In December 2022, we published a list of five marketing trends we expected to define 2023: authentic storytelling, video, guerrilla marketing, AI, and short-form social stories. Enough time has passed to check our own work. Two of those calls held up close to as written. Two pointed in the right direction but badly undersold what actually happened. One didn't hold up the way we framed it. Below is the original list, graded honestly, plus the one lesson each grade actually teaches about reading a trend forecast.
One caveat before the grades: this isn't a controlled study of our own campaign data from that period. It's an honest look back using what's publicly known about how the industry actually moved between 2023 and today. Where a specific number would be needed to prove a grade, we say so instead of inventing one.
KEY TAKEAWAYS
- Predictions describing a behavior shift already underway, like more short-form video, held up. Predictions naming a specific feature or platform, like Stories or guerrilla stunts, held up worse.
- Getting the direction right and getting the mechanism right are two different bets. The AI call is a clean example: right that AI would matter more, wrong about how.
- A prediction that's true most years, like "brands want to stand out," isn't really a forecast. It doesn't need a specific reason to sound plausible, which is exactly why it's hard to grade.
- Naming a platform bets on that platform surviving, not just on the behavior underneath it. That's a bigger risk than the trend itself usually gets credit for.
AUTHENTIC STORYTELLING: RIGHT INSTINCT, WRONG MECHANISM
Original line from the post: "Testimonials and stories of success and people doing good are really key. Influencers and paid spokesperson models are beginning to fade." Grade: partially right.
The instinct was correct. Audiences did move toward content that felt like it came from a real person rather than a produced ad, and testimonial and user-generated content became a standard part of how brands advertise instead of a niche tactic. The mechanism was wrong. Influencer marketing didn't fade, spending on it kept climbing through 2023 and beyond. What changed is what influencer content looks like. Brands leaned into creators who film on a phone and talk straight to camera, because that reads as authentic even when it's still a paid placement.
The lesson: when a trend is "people want something more real," don't bet on which existing channel gets replaced. Bet that the channel that survives will be the one that learns to look like the alternative.
WE GOT VIDEO MOSTLY RIGHT, BECAUSE IT WAS THE EASY CALL
Original line from the post: "expect to see an increased focus on video content across all platforms... Major brands are choosing to spend less on massive commercials and more on diverse videos across a variety of styles and platforms." Grade: accurate.
This is the easiest grade on the list. Video was already the fastest-growing content format in late 2022, so predicting more of it mostly extended a line that was already moving. Short vertical video, on Reels, TikTok, and YouTube Shorts, became the default format brands reach for first. Big single-commercial budgets kept shrinking in favor of smaller, cheaper videos built for a feed instead of a broadcast slot.
The lesson: extending a trend that's already visibly accelerating is the safest prediction available. It's worth stating, but it doesn't tell you much about the quality of your judgment.
GUERRILLA MARKETING NEVER BECAME THE TREND WE PREDICTED
Original line from the post: "Expect guerilla tactics that carry a surprise and an unscripted approach to become even more popular in 2023." Grade: missed.
Guerrilla marketing, in the flash-mob and physical-stunt sense the original post described, didn't become a defining theme of the years that followed. Marketing budgets tightened through 2023 as companies watched costs generally, and the kind of unscripted, resource-heavy stunt this prediction described needs discretionary budget that was harder to justify. The attention guerrilla marketing used to chase moved somewhere else: a single well-timed organic social post can now reach more people for less money than a physical stunt, so that's where the surprise-and-unscripted energy actually went.
The lesson: a real trend forecast needs a reason for "why now." Guerrilla marketing gets predicted most years because standing out is always appealing to a marketer, not because 2023 specifically was set up to be its moment. A prediction that would sound equally true in any other year isn't really a forecast.
THE AI CALL UNDERSOLD WHAT WAS ABOUT TO HAPPEN
Original line from the post: "AI technology allows marketers to get better insights into customer behavior... expect its use in marketing strategies to increase significantly." Grade: right direction, badly undersold the scale.
ChatGPT had launched only a few weeks before this post went live, and the original prediction had no way to know it. The AI framing here was about predictive analytics, behavior modeling, and using "bots" to test products, which was a reasonable read of where AI marketing tools stood in late 2022. What actually happened is generative AI moved into marketing workflows directly: drafting copy, generating images, building first-pass campaign briefs, and handling a growing share of customer-facing chat. The direction called correctly. The mechanism and the speed were both off by a wide margin.
The lesson: naming a direction, "AI will matter more," is a safer bet than naming the mechanism, "through better behavior prediction." A genuinely new capability can arrive and make the mechanism obsolete while leaving the direction intact.
BETTING ON STORIES WAS RISKIER THAN BETTING ON VIDEO
Original line from the post: "In 2023 expect even more brands utilizing Instagram stories... to keep customers engaged with your brand." Grade: partially missed.
The broader call, more short vertical video, was right, and it's the same trend covered above. Naming Stories specifically as the format was a narrower and riskier bet than it looked at the time. Engagement and ad dollars concentrated instead in algorithmic feeds: the TikTok For You page, Instagram and Facebook Reels, YouTube Shorts, rather than the separate Stories tray. Betting on a specific platform also carried a risk this prediction couldn't have flagged. TikTok itself became a genuine question mark in the U.S. once lawmakers moved to restrict it starting in 2024, the kind of platform-level risk a behavior-based prediction never has to worry about.
The lesson: naming a specific feature or platform is a narrower bet than naming the behavior behind it. Features get replaced and platforms get regulated far faster than the underlying shift in how people actually watch content.
WHAT THIS CHANGES ABOUT HOW WE MAKE CALLS NOW
We still make trend calls every year, internally and for clients, because writing an assumption down is what makes it possible to check later. The difference this exercise reinforced: we now separate the behavior we're confident about from the specific channel or feature we're betting it will show up in, and we say so out loud. That's a smaller, more honest claim, and it's the one that's actually useful to revisit next year.
FAQ: EVALUATING MARKETING TREND PREDICTIONS
HOW DO YOU KNOW IF A MARKETING TREND PREDICTION WILL HOLD UP?
Look at what the prediction is actually betting on. A forecast describing a behavior already in motion, more short-form video, more automation, tends to hold up because it's extending momentum that's already visible. A forecast naming a specific feature, app, or platform is a much larger bet, because those get replaced or regulated faster than the behavior underneath them changes.
IS IT A BAD SIGN WHEN A MARKETING TREND PREDICTION DOESN'T COME TRUE?
Not by itself. Some trends get predicted almost every year, guerrilla marketing and "the year video finally matters" are good examples, because they describe something that's always somewhat true rather than a specific shift tied to that year. Those aren't really forecasts, so there's not much to grade. The more useful test is whether a prediction changed an actual budget or campaign decision. That's the kind of call worth grading honestly a year later.
SHOULD A COMPANY STILL BOTHER MAKING YEARLY TREND PREDICTIONS?
Yes, treated as a planning exercise rather than a prophecy. The value isn't forecasting accuracy, it's forcing a team to write assumptions down in a specific enough form that they can be checked later. This post is a small example of what that looks like in practice.
WHAT SEPARATES A REAL TREND FROM SOMETHING THAT JUST SOUNDS NEW?
A real trend changes measurable behavior: what platforms people spend time on, what content an algorithm rewards with reach, how ad budgets actually shift. Something that sounds new without changing behavior at scale, a lot of the recurring guerrilla-marketing conversation falls here, tends to be evergreen commentary rather than an actual shift worth acting on.
