CMO playbook & advanced tactics for integrated media planning

A group CMO with several hundred million dollars of working media does not spend the quarter arguing about channel roles. The arguments that eat the job are narrower and far more expensive: who buys the media, what the buyer earns on it, and whether the numbers coming back were produced by anyone with a stake in the answer. Get the channel mix wrong and you lose a few points of efficiency. Get the governance wrong and a fifth of the budget can disappear without one line of the reporting looking unusual. What follows are the decisions only the CMO signs.

Where the money actually goes

The plan itself, in the sense the foundations lesson sets out, is the straightforward part. The contract layer underneath it is not. Two audits are worth committing to memory. PwC's 2020 study for ISBA in the UK followed advertiser money through the UK programmatic chain and could account for only about half of it arriving at publishers, with roughly 15% untraceable at any point in the chain. The ANA's 2023 programmatic transparency work in the US found campaigns scattered across tens of thousands of domains, with around a fifth of impressionsimpressionsThe total number of times an ad or piece of content is displayed, regardless of clicks. Each display counts as one impression, even to the same person.View full definition → landing on made-for-advertising sites: pages built to arbitrage ad dollars rather than to be read by anyone.

Now put that against a portfolio. On $400 million of working media, a 15% untraceable share is $60 million, larger than the full-year budget of most brands sitting inside your group. That is the number that makes media governance a board conversation rather than a procurement one.

Four contracts decide where that money lands, and all four need the CMO's signature rather than the CFO's:

• The remuneration model for planning and buying
• Whether the agency may trade as principal, and at what disclosed margin
• Who holds the ad tech contracts and, with them, log-level data rights
• Who builds and owns the measurement that grades the spend

Agency, in-house, or split

This is never all or nothing. Strategy, planning, activation, content production and ad ops can each sit in a different place, and the sensible groups split them. Procter & Gamble moved media planning and a large share of programmatic buyingprogrammatic buyingProgrammatic advertising is the automated buying and selling of digital ad inventory through real-time auctions and software, replacing manual negotiation with data-driven decisions.View full definition → in-house in the US while keeping agency partners on a fixed-and-flow model, and reported cutting agency and production costs by roughly $750 million from 2014 onward, with a further several hundred million targeted, having removed thousands of agency relationships from the roster. Unilever went at content instead: its U-Studio in-house studios now operate across more than twenty markets, the agency roster was cut by about half, the number of ads produced fell by roughly 30%, and the company reported around €500 million of savings in brand and marketing spend in 2017.

The arithmetic is less flattering than the case studies. In-housing captures the agency margin on the layer you take over, usually a low single-digit percentage of the spend that passes through it. Against that you carry fixed headcount, DSP and verification licences, and a hiring problem. Below roughly $20 million of annual programmatic spend, the fixed costs rarely beat the margin you reclaimed. It works worst where spend is seasonal, because you own the team in the quiet quarter whether there is anything to buy or not. The failure mode is predictable: an in-house desk built in a growth year, a budget cut, and a team of four now buying at rate card because the pooled negotiating leverage went out of the door with the agency.

Fee arbitration and the principal media question

Commission on spend pays your buyer more for spending more. Flat fees pay them the same for good and bad work. Output-based fees only function if you can define the output, which means agreeing the metric before the contract, not after the campaign. The harder arbitration is principal media buying: the agency buys inventory onto its own book and resells it to you at a margin it does not have to show you. The ANA's 2016 work with K2 Intelligence documented how widespread rebates and non-transparent trading had become in the US market, and the practice has grown since. You have three defensible positions: ban it, permit it with a disclosed margin and a cap on the share of spend it may cover, or permit it only in specified inventory classes. Silence is not one of them, because silence is consent.

Two second-order effects that CMOs miss. First, squeezing fees to the floor buys you a junior team, and that cost surfaces two years later as weaker planning nobody attributes to the fee negotiation. Second, a right-to-audit clause is decorative without log-level data access written into the same contract, because an audit of the invoices tells you what you were charged, never where the impressions went.

Frequency waste and brands bidding against each other

ReachReachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition → and frequency, as the foundations lesson defines them, are brand-level ideas. At group scale they turn into a governance problem. Marc Pritchard has said P&G found individual consumers being served the same brand's ads ten or more times a month, and that cutting the excess paid for incremental reach rather than saving money outright.

Worse is portfolio self-competition. Two of your brands targeting the same demographic in the same auction bid against each other: you pay the clearing price and the increment you created. Fixing it needs a group-level frequency and exclusion layer, shared audience keys or a single trading seat, and someone with the authority to tell a brand P&L owner to stand down. Most groups have the technology and not the authority, which is precisely why the waste survives audit after audit.

Who grades the homework

The cross-channel measurement setup belongs to the application lesson; the governance question is who owns it. Do not let the party selling the media build the model that values the media. LinkedIn's B2B Institute has published genuinely useful work on brand-versus-activation ratios and on the small share of buyers in market at any moment, and LinkedIn sells the inventory that work argues for. Read it, then verify it with your own holdout test. The same applies to an MMMMMMA statistical approach that estimates how each marketing channel and other factors drive sales, guiding budget allocation.View full definition → built by the agency that books the spend, and to platform-reported conversions of any kind.

Case 1: P&G and the cost of demanding proof

At the IAB Annual Leadership Meeting in January 2017, Pritchard described the media supply chain as murky at best and fraudulent at worst, and set conditions: MRC-accredited viewability, third-party verification, TAG certification against fraud, transparent agency contracts. P&G then cut roughly $140 million of digital spend in a single quarter that year, mostly inventory it could not verify as brand-safe, and reported the effect on business growth as immaterial. The lesson for a CMO is not that digital is wasteful. It is that a spend cut of that size was survivable because P&G had the measurement to know what the money had not been doing.

Case 2: Unilever and the limits of consolidation

Unilever's savings came from taking production and roster decisions away from individual brands. The exposure that creates is concentration: fewer agencies means fewer credible alternatives at the next pitch, and pitch leverage is most of what keeps rates honest. Keith Weed's 2018 Cannes intervention, threatening to withhold spend from platforms that would not deal with transparency and content standards, only carried weight because Unilever's spend was large enough to matter to the platforms. Concentration cuts both ways: you gain leverage over suppliers and lose optionality over your own agency.

Case 3: LinkedIn and the walled-garden edge case

LinkedIn clears at high single-digit dollars per click and CPMs well above the open web. That premium is defensible when the alternative is unverified open-web inventory with a fifth of impressions on made-for-advertising pages, and indefensible when the channel is being run as a direct-response engine at scale. The governance catch: your log-level data clause stops at the wall. Inside a walled garden you get aggregate reporting, not the impression-level trail your auditor needs, so the control has to shift to spend caps, holdout geographies and independent measurement rather than supply-chain forensics.

CMO action items

• Commission an independent supply-chain audit on your largest programmatic market, with log-level data access written into the agency contract first. If the contract does not grant it, that renegotiation is the action item.
• Decide your position on principal media in writing: banned, or capped at a stated share of spend with a disclosed margin. Circulate it to every brand team so nobody signs a local exception.
• Split the measurement contract from the buying contract, even if both sit with the same holding company, and fund at least one annual holdout test per major channel.
• Set a group-level frequency cap and cross-brand exclusion rules, then name the person who arbitrates when two brands want the same audience in the same week.

Common mistakes that cost real money

Treating in-housing as a cost programme. The reason to bring planning inside is control of data and speed of decision; the margin you recapture is small and the fixed costs are real. Groups that in-house to hit a savings target usually rebuild an agency relationship within three years, having paid twice.

Auditing the invoice instead of the impression. Invoice reconciliation confirms arithmetic. It cannot find the delta the ISBA and ANA studies both identified, because that gap sits between the buying platform and the publisher, in fees nobody bills you for directly.

Letting fee negotiation run through procurement alone. Procurement optimises the number it is measured on, which is the fee. The expensive variable is the quality of the media bought with 90-plus percent of the budget, and no procurement scorecard captures it.