Marketing Influenced Pipeline: What Counts as a Touch

A revenue dashboard carries two tiles side by side. One says marketing sourced a modest share of the quarter's pipeline. The other says marketing influenced almost all of it. Both tiles are computed from the same CRM, by the same person, over the same ninety days, and the second one is the one that gets screenshotted into the board deck.

The gap between those two tiles is not a measurement error. It is the difference between two claims that happen to share a word, and the influenced number is the one that grows quietly until it stops carrying information. Getting value out of it depends almost entirely on one decision that most teams never write down.

Sourced and influenced are two different claims

Marketing-sourced pipeline names the origin of an opportunity. Something marketing did produced the first contact that led to the deal existing, so each opportunity has exactly one source and the sources partition the pipeline cleanly. It is a statement about creation.

Marketing-influenced pipeline names presence. A qualifying marketing interaction appeared somewhere in the account's history before the opportunity closed or advanced, so many programmes can qualify for the same opportunity and the shares do not add to a hundred. It is a statement about participation.

Neither is more honest than the other. The sourced number understates programmes that reliably progress deals without ever starting one, which is the job mid-funnel content and events are usually doing. The influenced number overstates everything, by design, because it was built to count presence rather than causation. What breaks is presenting one where the other was asked for, which is how a marketing team ends up defending a figure nobody in the sales organisation recognises.

Marketing sourcedOrigin

• One owner per opportunity, so shares sum to the whole pipeline
• Answers which motions create pipeline that did not exist before
• Right for channel investment and for deciding what to scale
• Blind to programmes that progress deals without starting them
• Moves only when the first-touch definition or the routing changes

Marketing influencedPresence

• Many programmes per opportunity, so shares overlap and exceed the whole
• Answers what appears in the history of deals that progress
• Right for arguing that a programme deserves to keep existing
• Inflates without bound unless the qualifying touch and the window are defined
• Moves whenever tracking coverage improves, with no change in the motion

That last line is the one worth sitting with. Improve email tracking, add a webinar platform integration, start capturing content downloads that were previously anonymous, and influenced pipeline rises. Nothing about the go-to-market motion changed. The metric measures the instrumentation as much as it measures the marketing, and a rise that follows a tooling change is evidence about the tooling.

The qualifying touch is the entire metric

Everything that makes influenced pipeline useful or useless sits in one definition: what counts as a touch. Left undefined, the answer defaults to whatever the CRM happens to log, and the CRM logs email opens, form fills, page views, webinar registrations, event badge scans and list memberships without any opinion about which of those indicates a person considered anything.

A definition that survives challenge names four things.

Which interactions qualify. Passive signals and deliberate ones are not the same evidence. A page view is a person arriving; a demo request is a person deciding. Teams that include everything end up with a number that means the account existed.

Whose interaction counts. An opportunity involves several people, and only some of them are on the buying committee. A touch on a junior analyst who never entered the deal is not evidence the programme influenced it. Deciding whether influence is measured at contact level or account level changes the result substantially, and it has to be decided once rather than per report.

Over what window. Influence claimed on an interaction from two years ago, on a person who has since left, is influence in name only. A bounded window forces the claim to describe the buying cycle rather than the customer database.

Before which event. A touch after the opportunity was created is a different claim from a touch before it. Both can be legitimate to count, and counting them together makes the number unreadable, because one is about creation and the other is about progression.

Is your influenced number defined

• ~Depends: The qualifying interactions are listed explicitly, and passive signals are in or out on purpose
• ~Depends: Contact-level against account-level influence is decided and stated
• ~Depends: The lookback window is bounded and matched to the sales cycle
• ~Depends: Pre-opportunity and post-opportunity touches are separated rather than pooled
• ~Depends: The rule is versioned, so a change to it is an annotated event rather than a trend
• ~Depends: Tracking coverage is reported next to the number, so an instrumentation change is visible

Three ways the number inflates

The inflation is rarely deliberate and it is always structural.

Any-touch scope. The loosest definition counts any recorded interaction of any kind by any contact at the account, at any time. Under that rule a company with good tracking and a long customer history will find marketing influence in nearly every deal, and the number converges on the share of the pipeline that has any digital footprint at all.

Unbounded windows. Without a lookback limit the metric accumulates. Every additional year of history adds touches and never removes them, so the figure rises year over year on a business that is doing exactly the same things.

Double counting across programmes. Because many programmes can claim one opportunity, summing programme-level influenced pipeline produces a total larger than the pipeline. That is arithmetically fine and presentationally fatal, and it is the moment a sales leader stops reading the tile. Report each programme against the whole rather than adding them, and say in the caption that the shares overlap.

What the number is actually for

Influenced pipeline answers one question well: which programmes are present in the deals that progress, and which are present only in the deals that stall. Used that way it is a coverage and allocation instrument, and a genuinely useful one, because it surfaces the mid-funnel work that a sourced number cannot see.

It answers a second question badly, which is how much of the revenue marketing produced. There is no honest version of that answer available from presence data, and pursuing it is what turns the metric into a credit argument. The functions then optimise for appearing in deals rather than for progressing them, which is cheap to do and hard to detect.

Two disciplines keep it honest. Report sourced and influenced together, permanently, so the gap between them is visible rather than load-bearing. And treat the gap itself as the finding: a wide gap says a lot of progression work is happening that nothing is crediting, which is an argument for keeping those programmes and a poor argument for scaling them, since presence does not tell you what would have happened without them.

It is also worth separating this metric from the attribution model that usually sits next to it. First-touch, last-touch and multi-touch models exist to split credit for one opportunity across several programmes, so they produce fractions that sum to one. Influence is a binary test applied per programme, so it produces overlapping counts that do not. Running both on the same dashboard is fine and common; letting a reader move between them without noticing is how a programme appears to own a share of the pipeline under one tile and a much smaller one under the next.

The reconciliation worth doing once a quarter is short. Take the opportunities the influenced number claims, sample twenty of them, and read what the qualifying touch actually was in each. The exercise usually finds two things: a handful of claims resting on an interaction nobody would defend out loud, and one or two programmes doing real progression work that the sourced number had written off entirely. Both findings are more useful than the headline figure, and neither is available from the tile.

1. Step 1Write the qualifying-touch rule

   Which interactions, whose, over what window, and on which side of opportunity creation. Version it.

2. Step 2Pick the level and hold it

   Contact or account. Both are defensible; alternating between them is what makes two reports disagree.

3. Step 3Report sourced beside influenced

   Two named numbers, never one figure answering whichever question was asked.

4. Step 4Print the coverage

   State what share of accounts has tracking at all, so an instrumentation improvement is not read as growth.

5. Step 5Annotate every rule change

   A definition change is an event on the chart, not a data point in the trend.

Where outbound sits inside this

Outbound is where the sourced and influenced distinction is easiest to get wrong, because an outbound-created opportunity at an account that has ever received marketing is simultaneously outbound-sourced and marketing-influenced, and both statements are true. The mistake is letting the second one quietly become a claim on the first.

The workable convention is that source belongs to whatever produced the first contact that led to this opportunity, and everything else is influence. An outbound sequence that reached a cold account creates the opportunity even when that account downloaded something eighteen months earlier, and the download is influence if the qualifying-touch rule says a download inside the window counts. Written that way the two numbers coexist without competing, and each keeps the property that makes any reported rate readable, which is that the population it was measured over is printed beside it.

The same discipline the rest of the reporting stack needs applies here. Every rate carries its population, the qualification stages underneath the pipeline hold their MQL and SQL definitions consistently, and somebody owns the join between the systems producing the two numbers, which is the argument for revenue operations as a function rather than a rota.

One practical caution about acting on it. A programme with high influenced pipeline and low sourced pipeline is a candidate for keeping, not automatically a candidate for scaling, because scaling presence does not reliably scale progression. Test it the way you would test any acquisition input, against the cost of the pipeline it can be shown to create rather than against the pipeline it appears in, which is the same discipline cost per customer acquisition demands of every other channel line.

If the constraint underneath the reporting turns out to be the number of qualified conversations rather than the attribution around them, you can see what a first campaign produces.

The short version

Marketing-influenced pipeline counts opportunities where a qualifying marketing interaction appeared in the account's history, regardless of what created the opportunity. Marketing-sourced counts origin. Sourced shares partition the pipeline; influenced shares overlap and can exceed it, so they are never summed.

The metric is worth exactly as much as its qualifying-touch rule, which has to name which interactions count, whose, over what window, and on which side of opportunity creation. Undefined, the number drifts upward with tracking coverage and history rather than with performance. Report both figures together, print the tracking coverage beside them, annotate every definition change, and use the influenced number to decide what to keep rather than to decide who gets credit.