The 70:20:10 commerce portfolio

A practical model for “Now and Next”

Commerce growth rarely fails because teams don’t work hard. It fails because the system is unbalanced.

Some organisations over-invest in BAU activity, and it’s easy to see why. It’s measurable, safe, and delivers this quarter. Others however swing hard into “innovation” because it’s exciting, progressive, and because it might unlock breakthrough growth. Both patterns usually lead to the same place: plateau.

The fix isn’t another channel, platform, or dashboard. It’s a better operating model. That’s where 70:20:10 earns its place for modern commerce teams. Used properly, it becomes a portfolio approach to growth — one that protects what’s working today, systematically scales what’s emerging, and maintains a disciplined stream of experiments designed to build tomorrow’s advantage.

Where the 70:20:10 model comes from

The 70:20:10 framework is most commonly associated with Coca-Cola’s Content 2020 strategy, articulated by then-CMO Jonathan Mildenhall. Originally developed as an investment model for managing risk across content and channel innovation, it introduced a simple but powerful idea: protect proven performance, systematically build on what shows promise, and maintain a controlled pipeline of experimentation.

While its roots sit in marketing, the logic has endured precisely because it maps so naturally to how organisations should manage growth under uncertainty. As commerce has fragmented across retail media networks, platforms, formats, and markets, the need for this kind of portfolio discipline has only increased. What’s changed isn’t the principle. It’s the environments it now governs.

What 70:20:10 means in a commerce context

At its core, 70:20:10 is an investment split by risk and maturity:

• 70% — Core / Proven
  Activity with repeatable, predictable returns. Not glamorous, but compounding.
• 20% — Scale / Promising
  Activity that’s already working in pockets and needs structured scaling.
• 10% — Explore / Experimental
  Controlled tests with capped downside and asymmetric upside.

The power of the model isn’t the percentages. It’s the movement between them. What you learn in the 10% should graduate into the 20%. What proves itself in the 20% should earn its place in the 70% engine. Without that movement, 70:20:10 becomes a budgeting exercise — not a growth system.

Why commerce teams need this now

Commerce has never been noisier or more complex:

• Retail media is fragmenting across networks, formats, and buying models
• Measurement remains messy, with RoAS comfort metrics often masking true incrementality
• Margin pressure means “growth” that isn’t profitable is a trap
• Availability, fulfilment, and content fundamentals still decide outcomes — even in the most advanced stacks

70:20:10 gives teams a way to say: “We will deliver now, without mortgaging next.”

It introduces structure in an environment where discipline increasingly outperforms activity.

What goes in each bucket

1: The 70%: fundamentals that compound

This is your predictable growth engine. It should include:

• Availability & fulfilment
  In-stock performance, delivery promise, returns handling, marketplace health
• Digital shelf basics
  Catalogue integrity, PDP hygiene, taxonomy, imagery standards, A+ / rich content where available
• Always-on retail media with maturity
  Branded search defence, proven audiences, repeatable creative formats
• Review & trust systems
  Review generation, response playbooks, Q&A discipline
• Promotional mechanics you understand
  Clear calendar discipline, guardrails on depth and frequency, defined success targets

If these aren’t executed consistently, your 10% experiments are simply pouring water into a leaky bucket.

2: The 20%: scaling what’s already showing signs of life

This is where many brands under-invest — and where momentum is often lost. The 20% should focus on:

• Incrementality-led optimisation
  Holdouts, matched markets, and promo-media interaction testing
• Expanding proven plays
  Applying winners across retailers, regions, categories, or formats
• Creative iteration loops
  Not “new campaigns”, but systematic variation tied directly to conversion drivers
• Data integration
  Improving the quality of signals used to plan and optimise — profit, stock, contribution margin, not just top-line RoAS

This is where “next” becomes real — through replication and operationalisation.

3: The 10%: disciplined experiments with asymmetric upside

The 10% is not random innovation. It is controlled learning. Examples include:

• Piloting new retail media networks or ad products with clear test design
• Off-site retail media or CTV activated against retailer audiences where possible
• Creator or affiliate pilots tied to measurable commerce outcomes
• Emerging formats such as shoppable video or interactive PDP modules (retailer-dependent)
• AI-enabled content operations — faster variants, better localisation, higher relevance — with clear governance

The goal of the 10% is not scale. It is evidence.

The rules that make it work

A portfolio without governance fails quickly. A simple system keeps 70:20:10 effective.

Define success metrics by bucket

• 70%: efficiency and stability
  (profit-aware ROAS, conversion, availability, share)
• 20%: incremental lift and repeatability
• 10%: learning velocity and evidence of potential

Set graduation criteria

• 10 → 20
  When you see repeatable lift across at least two comparable contexts
  (e.g. two retailers, regions, or audience groups)
• 20 → 70
  When the activity is operationally repeatable and materially contributes to profitable growth

Run a cadence

• Monthly test read-outs (10%)
• Quarterly portfolio rebalances (70 / 20 / 10)
• Annual reset based on business objectives and category dynamics

A simple example: how ideas graduate

● 10%
Pilot a new retail media network with a clear hypothesis (incremental new-to-brand reach), using a limited budget and a holdout approach.
● 20%
Replicate in a second retailer, refine audience and creative, and prove performance at higher spend without efficiency collapse.
● 70%
Integrate into always-on planning with standard reporting, creative production cycles, and stock-informed pacing.

Closing thought

“Now and Next” isn’t a slogan. It’s a portfolio. If your plan doesn’t explicitly protect the core, scale the promising, and fund disciplined experimentation, you don’t have a growth strategy — you have a collection of tactics.

If you want a starting point, build a one-page 70:20:10 plan:

1. Your three biggest 70% levers
2. Your two most scalable 20% opportunities
3. Your single most valuable 10% experiment

That alone will force better decisions.

Sources and further reading:

• Coca-Cola Content 2020 presentation (Jonathan Mildenhall):
  https://www.slideshare.net/slideshow/coco-cola-content-2020/36588402
• Jonathan Mildenhall on “liquid and linked” marketing (Q&A):
  https://www.bmsg.org/resources/eye-on-marketers/qa-mildenhall-on-cokes-liquid-marketing/
• The Media Leader (Millward Brown): The 70:20:10 rule as a portfolio approach to innovation:
  https://uk.themedialeader.com/it-works-for-coca-cola-and-google-and-it-can-work-for-you-too-the-70-20-10-rule/
• Smart Insights: Using the 70:20:10 rule in marketing planning:
  https://www.smartinsights.com/marketing-planning/marketing-models/using-the-702010-rule-in-marketing/