Demand Gen for Pipelines That Close in Six Months

If a property enquiry becomes a sale 90 to 180 days later, a Demand Gen campaign assessed on this month's cost per lead gets switched off in week five, every time. So the measurement design comes before the media plan: a geo holdout across matched markets, run for at least one full sales cycle, read on total enquiries and branded search volume rather than on what the platform claims it caused. The test design: Markets are paired on baseline branded search volume, organic enquiry rate and population, not on convenience. In Ireland that usually means holding out one or two counties against matched counties rather than splitting Dublin, which is too large a share of national demand to give up for eight weeks. The campaign runs everywhere except the holdout. The read is the difference in total enquiries per thousand population between the two sets, plus the movement in branded search impressions in Search Console and in the account. Platform-reported conversions are recorded and are not the decision. Why there is a floor: Below roughly fifteen thousand euro across the test period, the gap between the two market sets sits inside noise and the result cannot be defended in front of a board. That is arithmetic, not a statement about my fee. Where the budget is smaller, the honest answer is to put it into search and revisit this next year, and you will hear that on the first call rather than after the invoice. Audiences: First-party seed lists through Customer Match, segmented by stage: past applicants, viewing attendees, enquiries that went quiet. Lookalikes start at the narrow similarity setting and widen only when narrow exhausts. Existing customers, current students and prospects lost on price are excluded, which reads as obvious and is missing from most accounts I inherit. Affinity and in-market segments shape reach; they do not define the audience. Creative direction, not creative production: Demand Gen inventory is vertical, muted and scrolled past. The direction is specific: name the market in the first two seconds, put the price or the entry requirement on screen rather than in a voiceover, and treat the asset as the answer to one question instead of a brand film. I write the briefs and review the cuts. Your team or your agency makes them. What happens each month: The test runs and is monitored. You get a monthly note on the holdout read with the working shown, audience refreshes as CRM lists update, and briefs for the next flight based on what the current one did. Where the read comes back flat after a full cycle, the recommendation will be to stop. That recommendation has been made before and it will be made again. Not included: Video or still production. Landing pages. Attribution software. And any claim that Demand Gen caused a particular conversion, because the channel is not built to prove that and the tools that say otherwise are selling you something. Who this is not for: Businesses whose sales cycle is short enough to read in-platform. If enquiries close inside a fortnight, run this as an ordinary performance channel and skip the measurement overhead entirely. Anyone who needs the channel judged on last click. And anyone whose organisation cannot leave the holdout dark for a full cycle, because a holdout switched on in week four to quiet a regional manager ends the test and costs you the quarter as well as the fee.