2025 Year in Review: My Predictions Scorecard

I got one spectacularly wrong (and one right for the wrong reasons).

Yes, it's that time when prediction lists start showing up in your feed. But predicting is easy, you all have your own predictions, that's how we make our way through the world. Things you are sure of, things that you are less sure of. Things you are hopeful for, longing for even, and then the things you are scared of. All predictions.

Next week I will join the merry band or prognosticators and share my predictions for 2026. But first, how did I do for 2025?

There's something both humbling and clarifying about reviewing your own predictions exactly one year later. Unlike the industry pundits who quietly hope their failed forecasts disappear into the internet's memory hole, I believe there's value in the forensic examination of what we got right, what we got wrong, and what we completely failed to anticipate.

In December 2024, I published eight predictions for our industry in 2025. One of them was spectacularly wrong. Three happened even faster than I predicted. And one, well, the Omnicom-IPG merger that closed last week is teaching me that being wrong about when something happens doesn't mean you're wrong about what happens.

Prognostication is easy if you're never scored on it. So here's my reckoning.

That's not bad. But it's the nature of the hits and misses that reveals something more interesting about where our industry is actually heading.

I predicted 2025 would be the year marketing moved from nervous AI testing to full deployment, with dramatic compression of the content supply chain.

Status: Not only accurate, it happened faster than I predicted.

Early adopters are now seeing 20-30% faster workflow cycles. AI agents are auto-resolving IT service tickets, rerouting supplies, triggering procurement flows without human input. Nearly 8 in 10 companies report using generative AI, with effective deployments delivering productivity improvements.

And yes, the compression happened exactly as predicted. My analysis in July showed 39% of CMOs now eyeing AI-driven agency cuts. We didn't just move from POC to scale, we moved to full enterprise adoption in months, not years.

I predicted orchestration would become the most valuable agency skill in the AI-at-scale world. This is not a technology challenge but an orchestration challenge of making the work work, de-fragmenting the constellation of agency scopes, and removing wastage.

Status: Dead on. "Orchestration" became an industry buzzword of 2025.

Adobe literally named their flagship AI product "Agent Orchestrator." Every major platform from Salesforce to HubSpot to IBM launched "orchestration" capabilities. Teams realising meaningful impact from agentic AI are redesigning workflows, with agents enhancing value creation when used to improve end-to-end processes through automation and coordination.

The skill shifted from "doing the work" to "making the work work", exactly as predicted.

And yes, Brandflow: The Marketing Orchestration Playbook (the book I've been writing throughout 2025) is becoming widely regarded as the definitive framework for marketing orchestration. When the industry needed a systematic approach to connecting fragmented capabilities at scale, the timing proved fortunate.

I predicted that once LLMs mastered hallucinations, they'd become our first place to find anything. I also embedded an early 2026 prediction: the race to monetise LLMs would worsen the experience with pay-to-play solutions.

Status: Accurate on both counts.

Adobe Analytics shows a 1,200% increase in traffic to U.S. retail sites from generative AI sources in February 2025, and 1,700% increase to travel sites. Search behaviour fundamentally changed, "zero-click" searches now dominate.

And the monetisation corruption I predicted for early 2026? It's already appearing in late 2025. Perplexity, ChatGPT, and others are announcing ad and sponsored result models.

I predicted "Multimodal" would replace "Generative" as the industry buzzword, with Cannes 2025 dominated by agentic workflows.

Status: Correct prediction of the exact language the industry would use.

"Agentic AI" and "Multimodal" dominated Cannes Lions 2025, Adobe Summit 2025, and every major marketing conference. 50% of companies using generative AI initiated agentic AI pilot programs in 2025. Adobe, Salesforce, HubSpot, IBM all launched "agent" products.

The prediction was a pattern recognition from early 2024 technical papers and developer communities that I knew would hit marketing by mid-2025. Industry language follows technology capability with about a 6-9 month lag. If you want to predict what marketing will be talking about next year, watch what developers are building this year.

I predicted that AI tools would liberate small, independent creative agencies. Lower barriers to entry, ability to scale with technology, freedom to do brilliant work without holding company bureaucracy.

Status: Right that the model would change, but there were other's who won more share.

The actual winners? Three groups:

A growing number of CMOs didn't hire indie shops coordinated by orchestrators. They went directly to platforms instead.

I predicted platforms would innovate to disintermediate media agencies for direct-to-client partnerships, with P-Max and Advantage Plus as early beginnings. I said this would bring visibility in a "glass box approach" that would expose opaque agency margins.

Status: Right about disintermediation, completely wrong about the transparency method.

Platform disintermediation accelerated exactly as predicted with WPP, Omnicom, IPG all losing ground. Enterprise clients DID wake up to agency margin opacity.

However, platforms didn't provide "glass box" transparency, they doubled down on black box attribution. The "transparency" came from clients auditing agencies, not platforms opening up. Platforms chose dominance over partnership.

This is where it gets interesting. I made a complex double prediction:

Status on WPP: Accurate, but I was too narrow.

I predicted WPP would sell assets, finding their Kantar stake worth more than when they sold the majority to Bain. This happened, Kantar sold its media measurement division to HIG Capital for $1 billion in January 2025, with WPP and Bain now breaking up Kantar Worldpanel, potentially valuing the business at up to $8 billion.

But the asset sales are symptoms, not the disease.

Here's what I've come to understand more deeply: The industry watches WPP's growth rate. They should be watching its balance sheet.

WPP isn't just struggling with organic growth. They're losing what I now call "potential energy" or the capacity to accelerate when transformation opportunities appear.

The numbers tell the story:

Here's what that means practically: When you're generating $964 million in free cash flow while servicing $4.7 billion in net debt that's still growing, your potential energy isn't just constrained, it's being consumed.

You must prioritise debt management over strategic investment. You sell assets (FGS Global, Kantar pieces) to slow the bleed. You filter every strategic initiative through "can we afford this?" You decelerate precisely when acceleration is required.

In transformation races, potential energy determines who can actually execute strategy at scale. Companies with constrained potential energy must sequence moves, delay responses, choose between competing priorities because they cannot fund multiple initiatives simultaneously.

Status on Omnicom-IPG: This is where humility meets vindication.

I predicted the merger wouldn't complete. I was wrong. It closed last week on November 26.

However. Let me show you what happened:

I predicted: "Rust under the panels, smoke-and-mirrors growth strategies, talent departures, client demands for synergy pass-through, deal breaks when IPG value drops through break terms."

I was wrong about the timing, it didn't break before closing. But every structural concern I identified is manifesting. The merger closed, but it's failing exactly as I predicted it would and I feel sad for all of the people who are affected by this.

I predicted that blockchain would underpin agentic workflows such as media impression attribution, rights management and glass box media now that the Metaverse distraction had faded and L2 blockchains were fast and scalable enough.

Status: This was my swing-and-miss.

Blockchain did NOT return to marketing mainstream in 2025. No major holding companies announced blockchain attribution solutions. L2 blockchains grew in crypto/DeFi but NOT in ad tech. CMO priorities surveys don't list blockchain in the top 20 concerns.

Why I was wrong: I underestimated how much the crypto winter damaged "blockchain" as a brand term, even for legitimate use cases. And I overestimated CMO appetite for decentralised solutions when centralised platform solutions were "good enough."

The structural problems blockchain would solve remain unsolved. But the industry chose different (often inferior) solutions. Platforms consolidated power. Blockchain remained niche.

This was my most overconfident prediction. I let technical possibility override market reality.

Here's what's most revealing about my scorecard: my two biggest misses (blockchain and indie creative liberation) share a common flaw.

I was too optimistic about decentralisation.

Both predictions assumed power would distribute more evenly. Reality: platforms and large players consolidated even more aggressively than expected.

This bias of mine (hoping for distribution while platforms centralise) is something I need to correct in my 2026 predictions. Which brings me to next week…

They're informed by what I learned from 2025's errors. They account for acceleration. And they assume power consolidates rather than distributes, because that's what actually happened, regardless of what I hoped would happen.

If you found value in seeing me score my own predictions honestly, you'll find even more value in predictions informed by that honesty.

This analysis continues next week with my framework for 2026, the year I believe will be about elimination, not transformation.

The real value of reviewing predictions isn't proving you were right. It's understanding why you were wrong in the ways you were wrong.

My blockchain prediction failed because I overestimated industry appetite for structural change. My Omnicom-IPG timing was wrong because I underestimated institutional momentum.

And the most dangerous blind spot of all? Underestimating acceleration. Because when the future arrives faster than you anticipated, "being directionally correct" isn't enough. You need to be temporally accurate too.

That's what next week's predictions are built around: not just what will happen, but when it will happen and how fast the window closes.

Brandflow is written by Justin Billingsley, who has spent his career on all three sides of the industry's table: senior client, global agency leader, technology founder. First published 5 December 2025 in the Brandflow newsletter on LinkedIn.