Every senior marketer knows the moment. The CFO sets down the deck, looks across the table, and asks the question that decides the next 12 months.
“What, exactly, is marketing contributing here?”
It’s rarely hostile. It’s almost always sincere, in fact. Most importantly, it exposes a problem no dashboard can fix. Classify marketing as a cost, and the budget conversation is lost before it begins. Cost lines are benchmarked against other cost lines, squeezed when revenue tightens, and never expanded on the back of a good brand story or a solid pipeline quarter. The structural problem is not that marketers report poorly or attribute clumsily — it’s that the category itself is wrong.
The fix isn’t better reporting, it’s a different conversation, built on a different premise about what marketing is. In his OMR keynote, Tom Inbal, SVP of strategy and corporate marketing at Taboola, put it plainly. He refuses to treat marketing as a cost item and instead runs it like an investment function: capital deployed, returns measured. That single reframe changes the questions a CFO asks, the metrics that matter, and what happens to the surplus when a quarter goes well.
This article walks through how to make that shift.
Why the Cost Center Label Is Costing You More Than Budget
A cost is, by definition, something a CFO is paid to minimize. That’s the job. If marketing sits on the expense side of the ledger, every conversation about the function will be shaped by the gravitational pull of cost reduction. Efficiency gains are rarely celebrated as opportunities to reinvest. More often, they’re used as evidence that the same outcome should now require less spend.
The pattern shows up in predictable ways across every budget cycle:
- A campaign that hits its targets at 80% of planned spend is read as a 20% savings, not as 20% of capital available for redeployment.
- A team that automates its reporting and execution gets a smaller line item next year, not a bigger mandate.
- An efficiency gain becomes the baseline for next year’s ask, not a reason to expand the team’s scope.
The result is a self-undermining trap. The better marketing gets at running efficiently, the stronger the case becomes for cutting the budget that funded the efficiency in the first place.
None of this is a Finance problem. Finance is doing exactly what it’s supposed to do. The problem is upstream, in the framing. As long as marketing is evaluated as a cost, efficiency conversations will likely end in budget cuts. The only way out is to move the conversation from cost reduction to capital deployment. From, “How much are we spending?” to, “What is the yield, and where should we deploy more?”
That’s more than a change in language. It’s a reclassification of what marketing is inside the business.
The AI Efficiency Trap
The cost center framing was always vulnerable. AI makes it acutely so.
In a recent Taboola survey of several hundred marketers, 98% reported actively using or testing agentic AI solutions, and 76% said they were seeing meaningful performance improvements from those tools. Those numbers are real, and they’re good news for marketing teams. Under a cost center model, they are also an existential problem.
Here’s the chain of reasoning a CFO will follow, entirely rationally, if marketing is still positioned as overhead. AI is handling more of the briefing, more of the media buying, more of the creative variation, more of the measurement. The work product is similar or better. As Inbal said in his keynote, “If you’re not breaking free of the average, your CFO is thinking, ‘AI is doing 90% of this team’s job. Maybe they’re replaceable. Maybe they can be downsized. Maybe a few agents can do this work.’” In other words, the same output should be achievable with fewer people and a smaller budget. That conclusion isn’t an act of hostility, it’s simply the only conclusion available when marketing sits on the expense side of the ledger.
There’s a different way to read those numbers, though. If marketing is run like an investment portfolio — with an Exploit bucket of proven, reliable activity and an Explore bucket of deliberate bets on new opportunities — then AI-driven efficiency in the Exploit bucket is not a saving. It’s capital freed up to fund Explore, which is where the outsized returns come from. The threat to marketing is not AI itself. It’s the framing that lets Finance read efficiency as a reason to cut the budget instead of redirecting it.
Inbal made a related observation in his keynote. AI has compressed the variance in performance across the industry. The distribution of cost per action on the Taboola network narrowed by roughly 50% between 2023 and 2025. The average has gotten better, which means being average has gotten cheaper and easier, and being above average has gotten harder. If the response to that compression is to keep optimizing the proven layer, the team becomes more efficient and less differentiated at the same time. That’s the exact profile a CFO can replace with a few agents.
None of this is an argument against AI. Inbal is clear on this point in his keynote. “Anybody not deep into AI should be looking into their resume pretty quickly,” he said.
The marketers who win in this environment are the ones racing to adopt AI, not bracing against it. The question isn’t whether to use the tools, it’s whether you’re using them to optimize a cost line or to fund the bets that generate alpha.
What It Actually Means to Run Marketing Like an Investment Portfolio
As Inbal emphasized in his keynote, being courageous without a system is recklessness. Being courageous with a system is how you outperform. That’s the purpose of a portfolio: making bravery operational rather than accidental.
A portfolio is a very specific concept, and it’s worth being clear about what it is and what it isn’t.
- A portfolio is not a mix of activities that all carry an expectation of positive return. That’s a diversified spend plan.
- A portfolio is a structured allocation of capital across assets with different risk and return profiles, where some allocations are explicitly expected to lose money.
A portfolio without losses isn’t a cautious portfolio. It isn’t a portfolio at all.
This is the conceptual leap most marketing organizations haven’t yet made. Many marketing budgets are still built as if every line is expected to perform. Every campaign has a target. Every test is expected to validate. Every quarter is supposed to be green. That construction is what makes the marketing function look responsible from the outside, while also preventing it from generating the disproportionate returns that justify expansion.
Running marketing like a portfolio means budgeting from the start with two distinct buckets and two distinct sets of expectations.
The Two Buckets: Exploit and Explore
A portfolio works because its two buckets are doing fundamentally different jobs. The Exploit bucket is the proven layer. These are the channels, creatives, audiences, and tactics where the team has historical data, predictable return profiles, and confidence in the unit economics. If the team can say, “We know what this returns,” it lives here. This is where most of the budget sits, and where AI-driven automation delivers the most value. Efficiency is where the Exploit bucket shines. Let AI run the things it can demonstrably run better than a human operator. The role of the Exploit layer is to produce stable, predictable yield, not breakthrough performance.
The Explore layer, on the other hand, is where you put money behind things you haven’t proven yet. New channels, new formats, new audiences, and new creative territories. A few years ago, connected TV sat in Explore for most advertisers. Today, large language model platforms sit there. The point of Explore is not innovation theater layered over a steady Exploit operation. The point of Explore is to get there first. The bets only pay off if the rest of the market hasn’t already crowded in, because once everyone is doing it, the edge disappears.
Some of these bets will fail. That’s the model working as designed. The ones that succeed generate the alpha — the return above the market average — that the Exploit layer can never deliver on its own.
Why a Real Portfolio Expects Some Losses
Most marketing leaders can handle the math of a portfolio model. What trips them up is the discomfort of planning for failures they know are coming.
This discomfort isn’t a character flaw. Performance reviews, quarterly check-ins, board updates, and CFO conversations are all built around the assumption that every line item should be working. A failed test is a thing to be explained, not a thing to be expected. Under that operating logic, Explore is impossible. The first time a bet doesn’t pay off, the budget is at risk.
“You expect only some of it to work out,” Inbal said in his keynote. “The part that pays off, that’s where your alpha is going to come from. That’s where you’re going to outperform, not by optimizing your Exploit.”
The CFO conversation has to include this expectation explicitly, in advance. The loss rate inside Explore is not a defect of the strategy, it’s the price of access to outsized returns. If that isn’t settled at the front end of the budget cycle, every Explore failure becomes a threat to the entire allocation.
Changing the CFO Conversation: From Spend to Yield
The budget conversation is where this reframe gets real. When marketing is treated as a cost, the conversation centers on, “How much are we spending on marketing?” When it’s treated as a portfolio, it centers on, “What’s the yield on our marketing capital, and how is it allocated?”
That shift requires a different reporting structure, a different set of metrics, and a different relationship between the marketing team and Finance. It’s more demanding for marketing leaders, not less. It’s also the only version of the conversation that ends with marketing being treated as a value driver.
Replacing the Budget Request With a Capital Deployment Plan
A budget request is a cost justification. A capital deployment plan is an investment thesis. The two documents look similar on the surface, but do very different work.
A capital deployment plan spells out three things:
- How much capital goes into Exploit, with the expected return range and the confidence level behind that range.
- How much goes into Explore, with the expected loss rate, the criteria for what counts as a viable bet, and the upside if those bets land.
- What the reinvestment mechanism looks like when the Exploit layer overachieves — and that mechanism needs to exist before the overachievement happens, not after.
That puts the marketing leader in roughly the same conversation a fund manager has with a board. Capital’s being deployed against a stated risk-and-return profile. Some allocations are expected to outperform, some to underperform, and the portfolio as a whole is evaluated against a benchmark. That’s a fundamentally different posture than defending a cost line.
Reinvesting Surplus: The Mechanism That Changes the Dynamic
The most concrete piece of the model is what happens when things go well, as Inbal described in his keynote. “Last year, my teams overachieved on their goals,” he said. “We didn’t just kick back that extra surplus to the CFO. We came in with a business plan, and we funded additional exploration.”
That single behavior is the test of whether the portfolio framing has actually taken hold. If marketing is still classified as a cost, surplus becomes a cut. The logic is often unavoidable: if the team hits its goals with less, it can hit them with less next year, too. If marketing is classified as a capital deployment function, surplus becomes fuel for the next round of bets. The same dollars, framed differently, produce opposite outcomes.
This mechanism is only available to marketing leaders who’ve already done the upstream work with Finance. You can’t retrofit it after a good quarter: it has to be the default assumption built into the budget cycle from the start.
How Realize Enables the Portfolio Approach in Practice
The portfolio framework isn’t just a conceptual reframe. It needs to be auditable and reportable, because CFO buy-in doesn’t come from a good argument, it comes from your ability to show that the allocation is intentional, measurable, and tracked over time.
That’s where Realize earns its place in the model. Realize+ runs as the autopilot for the Exploit layer, handling proven campaigns automatically and protecting the efficiency that generates the stable returns funding Explore. You aren’t babysitting the proven layer. The machine runs it, and you get the capacity back for the bets that actually move the needle.
The Realize dashboard and autonomous budgeting tools handle the structural side. Teams can split spend cleanly between Exploit and Explore from the start of the planning cycle, track the performance of each bucket separately, and show Finance that the allocation is deliberate. The Exploit layer gets reported against efficiency and stability metrics like cost per action (CPA), return on ad spend (ROAS), and contribution margin. The Explore layer gets reported against learning velocity, bets initiated, signal strength, and the rate at which Explore bets graduate into Exploit programs.
The point isn’t that Realize makes the portfolio strategy possible in theory. It makes it auditable and reportable, which is what CFO buy-in actually requires.
Key Takeaways
The CFO sets down the deck again. The same question, the same table, the same decision about the next 12 months.
“What, exactly, is marketing contributing here?”
This time, though, the marketing leader doesn’t defend a cost line. Instead, they present a capital allocation review. The Exploit layer is performing. The Explore pipeline is healthy. Losses are inside expected parameters. Two programs are signaling strongly enough to graduate into Exploit next quarter, and the surplus from Exploit overperformance is going into a third Explore allocation.
That conversation doesn’t get cut. It gets expanded.
The shift from cost center to investment portfolio is partly a financial reframe, but it’s also a career strategy. The marketers who become irreplaceable in the AI era are the ones who take ownership of outcomes, not just activities — who are willing to be measured on yield, comfortable with planned failure, and able to defend their portfolio in the language Finance already speaks.
Frequently Asked Questions (FAQs)
How do I decide what percentage of budget should go to Exploit vs. Explore?
There’s no universal ratio. The right split depends on the maturity of the business, the competitive intensity of the category, and how much risk the organization is willing to take. The starting principle is that the Explore allocation needs to be big enough to generate meaningful signal across multiple bets, but small enough that the expected loss rate doesn’t breach a threshold Finance won’t tolerate. The Exploit layer also has to be robust enough that the team can take risks elsewhere without the whole budget feeling threatened. For teams new to the portfolio, an 80/20 split between Exploit and Explore is a defensible starting point, adjusted as the team learns its actual hit rate.
How do we report on Explore investments without the CFO treating every failure as a red flag?
This is a conversation that has to happen before the budget is allocated, not after a loss. Sit down with the CFO upfront and agree on three things: how often Explore bets are expected to miss; how much any single bet can lose before it gets pulled; and what actually counts as a win, a wash, or a failure. Once those numbers are on the table, losses that fall inside them aren’t bad news, they’re proof that the model is doing what it’s supposed to do. Try having that conversation after a bet fails, and the framing will land very differently.
Should CFOs reclassify marketing spend as a capital investment on the balance sheet?
Formal reclassification is a finance and accounting decision that varies by jurisdiction, business model, and audit treatment, and it isn’t something that a marketing leader can drive unilaterally. The more immediately actionable point is that the mental model can shift even when the accounting treatment doesn’t. Treating marketing spend as capital deployment changes the questions Finance asks, the metrics the team tracks, and the default assumption about what happens to surplus. Marketing leaders who put in the upfront work with Finance can shift the conversation immediately. The balance sheet can catch up later, or not at all.
