Marketing Strategy / Field guide
How Much Should a Startup Spend on Marketing? A Founder's Method
AlexPublished · Updated
How to set a startup marketing budget when you have little data: three methods, working backwards from customers, how big a channel test needs to be, the costs beyond ad spend, and when to scale, hold or cut.
The short answer: budget for learning first, then for growth
"How much should we spend on marketing?" rarely has a single right number. Rules like "spend a set percentage of revenue" assume steady revenue, and most startups don't have that yet.
A better way to think about it: a startup's first marketing budget is mainly a learning budget. Its job is to find out which channels bring customers at a cost you can afford. Only then does spending more make sense.
So split the question in two:
- Learning budget: how much do you need to test two or three channels properly?
- Growth budget: once a channel works, how much can you put behind it while it stays profitable?
Two terms used throughout: cost per customer is the spend in a period divided by the new paying customers it brought. Customer value is the profit a customer brings you, ideally over the whole time they buy from you.
Three ways to set a marketing budget
Founders usually choose between three methods. For most startups, the answer is a mix of two of them.
Spend a fixed share of current or expected revenue
Good forEstablished businesses with steady revenue
WeaknessDoesn’t work pre-revenue; ignores your growth goal
Start from the customers you need and the cost to win each
Good forStartups with a target and rough cost estimates
WeaknessEarly estimates are often wrong; revisit monthly
Set an amount per channel to learn whether it works
Good forPre-revenue or brand-new businesses
WeaknessShows what works, not how far it can scale
For most startups: goal-based to size the plan, plus a fixed test budget per channel to check the assumptions.
Whichever you choose, write it down as part of a plan. The US Small Business Administration's marketing and sales guide recommends a marketing plan with its costs broken down. Our guide to building a marketing plan for a new business shows how to write one on a single page.
Work backwards from the customers you need
The most useful budget starts from a customer target, not a round number. You need four inputs:
- New customers you want in a period, such as per month.
- What a customer is worth: at least the profit from a first purchase, ideally over the time they stay.
- Lead-to-customer rate: how many enquiries become paying customers. Start with a cautious guess.
- Cost per lead for a channel: estimate it with tools like Google's Keyword Planner, then replace it with real data.
The formula: customers needed ÷ lead-to-customer rate × cost per lead = monthly budget.
- New patients needed per month15
- Enquiries that become patients1 / 4
- Enquiries needed15 ÷ 0.25 = 60
- Cost per enquiry from search ads (estimate)$40
- Monthly ad budget60 × $40 = $2,400
- Cost per patient$2,400 ÷ 15 = $160
A patient brings $600 profit. $160 per patient is comfortable: test the plan.
A patient brings $120. Each costs $160: a loss. Find a cheaper channel, a better conversion rate or a higher-value offer.
Illustrative numbers, not client data. The same maths shows a plan that can’t work before you spend anything on it.
Want to try it with your own numbers? The budget calculator in our marketing plan guide runs the same maths.
Size a test budget for each channel
Give each channel a fixed amount and a fixed period, big enough to give a clear answer. A test that ends after three leads tells you nothing.
The rule we plan with:
- Aim for roughly 20–30 conversions (leads, bookings or sign-ups) per channel, or enough spend to be confident it won't get there.
- Run it for 4–8 weeks so you see more than one good or bad week.
- Test two or three channels at most at once. Six channels on a small budget give you six unclear answers.
- $15 per lead$375
- $30 per lead$750
- $60 per lead$1,500
- $120 per lead$3,000
Maths: 25 leads × cost per lead. Test two or three channels at most at once.
Aim for about 20–30 conversions per channel over 4–8 weeks. If the test would cost more than you can afford, start with a cheaper channel.
Platforms usually take a daily budget. For Google Ads, Google suggests dividing a monthly budget by 30.4 to get the average daily amount. Our Google Ads guide for small businesses covers the setup.
Don't forget the costs outside ad spend
Ad spend is only part of a marketing budget. The SBA's guide to calculating startup costs lists advertising, market research, printed materials and a website as separate expenses, and founders often forget everything except the ads.
- Ad spend (tests)$1,500 · 50%Two channels, $750 each
- Website & landing pages$500 · 17%Upkeep, a page per offer
- Tracking & CRM$250 · 8%Analytics, conversions, lead records
- Creative$450 · 15%Photos, video, ad designs, copy
- Tools$100 · 3%Email, scheduling, design
- People$200 · 7%Freelancer hours or agency support
TakeawayHalf the budget isn’t ads. If little is left for tests after the rest, start with the website and SEO instead.
Illustrative split for an early-stage service business. Plan the non-ad costs first, then see what’s left for testing.
Plan these first: the website and landing pages, tracking (analytics, conversion tracking and a CRM), creative, tools and people's time. Then decide how much is left for testing channels. If very little is, a strong website and SEO may be a better first investment than ads.
How to split the budget over time
Once something works, put most of the budget behind it and keep a smaller, fixed share for tests. A common planning habit is 70/20/10:
- About 70% on channels that already bring customers at an acceptable cost.
- About 20% on promising channels you're still optimising.
- About 10% on new experiments.
At the very start nothing is proven, so almost everything is test budget. The split becomes useful from about the third month.
- M1100%
- M230%70%
- M340%40%20%
- M460%25%15%
- M570%20%
- M670%20%
- Proven
- Optimising
- Testing
Illustrative. The 70/20/10 split is a planning habit, not a benchmark; it only makes sense once a channel is proven, usually from month three.
When to scale, hold or cut a channel
Decide with cost per customer, not clicks or likes. At the end of each test period, compare it with what a customer is worth:
- Cost per customer well below what a customer is worthScaleAdd 20–30% a month and watch whether costs rise
- Cost per customer close to the limitHoldImprove the landing page, offer or follow-up before adding budget
- Few conversions after a full testCutOr rethink the offer and audience before testing again
- Lots of leads, few customersFix follow-upCheck reply speed and lead quality before blaming the channel
Cost per customer = total spend on a channel in a period ÷ new paying customers it brought.
The last row is the one founders miss most. If enquiries arrive but don't turn into customers, more budget only makes the problem bigger. Our guide to lead follow-up and CRM covers how to fix it.
Our take: a small, measured budget beats a big, blind one
The founders who waste the most money usually aren't the ones with small budgets. They're the ones who can't see which spend brought which customer. A $1,500 test with conversion tracking and a CRM teaches you more than $10,000 spread across channels with no tracking.
So before recommending any budget, we check three things:
- 1Can we measure it?
Conversion tracking is on, and every lead’s source is recorded in a CRM or sheet.
- 2Is the destination ready?
A landing page that repeats the offer, loads fast and has one clear action.
- 3Can the team handle the leads?
Someone owns replies, with a target time, every working day.
If one is missing, fix it first. It’s cheaper than paying for leads you can’t see or don’t answer.
Key takeaways
- A startup's first marketing budget is a learning budget.
- Work backwards from customers needed, lead-to-customer rate and cost per lead.
- Give each channel a fixed test aimed at 20–30 conversions over 4–8 weeks, and test two or three channels at most.
- Include website, tracking, creative, tools and people, not just ad spend.
- Scale only channels with an acceptable cost per customer, and fix follow-up before blaming a channel.