Because nothing in your marketing is allowed to live long enough to compound. The median American worker has been with their employer 3.9 years, the lowest since 2002 according to the Bureau of Labor Statistics in January 2024. Spencer Stuart’s January 2026 study puts S&P 500 CMO tenure at 4.1 years, the shortest core C-suite role. The ANA and 4As found the average media agency relationship lasts 3.7 years. Every vessel that could hold what a company has learned about its buyers is emptied and refilled roughly every four years, and the Spring 2026 CMO Survey found more than 70% of marketing leaders prioritising immediate results over long-term gains. Compounding requires a base that persists. Most companies have arranged not to have one.
What compounding would look like, and why you rarely see it
Compounding in marketing has a precise description. Binet and Field’s analysis of 996 campaigns in the IPA Databank, spanning 1980 to 2010, describes it precisely: activation produces a sharp lift that decays within months, while brand building produces a slow lift that accumulates, and the total business effects of a campaign rise steadily the longer it runs. The Ehrenberg-Bass Institute’s 2018 hiatus study shows the other side of the same coin. When brands stopped advertising, big brands with years of accumulated memory coasted for a year or two; small brands with none fell 16% in the first year. Memory is the asset. It is built slowly, spent slowly, and only exists if something is left running long enough to build it.
Now look at what a mid-market company actually keeps running. Spencer Stuart tracked 218 CMO exits between 2021 and 2025 and found the average S&P 500 marketing chief lasts 4.1 years, against 5.0 for the C-suite overall, with consumer companies at 3.5. Below the index, where 31% of S&P 500 firms already have no CMO at all, the top marketing job is frequently a director who lasts less. The agency that holds the creative memory turns over every 3.7 years if it is a media shop, according to the ANA and 4As in April 2025, and every 5.8 if it belongs to a holding company. The coordinator who knows where the brand files are follows the BLS median out the door in under four years. Each departure takes with it the one thing that cannot be written down, which is why the last thing worked.
The CMO Survey adds the final piece. Its Spring 2026 edition found the median duration of a marketing investment’s impact on customers has lengthened to six months, with a meaningful shift toward a year or longer. Marketing is compounding more than it used to. But over 70% of the leaders running it are optimising for the quarter, training budgets have fallen to 3.8% of spend, and headcount growth halved in a year. The asset is appreciating and the owners keep selling it at the bottom.
Why the reset happens every time
Every new marketing leader, agency and coordinator arrives with the same rational incentive: to be seen to have changed something. A new CMO who keeps the predecessor’s positioning, agency and visual identity has, on paper, done nothing in year one, and Spencer Stuart’s data says year one is a quarter of the tenure they will get. So the logo is refreshed, the tagline is retired, the agency is reviewed at an average cost of $408,500 to the client, and the market, which had just begun to recognise the old version, is asked to start again. None of the people involved are acting badly. The system rewards the reset and penalises the steward, and it does so on a four-year clock.
The measurement culture reinforces it. The CMO Survey found more than 70% of leaders prioritising immediate results, and Moorman noted that most respond to pressure for proof by building stronger performance tracking rather than deeper customer insight. Performance tracking measures the sharp, short-lived effects almost perfectly and the slow, accumulating ones almost not at all. A company that manages to its dashboard will, with complete sincerity, cut the only activity that was compounding, because it was the one line that never looked urgent. What gets measured gets cited, and what gets measured quarterly gets cut annually.
There is a fair objection here. Markets change, and a company that never resets can calcify around a positioning its buyers stopped believing a decade ago. Some resets are correct. The distinction is between changing the expression and changing the memory. Apple has changed its advertising continuously and its point of view almost never; that is why the CMO Survey’s own respondents have named it the standard-setter for seventeen straight years. The reset that destroys compounding is the one that throws out the point of view along with the campaign, usually because nobody left in the building can articulate which was which.
The asset is appreciating and the owners keep selling it at the bottom, every four years, to whoever just arrived.
What you can actually do about it
A CEO cannot lengthen the BLS median or repeal the four-year clock. What a CEO can do is move the memory out of the people who leave and into something that stays. The Signal has written about what a marketing operating system is; this is the reason a company needs one.
Write down the point of view, and make it the CEO’s
One page: who you are for, what you believe about their problem, and how you say it. Owned by the CEO, not the marketing lead, so it survives the next 4.1 years. New leaders may change the expression. They may not change the page without a board-level conversation.
Keep a ledger of what worked and why
Every campaign, channel test and message gets a one-paragraph entry: what was tried, what happened, what was learned. Ten minutes a month. In four years it is the only institutional memory the next person will inherit, and it turns their year one from a reset into a continuation.
Measure one slow number alongside the fast ones
Unprompted brand recall in your category, branded search volume, or share of buyers who name you first. Reviewed quarterly, judged annually. It is the line on the dashboard that represents the compounding, and its job is to make the cut visible before it happens.
Reward the steward as loudly as the reformer
When a new leader or agency proposes to keep what is working, say so in public and in their review. The four-year reset is a response to incentives. Change the incentive and the clock loses some of its power, which is the most a CEO can do about a labour market.
The part nobody wants to hear
Most companies will not do any of this, and indifference has little to do with it. Writing down a point of view forces a CEO to have one, in words specific enough that a new marketing leader could be told no. Keeping a ledger means admitting how much was tried and forgotten. Measuring a slow number means watching it not move for a year while the fast numbers demand attention. Each of these is a standing commitment with no deadline and no natural owner, and in a mid-market company that describes the ten free hours a week that belong to no one.
So the four-year clock keeps running, and each new arrival is handed a folder of logos and a login to the analytics, and the buyer, who has been there the whole time, is introduced to the company again. Aware, adaptive, alive was never a slogan about speed. It was about a company that remembers what it learned, changes what should change, and keeps what should be kept, on purpose, across the departures of the people who learned it. Machines can hold the files. Only an operating system, with a human at the centre who is not leaving in 4.1 years, can hold the meaning. The companies whose marketing compounds are the ones that decided the memory belongs to the company and not to whoever happens to be in the chair.
Marketing fails to compound because every vessel that holds its memory is replaced within four years; the only fix is to move the memory into something that stays, and the CEO is the only person who can order it.