First off, a very happy 2026 to all of you! We're kicking off this year's first edition a little differently. I scan through and read a large volume of articles every week, and have decided to run an analysis on a large sample of articles from 2025. I've managed to collect a dataset of close to 3,000 articles that went live between January - December 2025 and have looked at the various topics, themes and discussions that have shaped the year.

Before we get into the thick of it, I want to provide some pretext in the shape of two tables I've put together:

First, the table below visualises the segmentation of various topics and volume share across these articles, as well as what the topic represents in terms of coverage.

The second table provides us with an idea of the level of editorial quality (according to AI) for each outlet, as well as the volume and what the outlet tends to focus on in terms of topics.

Utilising various AI tools, I've started to review the dataset covering 1,000+ different outlets and close to 3,000 articles and have created five key topics/themes that have dominated online media conversations in 2025. I've specifically omitted Unilever as a standalone theme, and have tied it to the existing themes - and have steered away from event-based topics such as the Super Bowl, Cannes, etc.

1) Professionalisation of influencer marketing

Our dataset shows that 2025 marks a clear shift in how influencer marketing is discussed: coverage moves away from why influencer marketing works and toward how it is structured, measured, and governed. Influencer marketing is no longer framed as an experimental, creativity-led channel, but as a business discipline that must operate consistently and at scale.

Coverage increasingly treats influencer marketing like other core marketing functions: something that requires performance accountability, governance, and integration into existing systems.

The Unilever announcement of its increased influencer marketing efforts early in the year reinforces the idea that influencer marketing is no longer being tested at the margins, but rather becoming institutionalised at scale.

Signals from our dataset

Professionalisation in the dataset is driven by structural, not organisational, signals:

2) Influencer marketing as a business function

We're seeing that more and more articles are describing influencer marketing as something that:

Influencer marketing is no longer framed as a channel you "activate," but as a proper business function.

It surfaces in 2025 because brands reached a point where influencer marketing touches too many parts of the organisation (brand, growth, commerce, legal, platforms) to not be taken seriously.

Signals from our dataset

3) AI becomes a scaling layer

AI shows up frequently in our dataset, but rarely as a standalone narrative about industry transformation. Instead, it appears embedded inside other stories: tooling, workflows, creator discovery, content production, reporting, and risk. That's the signal.

AI isn't necessarily framed as "the future of the creator economy" but instead positioned as the solution for scaling, as the operating costs of scaling through simply increasing headcount is becoming less viable and efficient.

Signals from our dataset

4) Trust and authenticity as a baseline requirement

Trust and authenticity are not necessarily dominant topics by volume, but they are structurally present when scale and risk are discussed. Trust surfaces most clearly at points of friction: regulation, enforcement, brand safety incidents, platform intervention, and debates around AI within influencer marketing infrastructure. Trust is no longer discussed as a way to improve influencer programmes, but as a minimum condition for operating at scale - a baseline requirement: if it is present, nothing special happens; if it is missing, your programme may not function properly.

Signals from our dataset

5) Continued consolidation

Funding rounds, acquisitions, agency roll-ups, and platform buy-and-build stories are consistent in 2025.

The dataset shows that capital is moving into the creator economy, not to discover what's possible, but to scale what already works. Agencies are acquired to increase capability, platforms are buying tools to control workflows, and investors back infrastructure-like businesses rather than creator-led experiments.

Signals from our dataset

Bonus: what to expect in 2026

This has been a really interesting exercise, and a great way to use AI tools & Python to play around with large data sets. I've made a conscious effort to avoid making predictions, but I've put the dataset through a few AI models and asked it to identify what we can expect in 2026 in terms of topics.

One thing that I agree with - is that we're likely to see more and more conversations around infrastructure and ecosystems, as it matches the scaling and AI conversations that have become mainstream in 2025.

I've asked AI to come up with a few topics we may see in 2026: