Measuring ROI when the buying cycle is nine months long.

Stop asking last-click attribution to answer a question it cannot answer

Last-click attribution assumes the channel that touched the lead right before the deal closed did the work. In a nine-month cycle, that channel is often just the one the buyer used to book the final call, after months of research on other channels the report never credits.

Multi-touch models are not perfect either, but a simple linear or position-based model across the full journey will tell you a truer story than last-click ever will. The point is not precision, it is not systematically starving the channels that create demand in favour of the ones that harvest it.

Track pipeline velocity, not just pipeline volume

A campaign that generates leads which sit in a stalled stage for four months is not the same as one that generates leads which move steadily toward close. Segment your pipeline by source and measure how long each source's leads take to progress through each stage, not just how many leads arrived.

This is where most B2B teams get their ROI wrong. They celebrate a lead-gen campaign for volume while ignoring that its leads convert at half the rate and twice the time of a smaller, better-targeted campaign.

Use cohort-based reporting instead of calendar-month reporting

Grouping revenue by the month it landed and comparing it to the month's ad spend produces nonsense when the buying cycle spans three quarters. Group leads by the month they entered the funnel, then track that cohort's revenue as it closes over the following months.

This requires patience from finance and leadership, who often want a monthly number regardless of whether the business model supports one. Part of your job is educating stakeholders on why a nine-month cycle needs a nine-month reporting window, not a monthly scoreboard.

Build proxy metrics for the gap between spend and close

Waiting nine months to know if a campaign worked is too slow to act on. Identify earlier signals that correlate with eventual close, such as content engagement depth, number of stakeholders involved, or sales-qualified status, and use those as interim health checks.

Validate these proxies against actual closed revenue every quarter. A proxy that stops correlating with real outcomes is worse than no proxy at all, because it gives false confidence.

Report a range, not a false-precision number

Long cycles carry more uncertainty than short ones. Presenting ROI as a single decimal-point figure when half your pipeline is still nine months from closing overstates your certainty. A range with the assumptions stated plainly is more honest and more useful for planning.