Funding What You Cannot Measure: A Budgeting Framework for Word-of-Mouth and Dark Social

Ask a CMO how much of the budget goes to paid search and you will get a number to the decimal point. Ask how much goes to word-of-mouth, private community spaces, and the Slack groups and group texts where buyers actually compare notes, and you will usually get a shrug. The spend happens — someone runs the podcast sponsorship, someone hosts the customer dinner — but it lives in a gray zone, funded by persuasion rather than a model.

That gray zone is where a growing share of B2B and consumer decisions are made. The trouble is that a channel you cannot prove is a channel that is always first on the chopping block when a quarter goes sideways. Teams need a way to fund the unmeasurable on purpose, not by accident.

Why the Untrackable Channel Always Loses the Budget Fight

Attribution systems are built to reward what they can see. A retargeting ad that intercepts a buyer who already heard about you through a colleague's recommendation takes full credit for the conversion. The recommendation — the thing that actually opened the door — receives nothing, because it happened in a DM, a community thread, or a hallway conversation.

The predictable result: teams over-invest in channels that capture credit and under-invest in channels that create demand. Over time, the provable channels get more expensive (you are paying to intercept buyers later in their journey), and the unprovable ones get thinner, quietly shrinking the pipeline the provable channels depend on.

The Fixed-Percentage Experimentation Model

The most workable answer we have seen across US mid-market teams is blunt: ring-fence a fixed share of working budget for channels that will never fully justify themselves in a dashboard. For most organizations, 10 to 15 percent is the right band. Below that, the investment is too small to compound. Above it, finance will (fairly) demand proof the discipline cannot provide.

The structure matters as much as the number:

• Lock it annually. Decide the percentage during planning, not mid-quarter. A budget that can be raided in week nine of Q3 is not a budget; it is a suggestion.
• Treat it as a portfolio, not a bet. Split the allocation across three or four untrackable plays — community programming, creator or customer advocacy, niche sponsorships, offline events — so no single unprovable bet has to carry the whole case.
• Define leading indicators, not conversions. You cannot measure pipeline from a Discord server, but you can measure qualified membership growth, repeat attendance, share of branded search, and direct traffic. Set targets on those.

Proxies That Survive a Finance Review

Fixed-percentage funding does not mean abandoning evidence. It means swapping conversion attribution for directional signals that a CFO can accept:

• Branded search volume as a proxy for word-of-mouth accumulation. When people hear about you in channels you cannot see, a portion of them Google you.
• Direct traffic and "none of the above" attribution responses. Rising unattributed conversions, paired with rising brand search, is the signature of dark social at work.
• Self-reported attribution. One field on your demo form — "How did you hear about us?" — is unglamorous and consistently among the most honest signals available. US teams that tabulate it quarterly routinely find community and podcast mentions outranking paid channels.
• Win/loss interviews. Sales conversations surface the private recommendations attribution never will. Ten interviews a quarter is enough to spot patterns.

None of these prove causation. Together, they make a defensible case that the unmeasured channel is not a void.

When the Model Breaks Down

Two failure modes deserve honest mention. The first is the experimentation fund becoming a slush fund — vague initiatives with no leading indicators at all. Every dollar in the unmeasurable bucket still needs a hypothesis and a review cadence. The second is the mirror problem: letting the provable channels shrink so far that short-term pipeline suffers while the dark funnel compounds. The 10–15 percent band exists precisely to prevent both drifts.

The Strategic Payoff

Teams that fund the dark funnel deliberately gain something subtler than incremental pipeline: negotiating leverage. When a third of demand originates in channels no ad platform can claim, your dependence on auction-priced attention drops, and so does your exposure to CPM inflation. In a market where every provable channel gets more expensive each year, the unmeasurable ones are where durable cost advantage hides.

You will never get a clean ROAS number for a customer recommending you at a barbecue. You can, however, decide in advance that such recommendations are worth funding — and set aside the money to earn more of them.