"The general who wins a battle makes many calculations in her temple before the battle is fought. The general who loses makes but few calculations beforehand. Thus do many calculations lead to victory, and few calculations to defeat."
Sun Tzu, The Art of War, Chapter 1: Laying Plans.

Planning season is here. The templates will produce templates. Below are some things the templates won't catch, and what a CMO ought to be fighting for before the plan goes to the board.


1. A legibility line, not a legibility project

Most 2027 plans will list "AI readiness" as a project, with a kickoff, a workstream, and an end date. Legibility isn't a project. The machines read your surfaces continuously, and your surfaces drift continuously. The budget line for this belongs with investor relations, not with the website refresh. It's maintenance, in perpetuity.


2. A cohesion review, not a consistency audit

Consistency audits check whether your visual system looks the same across the surfaces you control. Cohesion checks whether what you say on those surfaces matches the evidence of what you do on the surfaces you don't control. The second is where the machine's gap-detection lives. The first is a 2015 exercise. Put the second on the plan. Make it quarterly.


3. An explicit choice: top or edges

The middle is being compressed in every category by AI. You can't plan your way out of this. You can plan to be at the top of the old skill as it becomes rarer and more valuable, or you can plan to find and own a long-tail niche the mass version missed. The 2027 plan should name which one. Both is a cop-out.


4. A metric that describes the medium

Review every KPI on the brand dashboard against the medium it was built to measure. Any metric whose medium has shrunk by more than half is a comfort, not a signal. TV ratings. Brand-tracker awareness among surveyed panellists. NPS. Replace or supplement each with something that describes the medium the buyer is actually in. Share of model. Discourse volume. Agent recommendation share.


5. A middleman scenario

For every category where your buyer currently encounters you through a human intermediary. A retailer, a comparison site, a buying agent, an analyst, a reviewer. Write the version where the agent has taken that intermediary's job. What does your acquisition cost look like. What does your retention look like. Who, specifically, in your organisation is responsible for that scenario arriving. Put the exercise on the plan.


6. A distinctive asset investment

The piece of brand that still compounds in the AI era is the fast-recall shortcut. Logo, colour, slogan, sonic, scent. The 2027 plan should name which of these you're deliberately investing in and which you're allowing to degrade. Most plans will leave this to the agency to handle quietly. It's a board-level question.


7. A proof-first comms strategy

The machine checks the claim against the evidence. If your communications function is still producing claims your operations can't verify, the gap is the number the machine is now generating. Put the operational data on the plan. Marketing and ops are the same review for 2027.

That's the plan the template won't produce. The template will produce a budget allocation across channels and audiences, a measurement framework that re-states last year's framework, and a positioning statement that could belong to five competitors.

If your 2027 plan has none of the seven above, it's a comfort. If it has three, you're ahead of most. If it has all seven, you're either a Plan B client or you should be.

Plan B works on 2027 brand plans that take account of the medium the buyer is actually in. Get in touch.