The 70-20-10 Marketing Budget Rule (And 4 Alternatives)

The 70-20-10 rule is a simple way to split a marketing budget by how certain you are of the return: 70% goes to proven activity that reliably works, 20% to emerging activity that is showing promise but is not yet dependable, and 10% to genuine experiments that will mostly fail. It is a portfolio rule, not a channel rule — it tells you how much risk to carry, not which platform to buy.

It works best for teams that already have some performance history to protect. If you are pre-product-market-fit, or your category is shifting fast, an objective-based or zero-based method will usually serve you better. This guide covers all of them, with the arithmetic spelled out.

What the 70-20-10 rule actually says

Each bucket has a different job, and — this is the part most teams miss — a different success bar.

• 70% — proven. Channels and campaigns with a known, repeatable return. You defend this. It pays the bills.
• 20% — emerging. Things that have worked at small scale and now need budget to prove they scale. This is where last year’s 10% graduates to.
• 10% — experimental. Untested channels, formats, audiences, or messages. Most will not work. You are buying information, not immediate revenue.

The version marketers know best comes from Google’s描述 of how it split engineering effort between core business, adjacent projects and speculative bets, popularised by former CEO Eric Schmidt. Coca-Cola later used a similar split for content investment. Neither origin makes it a law — it is a rule of thumb that survived because the ratios are roughly right for most established businesses.

The genuine insight is not the numbers. It is that you should decide your appetite for uncertainty before you look at any channel, so that the experimental budget is protected rather than raided every time a quarter looks shaky.

A worked example: ₹1.2 crore and $1 million

Take an Indian D2C brand with an annual working media budget of ₹1,20,00,000 (₹1.2 crore, or ₹10 lakh a month).

Bucket	Share	Annual	Monthly	Example spend
Proven	70%	₹84,00,000	₹7,00,000	Google Search brand + non-brand, Meta retargeting, email/CRM
Emerging	20%	₹24,00,000	₹2,00,000	Creator partnerships that beat benchmarks last quarter, regional-language video
Experimental	10%	₹12,00,000	₹1,00,000	Connected TV, WhatsApp commerce, a new marketplace, a podcast sponsorship

Notice what ₹1,00,000 a month actually buys: roughly two to three real tests a quarter, not twelve. That is the honest constraint. A US team on a $1,000,000 budget gets $700,000 / $200,000 / $100,000 — same logic, more room, so perhaps four to five tests a quarter.

Set a graduation rule up front. Something like: an experiment moves to the 20% bucket if it hits 70% of the blended efficiency of the proven bucket at 3x its test spend. Without a written rule, every experiment stays an experiment forever, or gets killed by whoever shouts loudest in the monthly review.

The other 70-20-10 (and why it confuses people)

There is a second, unrelated 70-20-10 used in content and social: 70% of content is broad and low-risk, 20% is deeper and audience-specific, 10% is high-risk creative. Same digits, completely different subject — one splits money by risk, the other splits editorial output by ambition.

If someone in a meeting says “we run 70-20-10,” ask which one they mean. Teams have run parallel plans for months before discovering they were talking past each other.

Four alternatives worth knowing

1. The 60/40 brand-versus-activation split

Les Binet and Peter Field, working with the UK’s Institute of Practitioners in Advertising (IPA), analysed campaign effectiveness data and concluded that long-run growth comes from splitting spend roughly 60% brand-building and 40% short-term activation for consumer brands. LinkedIn’s B2B Institute has published a comparable figure closer to a 46/54 brand-to-activation split for business-to-business categories.

This answers a different question from 70-20-10. It splits by time horizon, not by certainty. The two are complementary: you can run 60/40 brand-versus-activation and still carve 10% of each side for experiments.

2. Objective-and-task budgeting

You start from the goal, not the pot of money. Define the target (say, 5,000 new paying customers), work back through your funnel maths (conversion rate, lead-to-customer rate, cost per lead), and the required budget falls out of the arithmetic.

Example: 5,000 customers ÷ 20% lead-to-customer rate = 25,000 leads. At a blended ₹800 cost per lead, that is ₹2,00,00,000. Now you have a number to argue about with finance — one built from assumptions anyone can challenge line by line.

3. Zero-based budgeting

Every line item starts at zero each planning cycle and must be justified from scratch. No “last year plus 8%.” It is the most rigorous method and the most exhausting; large consumer goods companies use it to strip out legacy spend that nobody remembers approving.

Run it every two or three years rather than annually, unless you enjoy the process.

4. Percentage of revenue

The crudest method and the most common: pick a percentage of revenue and spend it. Gartner’s annual CMO Spend Survey has put average marketing budgets in the region of 7–8% of company revenue in recent years, though the spread across industries is enormous.

Its flaw is that it is backwards-looking and pro-cyclical — you cut marketing exactly when sales dip, which is usually when you can least afford to disappear.

Comparison table: which framework fits you

Framework	Splits budget by	Best for	Main weakness
70-20-10	Certainty of return	Established teams with performance history	Assumes you already know what “proven” is
60/40 brand vs activation	Time horizon	Consumer brands playing a multi-year game	Hard to defend when finance wants quarterly proof
Objective-and-task	Required outcome	Startups, new market entries, launches	Only as good as your funnel assumptions
Zero-based	Justified need per line	Bloated budgets needing a reset	Very heavy to run; risks cutting brand spend
Percentage of revenue	Company size	Board-level planning, quick sanity checks	Ignores opportunity and competitive pressure

Most mature teams stack two: a percentage-of-revenue number to set the total, then 70-20-10 or 60/40 to divide it.

How to split the 70% across channels

70-20-10 gets you three buckets. It does not tell you whether the proven ₹84 lakh should lean Search or Meta. For that you need evidence about incremental contribution, and there are three practical options.

1. Platform attribution — fast, free, and systematically flatters whichever platform is reporting. Useful for in-channel optimisation, unreliable for cross-channel allocation.
2. Geo holdouts and incrementality tests — switch a channel off in a matched set of cities and measure what actually changes. Cheap, causal, and the single highest-value thing most mid-sized Indian teams are not doing.
3. Marketing mix modelling (MMM) — a statistical model of how spend across all channels, plus price, seasonality and distribution, drives sales. Privacy-proof, covers offline, but needs two to three years of data and does not settle small tactical questions.

A sensible sequence: use MMM to set the channel-level shares once or twice a year, use incrementality tests to validate anything the model finds surprising, and use platform data for day-to-day optimisation inside a channel.

Common mistakes

• Judging the 10% by the 70%’s standards. If an experiment must hit the same return on ad spend as your best proven channel in month one, you do not have an experimental budget. You have a rounding error with a nice name.
• Raiding the experiments first. When targets slip, the 10% is the easiest thing to cut and the most expensive to lose — it is your pipeline of future proven channels.
• Confusing working and non-working spend. Agency retainers, martech subscriptions and production costs are not media. Decide explicitly whether your ₹1.2 crore includes them; teams routinely discover mid-year that it did.
• No graduation rule. Without a written promotion threshold, the 20% bucket becomes a graveyard of things nobody will kill.
• Rebalancing monthly. Brand effects take quarters to show up. Reallocating every month means you are mostly reacting to noise.
• Applying it to a ₹5 lakh budget. Ten percent of ₹5 lakh is ₹50,000 a year, which funds roughly one weak test. Below about ₹50 lakh annually, run objective-and-task and reserve one deliberate quarterly experiment instead.

What this means for you

• Write down which framework you are using and why, in one paragraph, before the next planning cycle. Most budget arguments are actually disagreements about the framework, not the numbers.
• Set the total with percentage-of-revenue or objective-and-task; divide it with 70-20-10 or 60/40. Do not try to make one method do both jobs.
• Define your graduation rule in writing this quarter — the specific threshold at which an experiment gets more money.
• Ring-fence the experimental 10% for the full year and report on learning, not return on ad spend.
• If your budget is under roughly ₹50 lakh a year, skip the ratios and run one well-designed test per quarter.
• Before your next annual plan, run at least one geo holdout on your largest channel. It will change how you argue about the 70%.

Frequently asked questions

What is the 70-20-10 rule in marketing?

The 70-20-10 rule allocates 70% of a marketing budget to proven activity with a known return, 20% to emerging activity that is showing promise, and 10% to untested experiments. It divides spend by how certain the return is, so that innovation gets funded on purpose rather than with whatever is left over.

Is 70-20-10 the same as the 60/40 brand-versus-activation rule?

No. 70-20-10 splits budget by certainty of return; the 60/40 rule, from Les Binet and Peter Field’s IPA effectiveness research, splits it by time horizon — long-term brand building versus short-term activation. They answer different questions and can be used together.

What percentage of revenue should go to marketing?

There is no universal figure. Gartner’s CMO Spend Survey has reported averages in the region of 7–8% of revenue in recent years, but the range across industries is wide — subscription software and consumer brands often spend far more, while low-margin distribution businesses spend far less. Treat any benchmark as a sanity check, not a target.

How often should I rebalance my marketing budget?

Review the split quarterly and rebalance meaningfully once or twice a year. Brand and category effects take months to appear in results, so monthly reallocation usually means chasing noise. Reserve in-month shifts for genuine breakages, such as a channel’s cost per acquisition doubling.