The 70/20/10 portfolio for modern marketing teams.

A practical framework for balancing proven marketing, promising bets, and weird experiments — without wasting budget.

Your CFO is not wrong to be nervous.

Marketing teams love to talk about experimentation, innovation, and trying new things. Finance teams usually hear something different.

They hear: unclear spend, uncertain return, loose accountability, and another reason the budget might get messy.

Fair.

But here is the other side of the argument: a marketing team that only funds what is already proven is not being disciplined. It is quietly becoming fragile.

Channels mature. Costs rise. Creative fatigues. Competitors copy. Algorithms change. Buyers adapt. The thing that worked last quarter may still work this quarter, but it rarely works forever without pressure, improvement, and new thinking.

That is why modern marketing teams need a portfolio.

Not random experimentation. Not shiny-object chasing. Not “let’s test AI because everyone is talking about AI.”

A real portfolio.

One that protects the core business, explores what might work next, and leaves room for ideas that sound strange before they become obvious.

The simplest way to think about it is this:

70% proven.
20% promising.
10% weird.

The Problem With Running Marketing at 100% Proven

The safest marketing plan is often the riskiest one long term.

When 100% of the team’s effort goes into proven work, the business gets predictability. That feels good. It makes planning easier. It makes reporting cleaner. It makes the dashboard easier to defend.

But it also creates a problem.

If every dollar, hour, and idea goes into what already works, the team stops learning. The system becomes efficient at yesterday.

That might be fine for a while. Until the channel gets more expensive. Until a competitor copies the campaign. Until the creative fatigues. Until buyer behaviour shifts. Until the algorithm changes. Until the team realizes it has no next move.

Marketing effectiveness research has long made the case that not all valuable marketing works on the same timeline. Binet and Field’s work, often summarized through the 60/40 principle, argues for balancing long-term brand building and short-term sales activation. The IPA has noted that the original “sweet spot” was roughly 60% brand building and 40% activation, with later work showing the balance may shift by context.

The point is not that every marketing budget needs the same split.

The point is that marketing leaders need to manage time horizons.

Some work pays back now. Some work builds future demand. Some work teaches you what the next advantage might be.

If you only fund the first category, you do not have a strategy. You have a production schedule.

The 70/20/10 Portfolio for Modern Marketing Teams

The 70/20/10 idea is not new.

Harvard Business Review’s “Managing Your Innovation Portfolio” describes a version of the model where high-performing companies tend to allocate about 70% of innovation resources to core offerings, 20% to adjacent opportunities, and 10% to transformational initiatives.

For modern marketing teams, I would translate it this way:

70% proven: the work that reliably drives the business today.

20% promising: the work showing early signs of traction.

10% weird: the work that is uncertain, original, and hard to justify with yesterday’s data.

This is not a rigid formula. McKinsey has warned that companies can rely too heavily on rules of thumb like 70/20/10 if they are not connected to real innovation aspirations and strategic priorities.

That is the right caution.

The ratio matters less than the discipline.

The goal is not to worship 70/20/10. The goal is to stop letting the urgent consume everything original.

70% Proven: Protect the Engine

The proven 70% is the work that keeps the business moving.

This is where teams should spend most of their energy because it funds the system. It is not boring. It is necessary.

For most marketing teams, this might include:

• Core paid media campaigns
• SEO maintenance and optimization
• Lifecycle and nurture programs
• High-performing content formats
• Conversion rate improvements
• Reporting and measurement
• Sales enablement that already works
• Landing pages and web improvements tied to active demand

The goal of the 70% is reliability.

You are not trying to reinvent the system every week. You are trying to make the system sharper, cleaner, and more efficient.

How to Measure the Proven 70%

The scorecard should be clear:

• Revenue influenced
• Pipeline generated
• Cost per qualified lead
• CAC
• ROAS
• Conversion rate
• Lead quality
• Sales feedback
• Retention or repeat purchase

This is the bucket your CFO will understand fastest.

It has known inputs, known outputs, and enough history to make decisions with confidence.

But this bucket has a ceiling.

If all you do is optimize the proven 70%, eventually you will run out of meaningful improvements. You will squeeze the system instead of growing it.

That is why the other 30% matters.

20% Promising: Build the Next Core

The promising 20% is where early evidence exists, but the work is not fully scaled yet.

This is the bridge between proven and weird.

It might be:

• A new channel showing early traction
• A different creative angle
• A new content series
• A new audience segment
• A new nurture path
• An emerging SEO or AEO strategy
• An AI-assisted research workflow
• A better sales enablement system
• A new partnership or event format

The goal is not immediate perfection. The goal is progression.

This bucket should answer one question:

Could this become part of the core?

How to Measure the Promising 20%

You may not judge the 20% only by immediate return. You look for signals:

• Are people engaging?
• Is the cost improving?
• Is the audience right?
• Is sales hearing better conversations?
• Is the idea repeatable?
• Is the workflow making the team faster?
• Is the quality improving with each iteration?

This is where a lot of marketing teams get the process wrong.

They either scale too early because something looks exciting, or kill it too early because it does not perform like the proven bucket.

Promising work needs a different scorecard.

It needs enough room to learn, but enough discipline to prove it deserves more.

10% Weird: Buy the Right to Be Early

The weird 10% is the easiest bucket to cut.

It is also the bucket that may matter most when everything starts to look the same.

This is where teams test ideas that do not yet fit the dashboard:

• A strange creative format
• A contrarian point of view
• A founder-led content series
• A community experiment
• A live event concept
• A new AI workflow
• Original research
• A bold landing page structure
• A partnership no one else in the category is trying
• A “this might be stupid, but…” idea from someone on the team

The weird 10% is not random.

It is intentionally uncertain.

That distinction matters. You are not giving the team permission to waste time. You are giving the team permission to explore ideas that are too early to prove.

This is how originality survives inside a performance-driven marketing team.

The CFO should not hate the weird 10%. The CFO should hate marketing teams pretending everything is proven when it is not.

The weird 10% is a bounded risk.

It says: we are going to protect the core business, but we are also going to buy the right to be early.

Most weird ideas will not become big ideas. That is fine. The point is not a perfect hit rate. The point is learning velocity.

How to Measure the Weird 10%

You are looking for:

• New insight
• Qualitative response
• Earned attention
• Internal capability gained
• Strategic surprise
• Customer language
• Future potential

The weird 10% is not where marketing goes to play.

It is where the company learns before the market forces it to.

Different Buckets Need Different Scorecards

The biggest mistake with portfolio marketing is judging every idea by the same dashboard.

If you judge the weird 10% by immediate ROAS, you will kill it before it teaches you anything.

If you judge the proven 70% by how original it feels, you may underfund the work that pays the bills.

If you judge the promising 20% too loosely, you create a pile of interesting experiments that never become useful.

Each bucket needs a different question:

• 70% proven: Is it performing?
• 20% promising: Is it progressing?
• 10% weird: Did we learn something useful?

This is how you keep experimentation commercially grounded.

Innovation portfolio research makes a similar point: different types of innovation have different risk profiles, time horizons, and management needs. Harvard Business Review’s model separates core, adjacent, and transformational innovation because each type of work behaves differently and should not be managed as if it carries the same risk or expected return.

Marketing should work the same way.

A campaign optimization should not be evaluated like a category-shaping idea.

A strange creative test should not be evaluated like a mature paid search campaign.

A new AI workflow should not be judged only by immediate output. It should also be judged by whether it improves speed, quality, judgment, or team capability.

The portfolio only works if the metrics match the maturity of the idea.

AI Makes the Portfolio More Important, Not Less

AI will make marketing teams faster.

It will also make average work more common.

That is why the portfolio matters more now.

Google has said AI-generated content is not inherently against its guidelines, but the focus remains on whether the content is helpful, reliable, and people-first. Google's broader helpful content guidance also emphasizes content created to benefit people, not content created mainly to attract search rankings.

That is the tension for modern marketers.

AI can help the 70% by improving production, reporting, research, optimization, and analysis.

It can help the 20% by making it cheaper to test more variations, angles, workflows, and ideas.

But the 10% still needs human judgment.

AI can help you move faster. It cannot decide what is original in your category. It cannot tell you what your audience is tired of unless someone has the taste and context to see it. It cannot make a safe idea brave. It cannot turn a generic position into a point of view.

AI makes execution cheaper.

It does not make originality automatic.

That is why the weird 10% needs protection. Without it, teams will use AI to produce more of what everyone else is producing.

Faster sameness is still sameness.

How to Protect the Portfolio Without Losing Control

The 70/20/10 portfolio only works if it has operating discipline.

It cannot live as a slide in a strategy deck. It has to show up in planning, resourcing, reviews, and team conversations.

A Simple Operating Model

Define the buckets before planning starts. Decide which work is proven, promising, or weird before the quarter begins.

Allocate time, budget, or creative capacity. The portfolio does not need to be exact, but it does need to be protected.

Write a learning goal for every 20% and 10% experiment. If the experiment fails, the learning should still be useful.

Start small before scaling. The 10% should not get big money too early. It should get enough room to test the idea.

Review the portfolio monthly. Look at performance, progression, and learning separately.

Move ideas between buckets. Weird can become promising. Promising can become proven. Proven can decay.

Kill ideas cleanly. An experiment that teaches you something and ends is not a failure. It is a completed learning loop.

This is where the CFO conversation changes.

You are not asking for vague innovation budget.

You are showing a system for managing risk.

The proven 70% protects the business. The promising 20% develops the next growth layer. The weird 10% creates room for originality before you desperately need it.

That is a budget conversation worth having.

Originality Needs a Budget

Marketing leaders do not need to choose between performance and originality.

They need to allocate for both.

The 70/20/10 portfolio gives teams a way to protect what works, explore what might work, and leave room for the ideas that sound strange before they become obvious.

Because the future rarely arrives with a clean business case.

It usually starts as a small signal. A customer comment. A strange creative idea. A new tool. A shift in search behaviour. A campaign that does not fit the template. A junior team member saying, “This might be dumb, but what if we tried…”

That is where the next edge often begins.

If every dollar has to prove itself before it is spent, the team will only ever fund the past.

So protect the engine.

Build the next core.

And leave room for weird.

That is how modern marketing teams stay original without setting the budget on fire.