Marketing Budget Planning

The marketing director who turned a £180K budget cut into a £250K increase

In 2020, a SaaS company's board decided to cut marketing spend by 30% during the pandemic. The marketing director, Elena, had two weeks to respond.

Rather than arguing for budget preservation, she ran the numbers. She calculated the contribution of each channel:

• SEO content: £42 CAC, producing 34% of pipeline. Return: 8:1.
• Google Search: £88 CAC, 28% of pipeline. Return: 5:1.
• Events (now impossible): £240 CAC, 12% of pipeline. Return: 1.8:1.
• LinkedIn ads: £340 CAC, 8% of pipeline. Return: 1.2:1.
• Trade publications (print): £620 CAC, 3% of pipeline. Return: 0.7:1.

She presented the analysis to the board and made a counterintuitive proposal: cut events, LinkedIn ads, and trade publications (the low-return channels). Increase SEO and paid search.

Total budget decreased. Pipeline increased. By Q3, the board was asking her to spend more.

She'd transformed budget from an expense line into an investment portfolio. The pandemic cut became the most important strategic clarity she'd achieved in 5 years.

(Illustrative scenario based on patterns common in B2B SaaS marketing. Specific figures are representative of real-world outcomes — not a verified account of a specific named company.)

How to think about marketing budget

Marketing budget as investment: Every pound of marketing spend should have an expected return. CAC, LTV, and ROAS convert marketing spend from an expense line into an investment calculation.

Investment framing:
"We are proposing to invest £50,000 in Google Search advertising.
At our proven £65 CAC, this will acquire approximately 769 customers.
At our £390 average LTV, this represents £300K in expected lifetime value —
a 6:1 ROI on the investment."

This is how marketing should be presented to finance and leadership — not "we need £50K for marketing" but "this is the expected return on the investment."

The percentage-of-revenue model:

A common benchmark: marketing spend as a % of revenue varies by industry and stage:

Source: approximate benchmarks from Gartner CMO surveys, Deloitte CMO surveys, and SaaStr community data. Figures vary significantly by company stage, growth target, and competitive environment.

The percentage model is useful for benchmarking but not for planning. Plan from desired outcomes backward, not from revenue forward.

The outcome-based budget model

Step 1: Define the revenue goal "We need £1M in new customer revenue this year."

Step 2: Work backward to marketing requirements

• Average customer value: £5,000/year
• New customers needed: 200
• Historical lead-to-customer conversion rate: 15%
• Leads needed: 1,333
• Historical CPL across channels: £45
• Required marketing spend: £60,000

Step 3: Allocate by channel Based on channel performance data (CAC and LTV by source):

Step 4: Define contingency and testing budget Reserve 10–15% of marketing budget for testing new channels or creative. Without test budget, marketing stagnates — existing channels become the only options forever.

The quarterly vs. annual budget: Plan annually; review quarterly. Annual plans set the direction and resource expectations. Quarterly reviews allow reallocation based on actual performance: channels that over-perform get more; channels that under-perform get less.

Defending the marketing budget to leadership

The language of ROI: Translate every marketing investment into business outcomes. Not "we need more content team" but "content marketing currently generates £8 in customer revenue for every £1 invested, with 3-year compounding value as SEO rankings improve. An additional £30K in content investment is projected to add £240K in annual pipeline within 18 months."

The cost of not spending: If the marketing budget is cut, what doesn't happen? "Cutting the Google Search budget from £20K to £10K will reduce monthly qualified leads from 320 to 160, reducing projected new customer acquisitions from 48 to 24 per month. At £5,000 average ACV, that's £1.44M less ARR generated per year from this channel."

The benchmark comparison: "Industry average marketing spend for our stage is 15–20% of ARR. We are spending 10%. The gap suggests we may be underinvested relative to the growth rate our business targets."

The test-and-prove approach: For new channel requests: propose a contained test. "I'd like £5,000 to test LinkedIn ads over 60 days. Success criterion: CPL below £80 and 2+ customers acquired. If it hits the benchmark, I'll bring a proposal to scale to £20K/month. If it doesn't, we stop."

Back to Elena

Elena didn't convince the board with persuasion — she convinced them with arithmetic. The channel ROI data she presented wasn't new information; it had existed in her reporting for years. What changed was that she organised it in the language the CFO already spoke: investment, return, and opportunity cost. The £250K increase happened because she demonstrated that cutting the right channels could improve pipeline, not damage it. The budget increase was the evidence she already had, reorganised to answer the question the board was actually asking: "Is this money working?" The lesson isn't to fight for budget — it's to make the case so clear that leadership asks you to spend more.

Key takeaways

• Frame marketing as an investment, not an expense. CAC, LTV, and channel ROI translate spending into expected returns — the language finance and leadership understand.
• Plan from desired outcomes backward. Revenue goal → customers needed → leads needed → budget required. Not: "What's our budget?" → "What can we do with it?"
• Allocate based on proven channel performance. High LTV:CAC channels get more; low-performing channels get less. A portfolio of channels with clear performance data is defensible.
• Reserve 10–15% for testing. Without test budget, marketing optimises the same channels forever and never discovers better ones.
• Present new channel requests as contained tests with clear success criteria. A £5K test with defined pass/fail is easier to approve than a £50K channel investment with ambiguous goals.