The 70/20/10 rule for marketing is an easy method to split up your marketing budgets. Allocate 70% to creative ads that work for you, 20% to test different variations of successful ads, and 10% to untried concepts. And this is all there is to the rule. The hard part is sticking to it every month, because the 70% bucket always wants to eat the other two.
This guide walks through the 70/20/10 marketing budget one bucket at a time, shows a 70/20/10 rule example in rupees, and explains when a test creative earns its place in the proven bucket. For anyone who has ever asked themselves the question “How to budget ad spend?” this is probably the most common question for performance marketers.
What Is the 70/20/10 Rule in Marketing?
The 70/20/10 marketing rule is an allocation of a budget based on the level of trust that you have in each ad. The reliable ads will take 70% of the budget, the better versions of the ads take 20%, and everything new takes 10%.
Here is what each of the three budget allocations does:
- 70% for proven creatives: Ads with a track record. They’ve hit your target cost per result for weeks, and you’d be nervous switching them off.
- 20% for testing creatives: New hooks, angles and formats built on top of the ads that already work. This is your creative testing budget.
- 10% for experimental creatives: Ideas with no history at all. A platform you’ve never used, a message you’ve never tried, a format your competitors aren’t running.
Some teams label the same buckets core, growth and experimental. The names change, the logic doesn’t. You’re grouping ads by confidence rather than by channel or format, which is why the rule works on Meta, Google, LinkedIn or a mix of all three.
Students at Tiju’s Digital Marketing Institute meet this split in the second week of the paid ads module, before they touch a live account. This particular practice is a lot easier to develop early than to rectify later on, which is why it has earned this designation as the “best digital marketing course in Kerala” for those looking to advertise for a profession.
Who Created the 70/20/10 Rule?
Google made the split famous. According to Eric Schmidt, the chief executive officer of Google, their company’s innovation strategy consisted of allocating 70% of their resources to core search business, 20% to other products, and 10% to completely different things. This strategy was used by Coca-Cola somewhere around 2011, according to which they allocate 70% of the material to low-risk content, 20% to proven stuff, and 10% to high-risk content.
Advertisers borrowed the idea for budgets. In the 70/20/10 rule advertising version, the buckets hold creatives and channels instead of products, and the 70/20/10 rule digital marketing guides you see today are mostly this borrowed version. It sits close to the McKinsey three horizons framework as well, which separates today’s business from emerging opportunities and future bets.
The following are called the same thing but have no connection whatsoever to advertising:
- 70/20/10 rule for personal finance management (70% spending, 20% saving, 10% debt/giving).
- 70:20:10 model of learning in corporate training programs.
When searching for the simple phrase alone, both of these will come up overwhelmingly in search results. This paper will focus only on the advertising 70:20:10 formula.
The Power of the 70-20-10 Rule
The principle holds true since it rewards tomorrow’s winners while the winners of today are busy footing the bills. In the absence of such a principle, most advertising accounts fall prey to one of two vices.
The first habit is running the same three ads until they die. Every creative wears out. Ad fatigue shows up as rising CPM, falling CTR and a CPA that creeps up week after week. A clothing brand in Kochi that we studied in class ran one video ad for five months straight. By month four the cost per purchase had doubled, and nothing was ready to replace it. That is what a 100/0/0 budget looks like.
The second habit is the opposite: chasing every new format and platform with no stable base to fund the experiments. Neither extreme lasts long.
The 70/20/10 budget allocation fixes both by treating your account like a portfolio:
- Risk diversification: Most of the money sits in low-risk ads, a small amount in moderate risk, and a tiny amount in moonshots.
- A steady pipeline: Test creatives keep graduating into the proven bucket, so replacements exist before creative fatigue hits.
- Calculated risk: The 10% is capped, so a failed experiment can’t wreck the month.
- Always-be-testing discipline: Testing becomes a fixed line item instead of something you do when money is left over, which is never.
The ROI of this test-and-learn approach shows up months later, which is exactly why teams without a rule skip it.
The 70% Strategy: Scaling Proven Creatives
70% of your marketing budget is allocated to tested creatives, which are advertisements that have generated your desired return on ad spend or CPA for two to three weeks consistently. This is the budget that is used to scale your business; in other words, your goal here is not to fluctuate but to grow consistently.
What counts as a proven creative? An ad qualifies when it:
- Has left the learning phase (on Meta, roughly 50 optimisation events in a seven-day window).
- Has hit the target cost per result for two to three weeks running, not one lucky week.
- Holds its numbers after you raise the budget by 20 to 30%.
- Works across at least two audiences, for example, retargeting and a lookalike audience.
Each proven ad becomes a control ad, the benchmark every new variation has to beat.
Running the 70% well is mostly patience:
- Use CBO to make the platform allocate more of the budget to the winning creative on any particular day.
- Increase the budgets gradually; increase by 20% after every 3 to 4 days to avoid resetting the learning phase.
- Watch CPM and CTR weekly. When CTR drops by a third from its peak, the ad needs a creative refresh or a replacement from the 20% bucket.
- Keep performance data for each winner in a sheet: spend, CPA, CTR, hook rate, launch date. You’ll need it to judge challengers later.
In Tiju’s digital marketing course, students manage this bucket inside Meta Ads Manager on a live account for a local business, so they learn what “stable” looks like in real reports rather than in a slide.
The 20% Innovation: Testing Creative Variations
Twenty percent of the budget is your creative testing budget. It funds creative variations of ads that already work: same offer, same product, different execution. This is where your ad creative testing framework lives.
What to test in the 20% bucket:
- Hooks: The first three seconds of a video or the first line of a static. A new hook on a proven ad is the cheapest test you can run.
- Ad angles: Price, quality, speed, social proof, urgency. Same product, different reason to buy.
- Formats: Static vs video creatives, carousels, UGC ads, Reels ads. A winning message often works in a format you haven’t tried.
- Headlines, primary text and calls to action.
How to run the tests:
- Write a hypothesis first. “A price-led hook will cut CPA by 15% for cold audiences” is a hypothesis. “Let’s try a new video” is a wish. Hypothesis-driven testing is the difference between learning and spending.
- Change one variable per test. That’s plain A/B testing. Multivariate testing can handle two or three variables at once, but it needs far more spend to read clearly.
- Run tests under ad set budget optimisation (ABO) so each challenger ad gets equal spend instead of the platform picking a favourite on day one.
- Judge early results on hook rate, hold rate and thumb-stop rate. Judge final results on conversion rate and cost per result against the control ad.
- Cap spend per test at two to three times your target CPA. Stop at the cap or at statistical significance, whichever comes first.
Graduation criteria are your rules for promoting a challenger into the 70% bucket. A common one: beat the control on CPA for seven consecutive days at equal spend. Kill criteria are the reverse: two times target CPA with no purchases, switch it off. Write both down before launch so nobody argues after.
Creative volume matters too. Most accounts need four to six fresh variations a month to stay ahead of fatigue, and this creative iteration loop never really stops. In our weekly digital marketing classes, students build two or three variations every Monday, launch on Tuesday, and read results the following week. After a few cycles, choosing the next test becomes automatic.
The 10% Experimentation: Experimental Creatives
Ten percent is your experimentation budget, sometimes called the innovation budget. It’s for ideas with no history in your account: a channel you’ve never used, a message that’s a real departure, a format that’s new to your industry.
Examples of what belongs here:
- First LinkedIn ads for a business that has only run Meta.
- A YouTube ads test when all your video has lived on Instagram.
- A Performance Max campaign or an Advantage+ creative test to see how far the platform’s automation gets on its own.
- A humour angle for a brand that has always played it serious.
- Creator-style UGC from a real customer instead of studio footage.
- A programmatic display test for cheap reach.
Most of these will fail, and that’s fine. You’re buying information. One hit out of every five or six experiments is a healthy rate, and a single hit often pays for the whole year’s 10%.
A few rules keep this bucket from becoming a money pit:
- Cap the spend before launch and don’t top it up mid-test.
- Set a review date on day one. Thirty days is enough for most social tests.
- Every experiment ends with one of three decisions: promote to the 20% bucket, hold for another round, or stop.
- Log what you learned, including failures. “Humour doesn’t work for our audience” is a useful result.
- The proven versus experimental creators distinction is all about expectations: you expect the 70% to perform and the 10% to educate you.
The 70:20:10 Rule for Media Investment in Digital Channels
In fact, the same proportion applies on a higher level of channels. A 70:20:10 media investment strategy invests 70% of your digital media budget in proven channels, 20% in emerging channels, and 10% in experimental channels.
A media portfolio for a furniture store in Kottayam may include:
- 70%: Google Search on brand and product terms, plus Meta retargeting. Both have months of data.
- 20%: Meta prospecting to lookalike audiences, which started working last quarter but isn’t fully stable yet.
- 10%: A first YouTube campaign and a small Pinterest test.
Two things matter in media planning at this level:
- Attribution flatters some channels and punishes others. Retargeting always looks brilliant because it catches people who were already on their way. Judge channels on incrementality where you can: what happens to total sales when you pause a channel for a week?
- Channels move between buckets. This quarter’s 20% channel becomes next quarter’s 70%. A 70% channel can get crowded and expensive and slide back down.
At Tiju’s Digital Marketing Academy, media planning is taught as a spreadsheet exercise first and a platform exercise second, because the split has to make sense on paper before any money moves.
How the 70/20/10 Rule Is Used: A Worked Example?
Take a monthly ad budget of ₹1,00,000, a typical spend for a small brand in Kerala. This 70/20/10 rule example runs over two months.
Month one:
- ₹70,000 proven bucket. Three winning creatives under one CBO campaign. Target CPA ₹500, currently averaging ₹470.
- ₹20,000 testing bucket. Four variations of the best winner, each capped at ₹5,000: two new hooks, one static version of the video, one price-led angle.
- ₹10,000 experimental bucket. One test, a first YouTube Shorts campaign using the top creative, ₹10,000 cap, 30-day review.
By the end of month one:
- Two of the four test variations lost to the control and were switched off at their kill criteria.
- One matched the control and goes another round.
- The price-led angle beat the control at ₹410 CPA for nine straight days. It graduates.
- YouTube Shorts came in at ₹900 CPA. Too high to promote, but the hold rate was strong, so it stays in the 10% for one more month with a different hook.
Month two:
- The proven bucket now holds four creatives. The oldest winner, whose CTR has dropped 40%, gets paused.
- The testing bucket builds three variations off the new price-led winner.
- The experimental bucket reruns YouTube with the new hook.
The money never changed. What changed was which ads sat in which bucket. That is how to allocate ad budget in practice: the percentages stay fixed and the creatives rotate through them.
How Often Should You Rebalance a 70/20/10 Budget?
Review monthly if you spend under ₹2 lakh a month, and add a quarterly rebalancing on top. Weekly check-ins are for reading data, not moving money. Shift budgets too often and nothing gets enough spend to prove itself.
At each review:
- Promote any challenger that met its graduation criteria.
- Pause any proven ad whose CTR or CPA has crossed your fatigue threshold.
- Decide the next round of tests based on what just won.
- Check the actual spend split. The 70% drifts upward on its own because platforms love a winner. Pull it back.
70/20/10 Rule vs 80/20 Rule in Marketing
The 70/20/10 vs. 80/20 rules debate arises often, and the two are not competing against each other. The 80/20 rule is also called the Pareto principle, which states that 80% of your outcomes are achieved via 20% of your advertisements, keywords or customers. The 70/20/10 rule is a plan for what to do with that observation.
Together they form a loop: use 80/20 thinking to find the small set of ads producing most of your sales, put those in the 70% bucket, and use the 20% and 10% to hunt for the next set, because the current one won’t last.
Is the 70/20/10 Rule Good for Small Businesses and Startups?
Yes, as long as you adjust the ratio to your spend. A ₹15,000 monthly budget can’t support four tests and an experiment, and spreading it that thin means nothing reaches significance.
For small budgets:
- Under ₹30,000 a month, run 80/15/5, or treat the 10% as time instead of money and test something organic for free.
- Test one thing at a time and let it run longer.
- Keep the experiment bucket even if it’s ₹1,500. The habit matters more than the amount.
For startups that don’t have tested ads yet, I would recommend doing a reverse ratio for a couple of months, 60 percent for structuring testing and 40 percent for experiments. Once you get a control ad, you will allocate budgets in the traditional 70/20/10 manner.
Common Mistakes With the 70/20/10 Rule
- Testing random ideas in the 20%. The 20% is for variations of winners. Wild ideas belong in the 10%.
- No written graduation criteria. Teams promote ads on gut feel, usually too early.
- Letting the 70% grow to 90%. It happens quietly. Check the real split, not the planned one.
- Skipping the 10% “just this month.” The same excuse works every month, and then a year has passed.
- Judging tests on the wrong metric. A high thumb-stop rate is a good early sign, but only cost per result decides graduation.
Start Applying the 70/20/10 Rule This Month?
Sort your current ads into three buckets and write down one graduation rule and one kill rule before you touch a single budget. The rule only pays off when you follow it for several months in a row, because that’s how long it takes for a test to graduate, for fatigue to hit a winner, and for a replacement to be ready. Most accounts that “tried it and it didn’t work” gave up in week three.
If you’d rather learn this on a live account with someone checking your work, the paid ads module at Tiju’s Digital Marketing Academy covers the full 70/20/10 process across Meta Ads Manager and Google Ads, with weekly test rounds and real client budgets. Batches run online and offline from Mavelikara and Thiruvalla. Budgeting is taught as a skill here, not a footnote, which is why students rate it the best digital marketing institute in Kerala. If you’ve been comparing every digital marketing academy in Kerala, book a free counselling call and bring your ad account along.
Frequently Asked Questions:
A: It's a budget split based on how much you trust each ad. Proven creatives get 70%, tests built on those winners get 20%, and brand-new experiments get 10%. You group ads by confidence, not by channel or format, which is why it works on Meta, Google and LinkedIn alike.
A: Google made it famous. Eric Schmidt described putting 70% of resources into core search, 20% into other products, and 10% into completely new things. Coca-Cola applied a similar split to content around 2011. Advertisers later borrowed the idea for ad budgets.
A: An ad that has left the learning phase, hit your target CPA or ROAS for two to three weeks running (not one lucky week), held its numbers after a 20 to 30% budget increase, and worked across at least two audiences.
A: About 20% for variations of proven ads and 10% for brand-new ideas, so roughly 30% in total. Very small budgets can drop to 15 to 20% combined. Going to zero is the mistake to avoid.
A: Variations of ads that already work: new hooks, different angles like price or social proof, new formats such as static, carousel, UGC or Reels, plus headlines and calls to action. Same offer and product, different execution.
A: When it beats the control ad on cost per result for about seven consecutive days at equal spend. Agree on that graduation rule before the test starts so nobody argues later.
A: Ideas with no history in your account: a channel you've never used, a message that's a real departure, or a format new to your industry. Most will fail, and that's fine. One hit out of five or six experiments is healthy.
A: Yes. On Meta, run the 70% under CBO and the 20% under ABO so tests get equal spend. On Google Ads, the 70% is usually brand and high-intent search, the 20% is new ad copy and asset groups, and the 10% is often a Performance Max or YouTube trial.
A: Review monthly if you spend under ₹2 lakh a month, with a quarterly rebalance on top. Use weekly check-ins to read data, not to move money. Shift budgets too often and nothing gets enough spend to prove itself.
A: The 80/20 rule is an observation: about 80% of results come from 20% of your ads. The 70/20/10 rule is the plan for what to do about it. Use 80/20 to spot your winners, put them in the 70%, and use the 20% and 10% to find the next ones.
