influenced pipeline lacks:
- It moves weekly. Stage transitions happen continuously, so you get signal fast enough to act on.
- It is attributable to creative. You know which ads an account saw before it moved, so you can rank creatives by cost per movement rather than by cost per click.
- It cannot be inflated by definition. The thresholds are set once, and the number falls out.
It is also not revenue, and you should not pretend it is. It is a leading indicator with a known relationship to pipeline in your own historical data, which is the most any weekly marketing metric can honestly be.
How to present it
A structure that survives scrutiny, in the order the reader needs it:
- Coverage. What share of the target list was reached, by tier. Establishes that the programme ran.
- Movement. Accounts entering each stage, and cost per movement. The weekly-steerable part.
- Pipeline, with the definition stated in the same sentence as the number. "£2.11m influenced, where influenced means an engagement within 90 days of the open date."
- The strict comparison. One line. "On the two-function definition, £1.34m."
- What you are changing. Two decisions, with the evidence behind them.
The first time you present the strict comparison the number goes down and the credibility goes up. From then on, nobody asks whether marketing's numbers are inflated, because you have already answered it.