CMO playbook & advanced tactics for brand messaging

A CMO running 200 campaign teams does not write the messaging. What that CMO writes is the rule about who may change it. Right now, somewhere in your organisation, a regional demand-gen lead is rewriting the core claim because it does not translate, a business unit is running a claim legal has never seen, and an agency two timezones away is reviving a proof point from a 2019 case study. Nobody is acting in bad faith. All of them are producing message drift, and drift is expensive in ways that never appear as a line item. This lesson is about the arbitrations that hold one claim in place across regions, units and thousands of assets, and what each dodged arbitration costs.

Core concept

At small scale, messaging is a writing problem. Past roughly fifty people and two markets it becomes a permissions problem. The claim-and-proof system the foundations lesson sets out rarely fails because someone wrote a weak headline. It fails because forty people had the right to write a headline and nobody had the right to say no.

So the artefact a CMO owns is the licence, not the words: a written statement of what is locked, what may be adapted, and by whom.

• Locked: the core claim and the category frame. Changed by the CMO alone, and not on a campaign cycle.
• Licensed: the value narrative. Regions and units may reorder, translate and re-express it, against a named reviewer and a fixed clock.
• Local: proof and execution. Local customers, local numbers, local humour, full autonomy inside the locked claim.
• Off limits: inventing a claim the locked set does not contain. That is an escalation, not a rewrite.

This is a ranking of edit rights, not a ranking of claims (the frameworks lesson supplies that ladder). The two get confused constantly, which is how organisations end up with a beautiful messaging document that nobody has the authority to enforce.

Key sub-concepts

1. WHAT DRIFT ACTUALLY COSTS

Drift almost never presents as a failure. It presents as slightly worse everything. Paid mediaPaid mediaVisitors arriving via paid ads or sponsored placements, where you pay a platform to display your message rather than earning visits organically.View full definition → pays twice to build the same memory when two claims run in one market. Sellers spend calls reconciling a website promise with a deck promise. Organic search self-cannibalises across near-duplicate pages. Worst, your brand trackerbrand trackerRegular measurement of brand health metrics (awareness, image, preference, and purchase intent) over time, so shifts can be detected and linked to marketing activity.View full definition → goes blind: when the claim was delivered inconsistently, you cannot tell whether it underperformed or was never really tested.

Run the arithmetic with your own headcount rather than trusting a benchmark. Two hundred campaign teams, each losing three days a quarter re-litigating the claim and remaking assets that already exist, at a loaded day rate of $600, is about $1.4 million a year in rework before a single media dollar moves. That number is usually smaller than the media waste sitting behind it, and it is the one your CFO will believe, because it is payroll.

1. THE ESCALATION PATH IS THE PRODUCT

Governance dies from silence, not from refusal. A regional team asks for an exception, hears nothing for a week, ships its own version, and now you have precedent that is harder to reverse than the original request. So the operative commitment is a written answer inside a stated window, even when the answer is no with a reason.

Salesforce runs an unusually literal version of this with V2MOM, the vision, values, methods, obstacles and measures document Marc Benioff introduced in 1999 and cascaded so that individual plans ladder into the one above. Whatever you think of the format, it gives language a distribution mechanism instead of relying on people reading a brand portal. (Salesforce sells marketing and CRMCRMCustomer Relationship Management: software and strategy to manage and analyse customer interactions throughout their lifecycle.View full definition → software, so it has a commercial interest in the discipline it preaches.)

1. PROOF EXPIRY AND THE CLAIM-MARKET PAIR

A proof library needs six fields, not four: claim, proof statement, source, date, expiry, and approved markets. Without an expiry, a stale statistic outlives its accuracy and one team eventually publishes it into a regulated document.

The edge case that breaks most centralised models: in regulated categories, marketing does not own the claim. An HSBC financial promotion needs compliance sign-off market by market, and a sentence that is lawful in one jurisdiction is unusable in the next. A Siemens Healthineers performance claim is bound to what regulatory clearance permits in each country. The practical unit of governance in these businesses is the claim-market pair rather than the claim, which means the locked tier gets narrower and the local tier carries substantiation duties, with an audit trail.

1. WHEN CENTRAL CONTROL IS THE WRONG ANSWER

Force one claim across genuinely divergent buyer sets and you get a claim abstract enough to be worthless. The test is blunt: could a direct competitor put their logo on this sentence without changing a word? If yes, the centre has over-consolidated, and the real positioningpositioningThe mental space you want your brand to occupy in your target customer's mind relative to alternatives.View full definition → work has been pushed down to units that were never funded to do it. The alternative is fewer locked words and stricter locked proof standards.

Real-world cases

Case 1: siemens

Roughly 300,000 employees, activity in around 190 countries, and business units whose buyers share almost nothing: plant automation engineers, rail operators procuring signalling, hospital administrators, building owners. Siemens has held a single corporate claim over long periods ("Ingenuity for life" from the mid-2010s, then "Transform the everyday" from 2021) while letting divisions carry their own proof and vocabulary underneath it. The instructive part for a CMO is the counter-move: Siemens Healthineers was given its own name in 2016 and its own listing in 2018. Some units do not need better alignment, they need separation, and that call belongs to the architecture question the brand strategy foundations lesson sets out. Deciding it late is what leaves you refereeing claim disputes for years.

Case 2: hsbc

"The world's local bank" ran from 2002 and was among the most disciplined global claims in financial services: it licensed local expression by design, because local expression was the claim. It was retired around 2011, and the reason was structural rather than creative. HSBC was withdrawing from markets, eventually going from a footprint spanning more than 80 countries and territories to roughly 60, exiting US and French retail banking along the way. The promise had outrun the operating model, and the bank moved on, arriving at "Together we thrive" in 2018.

The second-order cost is the one leaders underprice. Retiring a decade-old claim means retraining tens of thousands of frontline staff, reworking signage across thousands of branches, and writing off accumulated recall that competitors did not have to pay for. When you weigh a refresh, price the change management alongside the creative, and price the recall you are deleting.

Case 3: Salesforce

Salesforce built its early position by attacking software itself, coining "the end of software" and staging protests outside a Siebel user conference in the early 2000s. That claim worked until it stopped being true: a company past $34 billion in annual revenue, selling through acquisitions including MuleSoft, Tableau and Slack, cannot run on an anti-incumbent frame.

So the ongoing arbitration is which acquired brands keep their own voice. ExactTarget was absorbed and renamed Marketing Cloud. Tableau and Slack kept their names and much of their tone. Every one of those decisions determines how many messaging systems the CMO has to fund, staff and police, and each retained brand is a permanent line in the marketing budget rather than a one-off integration task.

CMO action items

• Run a drift audit before your next planning cycle. Pull the homepage headline, the top sales deck, the last five paid campaigns and the local site for your three largest markets, then count distinct core claims. More than two, and you fund reconciliation before you fund new creative.
• Publish the licence in one page: what is locked, what is licensed, who reviews, and the response window. Name a single person with veto authority and a named deputy, because an absent owner is the most common source of unauthorised claims.
• Add expiry dates and approved markets to every entry in the proof library, and give one person the quarterly job of retiring stale entries rather than adding new ones.
• Instrument drift so it has a number. Track how many claim exceptions were requested, granted and refused each quarter, and how long each took to answer. A rising grant rate with slowing response times tells you the centre has stopped governing.

Common mistakes that kill results

Mistake 1: messaging by committee

When brand messaging is reviewed by product, legal, sales and the CEO before it goes live, it turns beige. Each stakeholder sands off the edge that made it work. Committees also produce no accountable owner, so the next region to disagree simply reopens the file. Give one person final authority with a defined review window instead of an open revision loop, and put contested claims on a named agenda rather than in a comment thread.

Mistake 2: confusing tone with message

Being funny, bold or minimal is an executional choice. Many CMOs pour energy into voice guidelines while the underlying claim, what the company says it does for the buyer, stays vague. Tone amplifies a message and cannot replace one. If you cannot state the claim as a plain declarative sentence with no adjectives, there is nothing yet for 200 teams to be consistent about.

Mistake 3: resetting the claim because you are bored of it

Memory structures build through repetition, and repetition only compounds when the words stay stable. Refreshing every quarter because a competitor launched something resets that accumulation and writes off the spend behind it. Twelve months is a floor for evaluation, not a review cycle. The mirror failure is worse and rarer to admit: holding a claim the business can no longer deliver, which is what HSBC eventually confronted.

Mistake 4: governance that fights the incentives

A regional MD is paid on regional pipelinepipelineAll active sales opportunities across the stages of the sales process, together with their combined potential value and probability of closing.View full definition →, so when the global claim converts poorly in that market, rewriting it is the rational act. Policing the words while the bonus rewards deviation produces quiet non-compliance you will not see until the annual audit. Either give regions a funded route to test adaptations inside the licence, or accept that your locked tier has to be short enough to survive contact with a quota.