The Creative Intelligence Maturity Model: Fixing the “Return Path” From Measurement to Better Ads

A creative intelligence maturity model assesses how reliably an organization turns creative performance measurement into better briefs, better assets, and better outcomes. Winterberry Group's September 2026 research found that adoption is near-universal, but only 6% of organizations use creative intelligence consistently at every stage they have adopted, and just 1% connect the data through to changing a creative decision. The gap is rarely tools. It is a broken “return path” between insight and action. Leaders separate themselves through governance: a written testing roadmap, a protected experimentation budget, a shared taxonomy, and clear decision rights when creative and media disagree. Creative analytics and synthetic audiences each address part of that gap when they are tied to a decision process.

Key takeaways

- Adoption is close to universal, but only 6% of organizations use creative intelligence consistently at every stage they have adopted, and just 1% connect the data to actual creative decisions (Winterberry Group, 2026).
- The common failure is a broken return path: roughly one brand in seven has a closed loop from measurement back to the next creative decision, on self-reported numbers.
- Leaders are distinguished by governance, not more tools: a written testing roadmap, a protected experimentation budget, a shared taxonomy, and clear decision rights.
- Creative analytics explains why assets perform; synthetic audiences test likely response before spend. Both only pay off when outputs reach someone with authority to change the work.

Introduction

Most marketing teams can now measure creative performance. Far fewer can reliably use that measurement to change the next brief and the next set of assets. Winterberry Group calls the missing piece the “return path”: the route that carries measurement data back into a changed creative decision. This post uses Winterberry's five-band maturity curve to show where that path breaks, what leaders do differently, and where creative analytics and synthetic audiences fit.

The real gap: measurement exists, action doesn't

Creative intelligence, as Winterberry defines it, is the ability to collect, structure, and analyze creative decisions against performance data, then apply what that shows, in the context of audience and media, to keep improving assets. It sits alongside media intelligence (where and when a message appears) and audience intelligence (who it is for). Media and audience have had established measurement standards for years. Creative has not.

The research suggests the analytics are not what is missing. Among the 200 US and UK brand marketers surveyed, 82% get performance data on a live creative within a day, 53% attribute performance to specific creative components across all channels, and 59% consistently separate creative lift from media and targeting lift. What they have not built is the mechanism that turns a measurement output into a changed asset or a changed brief.

Common symptoms of a broken return path include:

- Speed without action: of the brands receiving performance data within a day, 35% do not adjust creative variants in-flight at all.
- Reports that stop at the report: of the brands using creative intelligence in post-campaign analysis, 30% do not carry it into the next brief.
- Insights trapped in one team: where an insight can be actioned inside a single team, it mostly is. Where it has to cross into creative production or audience planning, it mostly stops.

It also explains why so much of the value shows up as media reallocation. A buyer can move budget alone. Recutting an asset takes several people agreeing to act.

The five-band maturity curve: how to self-diagnose

Winterberry scored each respondent on practice rather than opinion: which lifecycle stages they use, what data is connected, which tools and teams are involved, and whether creative's contribution can be separated from the media buy. That produced five bands:

- Lagging (9%): exploring, with no solution or concrete tests in place. Creative intelligence is a topic rather than a place, and no one owns it.
- Emerging (21%): limited tests in specific campaigns. Creative intelligence appears inside media platforms, marketing operations gets involved, and a written testing roadmap appears.
- Progressing (34%): the largest band. Solutions are in place and used consistently, and creative intelligence starts touching live, in-flight decisions.
- Established (26%): used by cross-functional teams across pre-flight and in-flight. Measurement becomes rigorous and governance artifacts are in place.
- Leading (11%): embedded across the organization. The report notes that nothing new is added at this level. What changes is that data is fully connected, creative contribution can be measured, and creative and media run one workflow.

A useful way to read the curve: below Established, the challenge is acquiring capabilities. Above it, the challenge is connecting them and proving creative's contribution. Leaders are not doing more. They are doing the same things more reliably.

To locate your own team, ask two questions:

- Consistency: where you have adopted a stage (brief, pre-flight, in-flight, post-campaign), do you use it every time, or only when time allows? For example, 75% of brands have applied creative intelligence at the brief stage, but only 40% of them do so every time.
- Connection: do insights predictably change what gets made, how it launches, and what gets tested next?

Beware the partial adoption trap. A team can look advanced because it runs analyses or tests while remaining effectively lagging, because results do not consistently change briefs, assets, or launch plans.

What leaders do differently: governance and operating discipline

Winterberry's central finding is that what separates leaders from laggards is not the technology they own but the discipline they wrap around it. All of the largest differences concern process and governance.

Checklist: a written testing roadmap and a protected experimentation budget

- Document a roadmap that prioritizes a small set of repeatable experiments. Every Leading organization has one, against 12% of Lagging organizations.
- Protect a dedicated experimentation budget so testing is not deprioritized when performance pressure rises. 96% of Leading organizations have one, against 18% of Lagging.
- Define how learnings roll into the next brief and production plan. Brands with both a roadmap and a budget are the ones most likely to have reached Established. Either one alone tends to stall at Emerging.

Checklist: an enforced taxonomy and naming convention

- Standardize naming for creative assets and versions, and apply taxonomy fields the same way every time.
- Add a pre-launch QA step to validate tags and names, with a named owner.
- Treat it as the gate on automation. 40% of the market lacks a taxonomy, a testing roadmap, or both, and brands with both unify their data 70% of the time, against 23% for those without.

Checklist: fast feedback and explicit decision rights

- Set a review cadence aimed at action, not just reporting.
- Decide in advance who has the final say when creative and media interpretations diverge. Among Leading organizations, 96% have a fully integrated cross-functional team.
- Choose any deliberate operating model (centralized, embedded, or hybrid). Which one matters far less than having one: no Established or Leading organization runs creative intelligence ad hoc.

Why creative analytics matters

Creative analytics connects the decisions made inside an asset, such as the message, format, hook, and structure, to the performance those decisions produce. That makes three things possible:

- Learning that is reusable. Element-level attribution turns “this ad worked” into a finding a team can apply to the next brief.
- Proof of creative's contribution. Every organization in the Leading band consistently isolates creative lift from media and targeting lift. Without that separation, creative stays a fixed cost rather than a measurable performance driver.
- Compounding over time. Analytics is only as reusable as the tagging underneath it, which is why taxonomy comes first.

Learn more about creative analytics.

Why synthetic audiences matter

Synthetic audiences simulate how a target audience may respond to creative so teams can test ideas before committing media spend. They are already mainstream: AI and panel-based testing, which includes synthetic data, is used by 58% of respondents, ahead of A/B testing, dynamic creative optimization, and demand-side platforms. Use rises from 12% of Lagging organizations to 90% of Established ones. Budgets are moving the same way. Winterberry's spend models show traditional research declining by about $3 billion in the US from 2025 to 2027, while data- and tech-driven research grows by about $10 billion.

The practical benefits are speed and cost, earlier testing (moving testing in front of the spend is a defining step from Progressing to Established), and a way to bring audience response into creative decisions earlier. That last point matters because combining creative and audience data is the weakest handoff in the study.

The caveats are real. Winterberry notes the category has not resolved what synthetic audiences are for, and one interviewee stressed they do not replace real human response or in-market results. Pre-flight testing is the least adopted lifecycle stage at 59%, and several interviewees described it being used as a “fig leaf” to defend a creative decision to leadership. The strongest use is the opposite: shaping the brief and narrowing options before money is spent, alongside in-market learning.

Learn more about synthetic audiences.

Fix the broken handoffs: creative plus audience, and isolating creative lift

The handoff most often broken is between creative insight and audience data. 55% of brands using creative intelligence in creative development, and 52% of those using it to inform media planning, do not integrate audience data. Winterberry calls it the least reliable handoff in the study and the last capability to scale at the top of the curve.

When creative and audience data live in separate workflows, teams cannot tell whether an asset underperformed because of its message or because it reached the wrong audience context. Integration does not have to be elaborate. It can start with a shared review that looks at creative performance with enough audience context to decide what to change.

A shared workflow keeps responsibilities clear:

- Creative: owns hypotheses about messaging and structure and translates learnings into the next brief.
- Media: owns delivery context, targeting, and the feasibility of tests.
- Measurement: owns comparability rules and avoids changes that confound interpretation.
- Operations: owns taxonomy enforcement and the handoff rhythm.

Underneath this sits an organizational root cause. Ten of the sixteen expert interviews identified the separation of media and creative, a legacy of agencies splitting roughly a decade ago, as the binding constraint on adoption and maturity.

Practical guidance: make testing change the work, not just validate it

- Start with a learning agenda. Pick a few questions whose answers would change creative choices, rather than testing to confirm a decision already made.
- Run fewer, repeatable experiments. Design tests to be comparable across campaigns, and apply them consistently across relevant spend.
- Confirm inputs, comparability, and decisions. Check that taxonomy is applied before launch, hold key variables steady when attributing differences to creative, and document what changed in the brief or asset as a result.

When these are in place, measurement becomes a decision system. The output is not a slide but a record of what changed and what gets tested next.

Evaluating tools: ask how outputs reach a decision

94% of brands already run creative intelligence inside systems they own, such as media platforms, measurement tools, or creative suites. Among Leading organizations it most often sits inside channel measurement and marketing mix modeling (96%) and media platforms (91%), while standalone platforms fall from 48% at Established to 32% at Leading. Providers such as AdSkate, which offers creative analytics and synthetic audience tools, fit into that picture as inputs to existing workflows rather than destinations of their own.

Before adding or expanding a tool, ask:

- Does the output reach someone with the authority to change the asset?
- Is it tied to the brief, or only to a post-campaign report?
- Does it connect to audience data, not just creative data?
- Does it work with your taxonomy and existing stack?
- Is there a written plan and protected budget behind the testing?

If several answers are no, a better tool probably will not fix it. The handoff will.

Frequently asked questions

Q: What is a creative intelligence maturity model?
A: A creative intelligence maturity model is a diagnostic framework for assessing how consistently a team turns creative performance measurement into action. Winterberry Group's version scores organizations into five bands (Lagging, Emerging, Progressing, Established, Leading) based on what they actually do across the campaign lifecycle, what data is connected, and whether creative's contribution can be separated from media's.

Q: Why do creative performance insights fail to change ads and briefs?
A: They fail when the return path is broken. Winterberry's research found the issue is organizational rather than analytical: insights stall when they have to cross from one team to another, such as into creative production or audience planning. Inconsistent taxonomy, slow handoffs, and unclear decision rights make it worse.

Q: What separates leading organizations from lagging ones?
A: Process and governance rather than technology. Leaders have a written testing roadmap, a protected experimentation budget, a shared taxonomy, an integrated cross-functional workflow, and a settled answer to who decides when creative and media disagree.

Q: What standards should a creative taxonomy and naming convention include?
A: At a minimum: a defined naming convention for assets and versions, required taxonomy fields applied the same way every time, and an owner plus a pre-launch QA step to ensure compliance. The goal is for results to be identifiable and comparable across campaigns.

Q: What are synthetic audiences used for in creative testing?
A: Synthetic audiences simulate audience response so teams can test creative before or alongside live spend. They are best used to shape the brief and narrow options early, as a complement to in-market learning rather than a replacement for it. Winterberry notes the category has not yet settled on a clear role.

Q: How do you separate creative lift from media or targeting lift?
A: Keep key variables stable when attributing performance differences to creative, and review creative and audience context together. The goal is to determine whether a shift is best explained by the asset, the audience, or delivery conditions, then document the decision that follows. Every organization in Winterberry's Leading band does this consistently.

Conclusion

Creative intelligence maturity comes down to reliability: whether an organization can repeatedly turn signals into changed work. Start by mapping your return path and finding where it breaks, especially at the handoffs between creative, media, measurement, and operations. Then prioritize the governance basics, a focused testing roadmap with a protected budget, and decision rights that keep learning loops fast.

Creative analytics and synthetic audiences can each help close the gap, as long as their outputs reach someone who can act on them. AdSkate offers creative analytics and synthetic audience tools. [Add link]

Sources

Winterberry Group, “Creative Intelligence: The Market Maturity Curve, Understanding the Criteria That Defines Leaders and Laggards,” September 2026. Survey of 200 US and UK brand marketers plus 20 supply-side expert interviews. https://winterberrygroup.com/creative-intelligence-the-market-maturity-curve

Note: the research was sponsored by vendors in the creative intelligence space, and the brand-side survey data is self-reported.