Strategy 11 min read

How much to spend on ads: the budget math from $3K test to $100K program

Marketing runs 7.7 to 9.4% of revenue in the 2025 surveys, and paid media takes about a third of that. Here's how to turn averages into your actual monthly number, and the floors below which ads can't work at all.

How much should you spend on ads? At the median, companies put 7.7 to 9.4% of revenue into marketing and about 31% of that into paid media, so roughly 2 to 3% of revenue goes to ads. For a B2B SaaS that usually lands between $2,000 a month (pre-revenue test) and $100,000+ a month (a $20M+ ARR program).

That's the survey answer, and it's the right sanity check. It's also the wrong way to set your budget, because percent-of-revenue rules are how boards audit spending, not how operators plan it. Your actual number comes from three harder constraints: what a customer is worth to you, what the platforms need before their delivery stabilizes, and how much proof you've already accumulated. This post walks all three, with the same numbers we baked into our free ad spend calculator.

The short version

  • The medians: marketing is 7.7–9.4% of revenue, paid media ~31% of marketing. B2B product companies budget 6.4% of revenue for marketing; B2C product companies 15.5%. DTC brands under $5M often put 20–30% of revenue into ads alone.
  • Set the budget from payback, not percentages. Deal size → acceptable CAC → media's ~40% share of it → target volume. The percentage is the check at the end, not the input.
  • Respect the floors. One channel run properly needs $1,500–$3,000 a month minimum. Meta wants ~50 conversions per ad set per week; Google wants 30+ before you trust Smart Bidding. $5/day is a lottery ticket.
  • A first test buys information, not customers. Fund 30–50 conversions in 4–8 weeks on one channel, or don't start yet.
  • Skip the brand carve-out early. Run performance-heavy until a few $M in revenue, then build toward the 30%+ brand baseline the 2025 WARC data supports.

The short answer, in dollars

Here are the monthly paid-media bands we use, by revenue stage and risk appetite. They're anchored to the survey data below plus practitioner ranges for seed-to-Series-B startups, and they assume the money is concentrated on one or two channels rather than sprinkled across five:

Revenue stageCarefulModerateAggressive
Pre-revenue$1,000–2,500$2,500–5,000$5,000–10,000
Under $1M$2,000–4,000$4,000–8,000$8,000–15,000
$1–5M$5,000–9,000$9,000–18,000$18,000–35,000
$5–20M$12,000–22,000$22,000–45,000$45,000–90,000
$20M+$30,000–55,000$55,000–110,000$110,000–220,000

"Careful" means you want fast payback and would rather under-spend than waste; "aggressive" means you're funded, you accept the 18-to-24-month payback venture norm, and speed matters more than efficiency. Most bootstrapped companies belong one column left of where their ambition wants them to be.

What the surveys actually say

Two big surveys publish marketing-budget-as-percent-of-revenue every year, and in 2025 they disagreed usefully. Gartner's CMO Spend Survey has budgets flat at 7.7% of company revenue for a second year, with half of CMOs at 6% or less, and paid media as the single largest line at 30.6% of the marketing budget (about 2.4% of revenue).¹ The CMO Survey (Duke/Deloitte/AMA) has marketing at 9.4% of revenue, rebounding from 7.7% in late 2024, with B2B product companies at 6.4%, B2B services at 9.0% and B2C product at 15.5%.² The gap is the sample: Gartner skews to $1B+ enterprises, The CMO Survey includes mid-market. Small companies spend a higher percentage; big ones spend bigger absolute dollars at lower rates.

For SaaS specifically, SaaS Capital's 2026 survey of 1,000+ private B2B SaaS companies puts median sales spend at 15% of ARR and marketing at 8%, with bootstrapped companies running about half the equity-backed levels.³ Benchmarkit's 2025 report adds the efficiency lens: the median company spends $2.00 of sales and marketing for every $1.00 of new ARR, and total S&M runs 37% of revenue at the median (47% for VC-backed).

E-commerce runs much hotter. Stage-based P&L data from DTC finance practice Eightx puts ad spend alone at 25–35% of revenue under $1M, 15–25% at $5–10M, and 7–15% past $50M. Triple Whale's platform data across 33,000+ brands agrees: median blended ad spend was 41% of revenue in 2025, which is a 2.4x marketing efficiency ratio. The pattern worth stealing from the DTC world: the ratio must fall as you scale. If you're still spending like a sub-$1M brand at $25M, you have a retention problem wearing a marketing costume.

Percent-of-revenue is a rearview mirror: it describes what surviving companies spent, not what your next dollar should do. Set the budget from payback, then check the percentage.

The floors the platforms enforce

Before any of the strategy matters, there's a mechanical constraint: every platform's delivery system is a learning machine that needs a minimum diet of conversion events. Under-feed it and you don't get proportionally smaller results, you get noise.

The official numbers: Meta says an ad set needs about 50 optimization events in a week to exit the learning phase, and the bar is per ad set, not per account. Google publishes no minimum budget at all (the "$10 to $50 a day" figure that circulates in blog posts isn't Google's), but Target ROAS bidding wants at least 15 conversions in 30 days and Google recommends judging Smart Bidding on 30 or more. LinkedIn's floor is $10 a day, but LinkedIn itself suggests $25 a day for new advertisers and $50–100 for real ones, and the best-known LinkedIn ads specialist, B2Linked, calls $5,000 a month the realistic entry point because at $10–16 CPCs you need roughly 300 clicks before conversion data means anything.¹¹ TikTok enforces $50 a day at campaign level and $20 per ad group.¹⁰

Translate those signal requirements into practical monthly minimums per channel and you get:

ChannelPractical monthly minimumWhat sets the floor
Google Search$2,00030+ conversions to trust Smart Bidding; SMB starting ranges
Meta$2,500~50 events per ad set per week; daily budget ≥ 5× cost-per-result goal
LinkedIn$3,000$8–15 CPCs; ~300 clicks before conversion data is real
TikTok$1,500$50/day campaign minimum; ~25–50 conversions/week for stability
Reddit$1,500$50–150/day to exit learning on conversion campaigns
X$1,200Cheap clicks, weak conversion signal; treat as experimental

This is why the budget bands concentrate: under $3,000 a month you run one channel, under $9,000 two, under $25,000 three. Spreading $4,000 across four platforms guarantees all four stay in learning-phase pricing forever. We covered where the dollars go once you can afford multiple channels in the budget split post; the point here is that the number of channels is an output of the budget, never the starting plan.

Work backwards from payback

Here's the method that actually sets the number, in five steps. It works for any B2B company; the DTC version swaps deal size for contribution margin and runs the same direction.

  1. Start with deal size and payback tolerance. Benchmarkit's payback norms by deal size: about 9 months under $5K ACV, 12 months at $10–25K, 18–24 months for enterprise.
  2. Derive the CAC ceiling. CAC ≈ ACV × gross margin × (payback months ÷ 12). A $10K ACV at 80% margin and 12-month payback gives you $8,000 of fully-loaded CAC to play with.
  3. Take media's slice. Fully-loaded CAC includes salaries, tools and creative. Paid media is roughly 40% of it as a planning figure, so $3,200 per customer in this example.
  4. Multiply by target volume. Want three paid-sourced customers a month? At an 8–9% demo-request-to-customer rate for cold traffic, that's about 36 demo requests, at $100–300 each on Google or LinkedIn. Budget: $5,000–$10,000 a month.
  5. Sanity-check the percentage. If that budget is 40% of revenue and you're not a funded rocket, the target volume is wrong, not the math.

Notice what did the work: deal size and conversion rates, not a revenue percentage. A $3K-ACV product and a $30K-ACV product at the same revenue should have completely different ad budgets, which is exactly what the percent rules can't see. We published the full funnel arithmetic behind step 4 in the cost-per-demo post, and the calculator runs this whole chain for your inputs, including the channel split.

The budget that comes out of payback math is allowed to be zero. If your ACV is $1,200, your margin is thin and cold leads close at 3%, the honest output is "ads don't work at your economics yet." That's the calculator telling you to fix pricing or conversion first, and it's cheaper to learn it from arithmetic than from $20K of spend.

Sizing a first test budget

A first ad budget has one job: buy enough signal to make a keep-or-kill decision. Size it in conversions, not dollars. The practitioner consensus clusters tightly here: fund roughly 50 conversions per ad set to clear Meta's learning bar (which back-solves to target CPA × 50, so a $30 CPA needs about $1,500 per ad set per week of learning), or on Google, target CPA × 20–30 conversions as the monthly floor, the way independent Google Ads consultant Sarah Stemen sizes it: a $50 CPA and 30 conversions means $1,500 a month, about $49 a day.¹³

Three corollaries that save people money:

  • If you can't fund 30 conversions in 8 weeks, don't start. A $500 test that produces 4 conversions tells you nothing except that you spent $500. Wait until the budget can buy a real answer.
  • Judge clicks and conversions on different budgets. $1,000 on LinkedIn tells you your CTR and CPC; it cannot tell you whether LinkedIn produces customers.¹¹ Don't kill a channel on conversion data it never had the budget to generate.
  • One variable at a time. A first test is one channel, one offer, two or three creatives. Every extra simultaneous test divides the conversion signal further below the learning thresholds above.

The brand carve-out question

Every budget conversation eventually hits the Binet & Field question: shouldn't 60% of this go to brand building? The 60/40 rule is from 2013, and the B2B-specific work put the optimum closer to 46/54 brand-to-activation back in 2019. The 2025 update is WARC's Multiplier Effect study, which drops the fixed ratio for a "brand baseline": at least ~30% of budget on equity-building ads, with 40–60% as the best-practice range, and a striking pair of findings on the way there: moving from performance-only to a mixed strategy raised median revenue ROI about 90%, and putting more than a quarter of the budget into search is flagged as a red flag, since a chunk of search clicks are generated by your other marketing anyway.¹² Meanwhile actual practice runs 31% brand to 69% performance,² so almost everyone is below the baseline.

My take for the companies in the budget table above: the brand baseline is right and also mostly premature. Under a few million in revenue, every dollar needs to come back inside a payback window you can survive, and "brand" at that size is your product, your founder's posts and your landing pages, none of which need a media budget. Start carving out an explicit awareness slice (5–15%) somewhere past $5M in revenue, funded or not, and grow it toward the WARC range as attribution stops being your binding constraint. What you should never do is run the 60/40 split on a $4,000 budget: $1,600 of brand media is homeopathy, and the $2,400 remainder is below every floor in the table above.

Frequently asked questions

What percentage of revenue should be spent on advertising?

Marketing overall runs 7.7 to 9.4 percent of company revenue in the 2025 surveys (Gartner and The CMO Survey), and paid media takes about 31 percent of that, so roughly 2 to 3 percent of revenue goes to ads at the median. B2C spends more: consumer product companies budget around 15.5 percent of revenue for marketing, and DTC brands under $5M often spend 20 to 30 percent of revenue on ads alone. Treat these as sanity checks, not targets: the right number comes from your payback math.

Is $5 a day enough for Facebook ads?

Not for conversion campaigns. Meta's system wants about 50 optimization events per ad set per week to exit the learning phase, so at a $30 cost per result that implies roughly $210 a day, and $5 buys a fraction of one conversion. At $150 a month the only honest uses are retargeting a small warm audience or buying click and CTR data, not purchases.

How much should a small business spend on ads per month?

A practical floor for one channel run properly is $1,500 to $3,000 a month: around $2,000 on Google Search, $2,500 on Meta, or $3,000 on LinkedIn. Below that, campaigns sit in the learning phase indefinitely and results whipsaw month to month. If that is more than the business can risk, it is usually better to spend nothing and invest in organic channels than to run $300 a month everywhere.

What's a good first ad budget for a B2B SaaS?

Size it to buy a real test: enough for 30 to 50 conversions in 4 to 8 weeks on one channel. If your target cost per demo request is $150, that is $4,500 to $7,500 of total test budget. Our budget bands put pre-revenue and sub-$1M companies at $2,000 to $8,000 a month depending on risk appetite, almost always concentrated on a single channel.

When should you increase your ad budget?

When the current budget is reliably hitting its numbers: CAC payback inside your target (the published median is around 16 months for B2B SaaS), cost per opportunity stable for 6 to 8 weeks, and campaigns out of the learning phase. Then raise spend in steps of 20 to 30 percent rather than doubling overnight, and add a second channel only once the budget clears about $3,000 a month.

Sources

  1. Marketing budgets at 7.7% of revenue, paid media 30.6% of budgets — CMO Spend Survey 2025, Gartner
  2. Marketing at 9.4% of revenue, sector breakdowns, 31/69 brand-performance split — The CMO Survey, Spring 2025 (Duke/Deloitte/AMA)
  3. Median sales 15% and marketing 8% of ARR, 1,000+ private SaaS companies — 2026 Spending Benchmarks, SaaS Capital
  4. $2.00 S&M per $1 new ARR, payback norms by deal size — 2025 SaaS Performance Benchmarks, Benchmarkit
  5. DTC ad spend as % of revenue by stage — Ad Spend as a Percent of Revenue by Stage, Eightx (2026)
  6. Median blended ad spend 41% of revenue across 33,000+ brands — Marketing Efficiency Ratio, Triple Whale (2025 data)
  7. ~50 optimization events per ad set per week to exit learning — About the learning phase, Meta (official)
  8. Target ROAS conversion thresholds, Smart Bidding guidance — About Target ROAS bidding, Google (official)
  9. $10/day floor and suggested $25–100/day budgets — Making the most of your budget, LinkedIn (official)
  10. $50/day campaign and $20/day ad group minimums — About budgets, TikTok (official)
  11. $5K/month entry point, ~300 clicks before judging — How much should I budget for LinkedIn Ads, B2Linked (2023)
  12. Brand baseline ~30%, mixed-strategy ROI +90%, search ceiling — The Multiplier Effect, WARC (2025)
  13. Budget = target CPA × 20–30 conversions monthly floor — How much to spend on Google Ads, Sarah Stemen (2026)
Robin Choy

Founder of Adside. Writes about the operational side of running ads at agency scale: what to automate, what to keep human, and what the data actually says.

Know when a channel earns more budget

Adside tracks spend, cost per lead and payback across Google, Meta and LinkedIn in one place, so raising a budget is a decision you make from data instead of a guess you make from hope.