Why Your Marketing Budget Never Came Back

Here is the figure that should concern every marketing leader, and for which almost no one has a credible explanation. If you were one of the marketing managers out there that has been feeling the perpetual marketing budget pinch year after year then it’s time to dig into what’s really happening and spoiler: it’s not just AI.

In 2020, marketing budgets averaged 11% of company revenue. In 2021, they fell to 6.4%, the lowest in the history of Gartner’s CMO Spend Survey. Then 9.5% in 2022. Then 9.1%. Then 7.7% in 2024, 7.7% again in 2025, and 7.8% in 2026.

Five years. No recovery. Gartner’s own framing is explicit: in the four years before the pandemic, marketing averaged 11% of revenue. In the four years since, 8.2%.

Everyone had an explanation for 2021. Almost no one has an explanation for 2026.

We told ourselves it was a shock and that shocks reverse. Revenue recovered. Profits recovered. Ad markets recovered. The budget percentage did not, and at some point “we are still recovering” ceases to be an explanation and becomes a refusal to investigate.

The explanation is not inside marketing. It is in the bond market, and it indicates the capital is not returning.

The Narratives That Do Not Hold

Three stories circulate. All three fail in the same way: they treat this as a marketing problem. What occurred was not a marketing problem.

“Marketing lost credibility with the board.” Partially true, but causally inverted. Board scrutiny of marketing increased after the budget declined, not before. In Duke’s CMO Survey, the share of firms reporting board-level scrutiny of marketing jumped from 33% to 50% in a single year. That is what happens to a line item already under pressure. It is not what created the pressure.

“AI made marketing cheaper, so budgets fell.” The chronology does not support it. Budgets collapsed in 2021, well before generative AI affected any material production workflow. And Gartner’s own data contradicts the efficiency narrative: labor rose from 22% to 25% of marketing budget between 2025 and 2026. AI is being funded by reallocating dollars inside the budget, not by reducing the budget.

“It’s just the ad recession.” Ad spend recovered. Marketing as a percentage of revenue did not. Those are distinct lines on distinct charts.

Each narrative treats a five-year structural reset as a marketing effectiveness issue. It is a capital allocation issue, and it affected every long-duration investment in the enterprise. Marketing simply happens to be the one with no structural defenses.

What Actually Changed

Start with the component most marketing leaders never examine, because it resides in a part of the business we have collectively decided is not our responsibility.

In February 2026, the Federal Reserve published a research note examining why far-forward Treasury rates had risen so dramatically. The finding: nine-to-ten-year forward rates rose approximately 200 basis points over five years, the largest move of this kind since the early 1980s. The driver was not inflation expectations, which remained anchored near 2%. It was a rising real term premium, driven by perceived supply-shock risk and fiscal sustainability concerns, with the Fed’s own economists citing CBO projections toward 120% debt-to-GDP.

That distinction is load-bearing for this entire argument. A term premium is compensation for bearing fiscal risk. It is not a monetary policy setting. The Federal Reserve can cut the short end of the curve and the long end may not follow, because the long end is no longer pricing the Fed. It is pricing the borrower.

The cost of long-dated capital therefore did not rise cyclically. It repriced structurally. And structural repricing does not mean-revert while you wait.

Why A Repriced Discount Rate Lands On Marketing First

Every internal investment decision is measured against a hurdle rate. Marketing rarely participates in setting it, which is a strategic error, because it is the number that determines your budget before anyone evaluates your plan.

Niels Gormsen and Kilian Huber constructed the most comprehensive dataset that exists on this. They hand-coded approximately 74,000 paragraphs from corporate earnings calls between January 2002 and September 2021, extracting what firms disclosed about their discount rates and perceived cost of capital. Approximately 2,500 large firms across 20 countries.

Two findings are enormously consequential and almost entirely unknown inside marketing.

First, the average firm’s discount rate was 15.7%. The average perceived cost of capital was 8.4%. A standing wedge of more than seven points, in place long before this rate cycle began. Companies were already applying an internal bar roughly double their actual cost of capital.

Second, the wedge is sticky asymmetrically. Only one-third of firms had a different discount rate after three to four years. Approximately 70% had updated their perceived cost of capital within one to two years. Organizations notice that capital has become more expensive quickly and pass it into the hurdle rate slowly, and once elevated, it remains elevated.

Gormsen and Huber are explicit about the consequence: these wedges “affect firm investment,” and increasing wedges “can account for the recent puzzle of ‘missing investment.'”

Missing investment. That is a paper about the bond market describing your marketing budget.

The transmission mechanism is broken in the direction that harms you. Rising capital costs reach your hurdle rate quickly. Falling capital costs reach it slowly, if at all. The bar your CFO will apply to your plan next year is not going to decline when rates ease, and given that the long end is being driven by fiscal risk rather than monetary policy, rate easing is not the base case in any event.

Marketing Is The Softest Target In The Enterprise

A higher hurdle rate compresses every long-payback investment. It compresses marketing most severely, for a reason that has nothing to do with marketing effectiveness.

Marketing is expensed, not capitalized. Build a factory and it becomes an asset with a depreciation schedule, and cutting it requires an impairment discussion with real accounting consequences. Build a brand and it hits the P&L the quarter you spend it. There is no asset on the balance sheet, no schedule to defend, no write-down to explain.

This is not theoretical. Joshua Pierce documented it in the Review of Financial Studies: multinationals’ U.S. advertising spend tracks their contemporaneous foreign cash flow. Ad budgets are rationed through the internal capital market like investment, not protected like fixed operating expense. Robert Hall’s NBER work found the same dynamic from the opposite side, showing ad spend falls disproportionately in downturns because it is the most deferrable item on the list.

Combined: Marketing behaves like capital investment when capital is being allocated, and like discretionary expense when capital is being cut. It receives the worst half of both treatments.

Then add the discount rate. A higher rate mechanically devalues distant cash flows. Brand building, category creation, multi-year enterprise land-and-expand: all are long-duration assets being valued at a higher discount rate by a bar that ratchets up and refuses to ratchet down.

That is the entire mechanism. Marketing budgets did not decline because marketing became less effective. They declined because the discount rate rose, and marketing is the largest long-duration investment in the enterprise with nothing on the balance sheet to protect it.

The Capital Markets Already Selected Their Metric

While this was occurring, investors settled the question of what type of growth they would finance.

SaaS and tech revenue multiples compressed from a peak around 17x in 2021 into a 5-7x band by 2023, and have compressed further since. The screening criterion shifted from growth at any cost to Rule of 40 and burn multiple. Boards inherited that screen and applied it downward into the operating plan.

You can see it in the benchmarks. Median CAC payback for B2B SaaS compressed to 16 months in 2025, from 18 the year prior, with top-quartile companies at six months or less. Duke’s data shows CFO pressure on marketing rising from 52% to 63% of firms and CEO scrutiny from 51% to 61%, alongside that board-level jump from 33% to 50%.

None of that is a judgment on marketing effectiveness. It is a cost of capital being enforced through whatever metric is closest to hand.

And Then We Made It Worse

This is the portion I find most difficult to write, because it is the portion where the profession is complicit.

The best evidence on marketing effectiveness, Binet and Field’s work across roughly 1,000 IPA Databank case studies, indicates the profit-maximizing allocation is approximately 60% brand building and 40% activation. Their explicit crisis guidance is that brand should be the last thing cut, with a floor near 40% even under severe pressure, because the damage from going dark surfaces three to six months later, well after the quarter in which you saved the money.

WARC surveyed over 1,000 marketers about their 2026 budgets. Among those facing cuts, 42% said they would protect performance marketing. 29% said brand.

The exact inversion of the evidence.

And we are aware of it. In the same research, 55% of marketers named short-termism as a serious problem, up from 25% in 2022. We can articulate the trap, describe the trap, and walk into it anyway, because performance marketing produces a number inside the payback window the hurdle rate is enforcing, and brand does not.

The discount rate is selecting against the very activity that builds pricing power. We are assisting it.

What To Do About It: Five Moves

Five actions, roughly ordered by how quickly you can execute them.

1. Discover your company’s actual hurdle rate, and when it was last set. Ask the CFO directly. On the Gormsen-Huber evidence, there is a high probability it is several years stale and sits materially above your true cost of capital. You are being evaluated against a number no one has revisited, and revisiting it is a conversation Finance is equipped to have and Marketing has never initiated. This is the highest-leverage hour you will spend this quarter.

2. Stop requesting budget in budget language. Capital inside your enterprise is being allocated in payback months, IRR, Rule of 40, and burn multiple. A brand proposal argued on reach and awareness competes against a capex proposal argued in cash flow, and it loses on vocabulary before anyone assesses the merits. Translate your asks. Assign brand investment a stated payback horizon and defend the horizon rather than pretending it lacks one.

3. Put the 60/40 versus 42/29 gap in writing, to your CFO and your Board. Not as a request for more capital. As a documented decision: the company is electing to optimize for the metric Finance is monitoring this quarter, at a known and evidenced cost to pricing power in three years. Require a signature. Decisions made by drift are the decisions no one owns when they mature, and this one matures precisely when the current leadership team is being evaluated on pricing power.

4. Establish a brand floor and hold it as a floor. Binet and Field place the crisis minimum near 40%. Roughly half of tested brands show measurable decline within a year of going dark. That is an empirical argument, not a preference, and it is the only type of argument that survives a hurdle rate discussion.

5. Stop planning for the cycle to turn.

That last point is the critical insight, so let me state it directly.

If the long end of the curve is being driven by a fiscal term premium rather than monetary policy, and if internal hurdle rates require three to four years to reflect rate changes at all and a decade to reflect them fully, then 7.8% of revenue is not a trough you are waiting out. It is the operating environment.

Which means the strategic question is not “how do we get back to 11%.” It is “what is the best marketing function that can be operated at 7.8%, permanently, by design.”

Those questions produce entirely different organizations. The first is a shrunken version of what you had, carrying the same structure with less capital in it, quietly starving every function equally while waiting for relief. The second is architected for the constraint: fewer, larger bets, more owned media, more compounding assets, less rented reach, and a rigorous honesty about which activities were only ever viable because capital was free.

Most marketing organizations right now are operating the first model and labeling it prudence.

The bond market has been signaling for five years that it should be the second.

---

Sources: Gartner CMO Spend Survey (2020-2026); Duke CMO Survey on board/CFO/CEO scrutiny; Federal Reserve research note on far-forward Treasury rates (Feb 2026), real term premium, CBO debt-to-GDP path; Gormsen & Huber earnings call dataset (74,000 paragraphs, 2,500 firms, discount rate 15.7% vs cost of capital 8.4%); Joshua Pierce, Review of Financial Studies on internal capital markets and ad spend; Robert Hall NBER work on deferrable ad spend; SaaS revenue multiples 2021-2025, Rule of 40 / burn multiple shift; B2B SaaS CAC payback benchmarks (2025 median 16 months); Binet & Field IPA Databank (60/40 allocation, 40% crisis floor); WARC 2026 budget survey (42% protecting performance, 29% brand, 55% short-termism).