TL;DR
- There is no universal marketing budget. The often-quoted 7 to 8% of revenue comes from US benchmarks and means little without context.
- Before you set a number, answer three questions: which channel brings the most money, which audience buys most often, and who you are actually advertising to.
- Use simple models as a starting point: a share of revenue checked bottom-up, the most you can pay for one customer, 70/20/10 and the balance between brand and sales.
- Track eight metrics, including AI visibility and brand searches, which in 2027 are no longer optional.
- Below: a quick calculator, what €100 and €1,000 a month can really do, the most common mistakes and a checklist.
Have you been in a meeting where someone asks how much you have for marketing, and you answer: “That depends on your recommendation”? It happens to me all the time with business owners. When we get to money, there is usually an awkward silence.
This guide is here so you can answer that question yourself, and stop handing the decision to someone who doesn’t know your business the way you do.
The short answer
Why averages mislead
| Benchmark | What it says | What to keep in mind |
|---|---|---|
| Gartner CMO Spend Survey 2026 | Marketing budgets average 7.8% of company revenue | Mostly companies with over $1 billion in revenue |
| US Small Business Administration | 7 to 8% of gross revenue | For businesses under $5 million in sales with 10 to 12% net margin |
| By business model | Roughly 8% for B2B products up to about 15% for consumer services | Big differences between industries and markets |
The 7 to 8% figure is usually quoted without its source, the US market. There, a small business means under $5 million in revenue. In many European markets, a company with a few million in revenue already calls itself medium or large. So start with the benchmarks, but interpret them for what you actually do.
Three questions before any number
These are the three questions I rarely get an answer to, regardless of how big the business is:
- Which channel brings you the most money? Not the most clicks or likes: money.
- Which audience buys most often? Who they are, in which situations they buy, what triggers the purchase.
- Who are you advertising to? If the answer is “everyone aged 18 to 65 in the country”, no budget will ever be enough.
You can’t plan 2027 well if you don’t know what 2026 brought. And most businesses don’t: in Nielsen’s 2025 research, 85% of marketers said they were confident in measuring marketing ROI, but only 32% evaluated their channels together. Confidence is not measurement.
Where to look first
Even a spreadsheet of clients, deals and inquiries. That's where the money is, not the clicks: who pays most, who returns products, where closed deals come from.
Where you appear, what brings clicks, and now also the reports for AI features. If you don't have them, install them right after reading this.
Precise for well-structured campaigns, only a rough guide for broad ones. Split campaigns by audience to find what works.
"How did you hear about us?" In the form, at the checkout, on the phone. It catches what no software sees, including AI recommendations.
Four ways to set the budget
None of these replaces your own data, and none is a law. Use them as a starting point.
| Model | How it works | Watch out |
|---|---|---|
| Share of revenue, checked bottom-up | Start with a percentage of revenue, then check how many new customers you need and what one costs | If the two numbers are far apart, the goal or the budget is unrealistic |
| The most you can pay for a customer | Customer lifetime value × gross margin ÷ 3 | The 3:1 ratio is a common rule of thumb, not a law |
| 70/20/10 | 70% on proven channels, 20% on growing ones, 10% on experiments | A practical rule, not research. Startups test more. |
| Brand and sales | Research by Binet and Field points to about 60% brand and 40% short-term sales for consumer brands; in B2B the brand share is smaller, but still substantial | In AI search, brand is no longer only image: AI recommends brands it finds mentioned and documented |
Channels where a customer costs more than your limit lose money. Channels below it make money. Move budget towards the second group.
Calculate your limit
Eight metrics to track
| Metric | What it tells you |
|---|---|
| Cost per customer by channel | Where a customer costs least |
| Where your best customers come from | Which channels bring people who stay and buy again |
| Payback period | How long your money is tied up before it returns |
| Total marketing return | Revenue against all marketing spend: does marketing work as a system, not campaign by campaign |
| Share of repeat purchases | How much to spend on keeping customers versus finding new ones |
| Traffic and revenue from AI assistants | How much of your business already comes through ChatGPT, Gemini and others |
| How often AI mentions your brand | Whether you are in the answers at all. In 2027, this is not optional. |
| Searches for your brand name | Whether brand investment works and people remember your name |
€100 or €1,000 a month
One of the most common questions I get. The question is set up wrong: what matters is not how much you spend, but what each euro has to achieve.
About €3.30 a day. At €0.50 per click and 2 buyers out of 100 visitors, that is around 200 visitors and 4 sales a month. Meta's algorithm needs about 50 conversions a week to learn. With 4 a month, it guesses. So use €100 for precise things: remarketing, ads on your brand name, local ads for a few neighborhoods, or boosting a post that already works.
Enough to test, but I see the same mistake as with €100: €1,000 split across two platforms and ten campaigns is €100 per campaign again. Start with one platform and one goal, and expand only when you have results.
Which platform first? Google captures demand that already exists: people searching for what you sell. Facebook and Instagram create demand: they show you to people who weren’t looking. If nobody searches for your product, Google won’t help much. If they do, start there.
Common mistakes
1. No remarketing
You pay to bring a visitor and then let them leave foreverPlan budget for bringing visitors back to finish an order or buy again
Fix first
2. Spending without measuring
No link between the ad and the sale, so no idea what worksMonths of budget split between channels blindly
Fix first
3. Advertising to everyone aged 18 to 65
When you talk to everyone, you talk to no oneKnow who buys, and what offer suits each group
Strategy
4. Everything on one platform
Prices can rise and you have nowhere to move the moneyThe same applies to businesses that live only on Instagram or only on referrals
Risk
5. Trusting each platform's own numbers
Facebook, Instagram, Google and now ChatGPT each claim the same saleCompare platform reports with your CRM and analytics
Fix first
6. Forgotten customers
Thousands of emails and phone numbers, and not a single campaign to themExisting customers are usually the biggest gold mine
Quick win
7. Cutting brand and content because "they don't sell"
Content feeds AI with information about youAnd helps AI-assisted sales
Long term
8. A budget set once a year
A rough calculation in November, then the whole year without changesReview against the plan every month
Process
At the start, the budget buys information, not sales. Run short tests with a clear goal and deadline: four weeks, one channel, one message, a KPI written down in advance. Set up measurement before the first euro, and give it time. A new site and a new ad account need weeks for the algorithm to learn who you are. Stopping ads on day 12 tells you nothing.
The risk is the opposite: inertia. Budgets get copied from last year because "it works". Ask what you could stop without losing anything. Test by switching off a channel or region for a while, invest more in retention and repeat purchases, and protect your brand, PR and product data, because they now decide whether AI recommends you or a competitor.
What is new in 2027
- GEO and digital PRIf you haven't prepared your business for AI search and AI shopping, now is the time. Start with at least one audit, so you know the scope of work ahead.
- Product dataDescriptions, prices, availability, reviews and feeds, updated and detailed: sizes, materials, how it's made, not just quantities.
- Testing ads in AIA small, conservative budget line. ChatGPT ads became available in 31 European markets in August 2026, and Meta plans to automate ad creation fully with AI.
- Measuring AISet up tracking for AI-assisted traffic, citations and sales now. Every month of delay is another month without the data you need to plan.
Your checklist
Six things to have in hand
My take
There is no right answer to €100 or €1,000. Both can be enough, and both can be too little. The real question is what each euro has to bring back, and whether you can measure it.
The businesses that will plan 2027 well are not the ones with the biggest budgets. They are the ones that know where their money came from in 2026.
If you split your budget in a way I haven’t mentioned here, I’d like to hear how.
For a neutral overview of the benchmarks and models, read Marketing Budget Benchmarks and Allocation Models, Explained on Egorithms.
About The Author
