The one-page marketing plan

Who names the beachhead and the trigger; not for names three exclusions with a reason each. The promise names the alternative it beats, the value, and the category frame. Channel and test names the one channel in the centre and the one test running in the middle ring, with its kill line. Budget is a monthly figure, split by channel, with the CAC that split is expected to hit and where that expectation came from. Guardrail is the maximum CAC, derived from payback or lifetime value, with the rule stated. Stop is a number and a date: if this is true on that day, the plan changes.

Every line has to be checkable by someone who was not in the room. "Grow awareness among researchers" fails that test; "sponsor two research newsletters, $4,500 a month, expected CAC under $1,080" passes it.

Allan Dib's The 1-Page Marketing Plan (2016) uses nine squares in three phases, as listed on his company's site: before the customer knows you (target market, message, media), during the relationship (lead capture, lead nurture, sales conversion), and after the sale (a world-class experience, lifetime value, referrals). The labels here are reported from that site's course page, which was read only as a summary, so treat the exact wording as approximate.

The seven lines above cover his first phase in depth and touch the others through the guardrail. His template is broader, reaching into nurture, onboarding and referrals; the seven lines are narrower and harder, because each carries a number. Neither is "the" plan. Use whichever your approver will actually read, and keep the numbers either way.

Les Binet and Peter Field's The Long and the Short of It (IPA, 2013) analysed campaigns from the IPA effectiveness databank and, as widely summarised, found the best long-run results when about 60 percent of the budget went to brand building and about 40 percent to sales activation. The data are award entries from 1980 to 2010: campaigns that agencies chose to submit because they worked. That is a real body of evidence and a selected one.

For B2B, a 2019 follow-up for LinkedIn's B2B Institute is summarised in two different ways: one summary reports about 46 percent brand and 54 percent activation, another about 50 to 50 and calls the findings tentative. The original report was not read for these pages, so the disagreement is shown rather than resolved.

What a plan can take from this: the question is real, the ratio is someone else's average, and a small company with a tight beachhead where word of mouth works may reasonably spend almost everything on activation for a while. Write which side you are on and why; that sentence is more useful than either number.

In a tight market, the word of mouth that session 1 tested for does part of the brand-building work for free: researchers tell each other which tools they use. That is an argument for leaning toward activation early. It is not an argument for never building memory: once Cadence moves next door to product teams, who have not heard of it from peers, the free part stops working.

The brand identity side of this, the name, the voice, the visual system, belongs to the AI branding course. Here the question is only how the money is split, and the plan's job is to state the split it chose.

A minimum sample. Session 4's display channel closed 5 deals and its rate could have been anywhere from 2.9 to 14.7 percent. A stop rule that fires on five customers is noise. Cadence's rule waits for 10 Teams accounts and allows one extension; your threshold should come from how many customers your budget can plausibly buy in the window.

A threshold from the guardrail. The session 5 guardrail, at most $1,080 a Teams account (12 months of payback at $90 a month of gross profit), is the "keep" line. A second, looser line, $1,440 (16 months), separates "fix one lever" from "stop".

A date. Day 90, in the calendar of the person who owns the plan. A review with no date is the same as no review.

The line is usually credited to John Wanamaker, and sometimes to Lord Leverhulme. Quote Investigator's trace finds no direct citation of Wanamaker himself: the earliest evidence linking him to it is a 1919 speech by someone else, and the Leverhulme version comes second-hand through his son in 1931. David Ogilvy repeated it in 1963, crediting both. So: attributed, not established.

The question inside it, which half, is the subject of attribution: how to give each channel credit for a sale. That has its own course on the shelf. The plan's smaller, sufficient answer is to name, for each budget line, the measurement that will judge it, before spending.