Rolling forecasts vs. zero-based budgeting: when to use each

In February 2017, Kraft Heinz CFO Paulo Basilio sat on an earnings call defending margins that had reached 23.8%, the envy of the consumer packaged goods industry. By August 2019, the company had taken a $15.4 billion writedown on the Kraft and Oscar Mayer brands, the CEO had been replaced, and the SEC was investigating its procurement accounting. The instrument blamed by analysts, board members, and eventually 3G Capital's own partners? Zero-based budgetingZero-based budgetingA budgeting method where every expense is justified from zero each cycle, rather than carrying last year's spend forward.View full definition → taken to its logical extreme.

Meanwhile, 4,000 miles east in London, Unilever CFO Graeme Pitkethly was quietly running quarterly rolling forecasts that allowed the company to reroute €500 million in trade spend within weeks of the 2022 European energy shock, without convening a single emergency budget meeting.

Two finance philosophies. Two outcomes. The question every CFO faces in 2026, with input cost volatility, OECD Pillar TwoOECD Pillar TwoOECD-backed rules imposing a 15% minimum effective tax rate on large multinational groups, jurisdiction by jurisdiction.View full definition → creating new effective tax floors, and CSRDCSRDEU directive requiring large companies to report standardized, audited sustainability data alongside financial results.View full definition → reporting consuming an estimated 15-20% of FP&AFP&AThe finance function that builds budgets, forecasts and analysis to guide business decisions and connect strategy to numbers.View full definition → bandwidth, is no longer "which is better?" It's "which one, where, and for how long?"

The theoretical divide (and why it matters more in 2026)

Zero-based budgeting: the clean-sheet discipline

Zero-based budgeting (ZBB), originally developed by Peter Pyhrr at Texas Instruments in 1970 and popularized at Jimmy Carter's Georgia governorship, requires every cost center to justify every expense from a base of zero each cycle. No "last year plus 3%." No baseline inertia. Each dollar competes against every other dollar.

The modern 3G Capital version, deployed at AB InBev after the InBev-Anheuser-Busch merger in 2008 and Kraft Heinz after the 2015 Kraft-Heinz merger, added two industrial-grade enhancements:

1. Cost-package ownership: Every expense (travel, IT, marketing, R&D) is owned by a single executive globally, not the local cost center.
2. Zero-based targeting (ZBT): Targets are set bottom-up using benchmarks (printing cost per page, T&E per FTE), not last year's run-rate.

When AB InBev applied this to Anheuser-Busch, SG&A as a percentage of revenue dropped from 35% to 26% within 18 months. Operating margin expanded by roughly 800 basis points. The playbook seemed unimpeachable.

Rolling forecasts: the adaptive lens

Rolling forecasts replace the static annual budget with a continuously updated 12-, 18-, or 24-month forward view, refreshed monthly or quarterly. The forecast horizon "rolls", when Q1 closes, you add a new Q1 to the back end, so you always see four to eight quarters ahead.

The discipline isn't new, it dates to Beyond Budgeting Roundtable work in the late 1990s, but the 2020s tooling (Anaplan, Pigment, Workday Adaptive, OneStream) and the post-COVID acceptance of plan volatility have moved it from niche to mainstream. PwC's 2025 Finance Effectiveness Benchmark found 64% of large-cap companies now run a rolling forecast in some form, up from 42% in 2019.

The philosophical difference: ZBB asks "what should we spend?" Rolling forecasts ask "what will happen, and how do we respond?"

Case study one: the kraft heinz cautionary tale

The 3G ZBB model worked brilliantly on cost, and catastrophically on growth. Between 2015 and 2019, Kraft Heinz cut roughly $1.7 billion in annual costs. EBITDAEBITDAEBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) measures a company's operating profitability before financing and accounting decisions, used to compare core performance across firms.View full definition → margins reached industry-leading levels. But marketing spend was cut from 4.7% of revenue to under 3.5%. R&D investment in new product categories essentially flatlined. Brand health metrics for Oscar Mayer, Kraft, and Maxwell House deteriorated quarter after quarter while the finance function celebrated cost-package "wins."

By 2019, organic revenue growth was negative 1.5%. The brands the company had bought weren't worth what they had paid because the ZBB discipline had stripped out the reinvestment that maintains brand equity. CEO Bernardo Hees was replaced by Miguel Patricio, who immediately announced a "consumer-first" reinvestment strategy, code for "we starved the brands."

The lesson isn't that ZBB is bad. It's that ZBB applied to growth-investment line items is corrosive. When you force a brand manager to justify a $40 million Super Bowl ad from a base of zero against a procurement manager justifying $40 million in packaging spend, the procurement manager wins every time. The savings are measurable next quarter; the brand erosion shows up four years later.

Where ZBB actually creates durable value

Looking across two decades of ZBB deployments, Mondelez under Dirk Van de Put, Reckitt under Laxman Narasimhan's tenure, Unilever's 2017 "Connected 4 Growth" program, the durable wins concentrate in four buckets:

• Indirect procurement (T&E, professional services, software licenses, facilities)
• Overhead functions post-merger, where redundancy is structural
• Print, telecom, and legacy IT, cost categories with clear benchmarks
• Shared services and back-office consolidation

Where it destroys value: customer-facing marketing, R&D pipelines, sales coverage in growth markets, and talent investment in scarce skills (data science, AI engineering, sustainability reporting under CSRD).

Case study two: unilever's rolling forecast engine

Unilever's transition to a rolling forecast model began in 2019 under then-CFO Graeme Pitkethly and was fully embedded by 2022. The architecture is worth understanding because it's now the reference model for several FTSE 100 finance functions:

• Horizon: 6 quarters forward, refreshed monthly for the next 2 quarters and quarterly beyond
• Granularity: Driver-based (volume, mix, price, FX) rather than line-item
• Cadence: A monthly "flash" (3 days post-close), a quarterly deep-dive
• Decision rights: Brand category presidents can reallocate up to 15% of trade and marketing spend without group approval if the forecast deviation exceeds ±5%

When Russian energy prices spiked in March 2022, the Unilever European business reforecast within 11 working days. €500 million of trade promotion spend was redirected from low-elasticity premium SKUs to value-tier products that consumers were trading down to. The full-year European margin came in only 90 basis points below the original plan, versus an industry average decline of 240 basis points.

Compare this to a static-budget competitor: by the time the annual reforecast cycle began in September, six months of share had been lost.

The hidden cost of rolling forecasts

Rolling forecasts are not free. The same PwC benchmark found that companies running monthly rolling forecasts spend, on average, 38% more FP&A hours per year than peers running annual budgets with quarterly reforecasts. If the forecast doesn't drive a decision, reallocation, hiring freeze, capexcapexCapital Expenditure (CapEx) is money spent to acquire, upgrade, or extend long-lived assets like equipment, property, or software that deliver value over multiple years.View full definition → deferral, you're paying for an expensive weather report.

The discipline that separates Unilever-quality rolling forecasts from cargo-cult versions: every forecast cycle ends with at least one explicit reallocation decision. No decision, no value.

The decision framework: when to use each

The binary "ZBB vs. rolling forecast" framing is wrong. World-class finance functions use both, on different parts of the P&L, at different times in the corporate lifecycle. Here's the framework I recommend:

Use zero-based budgeting when:

1. Post-merger or post-spinoff integration, where redundant cost structures are structural rather than performance-driven. Kraft Heinz's first 18 months were legitimately ZBB-appropriate; the problem was running it for five years.
2. Cost categories with external benchmarks and low strategic differentiation, indirect procurement, facilities, telecom, low-end IT. AB InBev's procurement organization remains the gold standard here.
3. Activist pressure or covenant stress, when you need a credible, defensible cost program that withstands board scrutiny. Elliott Management's playbook frequently demands ZBB-style reviews.
4. Functions that have been "growing with the business" without challenge, typically HR, finance, legal, and corporate communications after a long bull run.

Use rolling forecasts when:

1. Input cost or demand volatility is high, commodity exposure, FX-sensitive revenue, discretionary consumer categories. The post-COVID environment, ongoing tariff volatility, and 2026 energy price uncertainty make this nearly universal.
2. The business model has variable monetization, subscription churn, marketplace take-rates, platform GMV. Spotify, Shopify, and Atlassian all run weekly rolling forecasts on subscription metrics.
3. You manage a portfolio with active capital reallocation, private equity portfolio companies, conglomerates, multi-divisional industrials. Siemens AG runs 18-month rolling forecasts feeding directly into its capital allocation committee.
4. Regulatory or policy environments are shifting, banks adapting to Basel IV finalization, multinationals modeling Pillar Two top-up taxes, CSRD-affected companies reforecasting sustainability capex.

The hybrid most cfos should run

In practice, the answer for 80% of large companies in 2026 looks like this:

• Annual ZBB sweep on indirect cost categories (typically 12-18% of the cost base), rotating through categories on a 3-year cycle so each gets a clean-sheet review every third year
• Quarterly rolling forecast on revenue, gross margingross marginGross margin is the share of revenue left after subtracting the direct cost of producing goods or services, expressed as a percentage of revenue.View full definition → drivers, and growth investment
• Monthly flash on the top 5-7 KPIs the CEO actually steers by
• Strategic plan (3-5 year) refreshed annually, with explicit scenario branches

This is essentially the model Microsoft CFO Amy Hood runs, the model Diageo deployed under Lavanya Chandrashekar starting in 2023, and the emerging standard at most S&P 500 industrials.

Monday morning: how to diagnose your own function

Before you announce a ZBB program or buy an Anaplan license, run this three-question diagnostic:

Question 1: When the world changed last quarter, how long did it take to reallocate $10 million?

If the answer is more than 30 days, your forecasting cadence is the bottleneck, not your cost discipline. Rolling forecasts are your highest-ROIROIReturn on Investment: the ratio of net profit to the cost of an investment. A 300% ROI means each dollar invested returns $3.View full definition → investment.

Question 2: If I asked your category owners to justify the bottom 20% of their spend, could they?

If you get vague answers about "ongoing programs" or "strategic initiatives," ZBB on that category will pay back in one cycle.

Question 3: What percentage of your FP&A team's time is spent producing reports versus driving decisions?

The benchmark from Hackett Group's 2025 study is that world-class finance functions spend 55% of FP&A time on analysis and decision support. Median is