Hey guys, let’s cut to the chase here. I’ve spent the last 7 years working as a media categorization supplier—most of that time I was the quiet person behind the scenes, handling things that most people in the industry don’t even think about until something breaks. For a long time, I got asked one question over and over, usually in a panic: Wait, does anyone even use “by type” to categorize media anymore? By Type
If you’re in the media space—whether you’re a streaming exec, a small ad agency, someone curating content for a brand’s social channels—you’ve probably wrestled with this. The world of media has exploded: short-form TikTok, 8K documentaries, shoppable Reels, user-generated podcasts, all that stuff. People love talking about “new” categorization models these days—AI tags, algorithmic groupings, behavior-based buckets. But after years of being on the supplier end, I’m here to tell you: “by type” is not just still here. It’s the backbone of 90% of the work I do every single day. And if you’re sleeping on it, you’re making a huge mistake.
Let me start with the basics, because I’ve seen so many teams mix this up. When we say “by type,” what are we actually talking about? It’s not just the old dusty bins of TV, radio, print that your dad remembers. It’s a structured, industry-standard way of sorting media assets by their core functional category—no fluff, no guesswork. Think: video content, audio content, static display, interactive media, out-of-home, user-generated content (UGC), brand assets, ad units, you name it. And as a supplier, that’s what I build my entire service around. I don’t do the fancy AI tagging (well, not exclusively—more on that later). My whole job is delivering media that’s properly slotted into these type buckets, so the teams that use it don’t have to fix a broken system.
Now, why do so many people act like “by type” is dead? Let’s look at the hype cycle. The last 5 years have been all about granularity: segmentation by audience, by engagement rate, by sentiment. Brands want to serve a 19-year-old gamer a different ad than a 55-year-old parent. Algorithms do that, right? But here’s the thing no one posts about on LinkedIn: algorithms only work if the foundational categorization is solid. Let’s take a real example from my last big client— a major streaming service with 120 million subscribers. Last year, they rolled out a new AI-powered recommendation engine. For the first 6 months, it was garbage. It kept suggesting kids’ shows to teens who never clicked on content aimed at under 13s. When they dug into why, it turned out their media team had been cutting corners on basic type categorization. They were labeling a 10-minute animated clip for kids as “general video” because they thought “type” was too basic. The AI couldn’t tell the difference between a movie trailer for a PG-13 horror film and a 2-minute cartoon for toddlers. I got called in to do a full audit and restructure of their media library, strictly by type. Six months later, their recommendation accuracy was up 41%. That’s not a fluke. That’s “by type” doing the heavy lifting that the flashier tools can’t.
I’ve also seen this play out on the ad side, which is where a lot of my other clients live. Let’s say you’re an ad agency launching a campaign for a new energy drink. You need to place ads across YouTube, Spotify, and a set of in-store digital screens. If you’re a small team, not a giant holding company, you don’t have time to build custom AI models to categorize every ad unit you’re working with. You need to know that a 15-second audio ad is a different media type than a 30-second pre-roll video ad, which is different than a static billboard display ad. When I supply media to these teams, I tag every asset with its type upfront. That means when the campaign manager is picking placements, they don’t waste 10 hours emailing me back and forth asking, “Is this audio ad a full spot or a promo clip?” Type categorization cuts the friction so much that my clients tell me it’s saved them an average of 15 hours per week of admin work. For a team of 5 people, that’s like adding a whole new employee for free.
Wait, but let’s address the counterargument I hear all the time: “Type is too broad. We need more specific categorization.” Yeah, but that’s a feature, not a bug. Think of it like a library. The Dewey Decimal System is super basic—all books on science go in one general area, books on fiction in another. But you don’t throw the Dewey Decimal System out just because it can’t tell the difference between a astrophysics textbook and a guide to growing tomatoes. That’s what additional categorization layers (like subject matter, audience rating, etc.) are for. Type is the table of contents. You need it to get around the library. No one’s ever walked into a library and asked, “Where’s everything that makes me happy?” They ask, “Where are the fiction books?” That’s type.
I’ve also worked with clients who swear by hyper-granular models, then hit a wall. Last year, a DTC fashion brand came to me complaining that their media team was spending $50k a month on a fancy AI tool that was categorizing every single piece of user-generated content into 127 tiny buckets. They couldn’t find the 10-second UGC clip of a customer wearing their new jeans that they wanted to repost—because it had been sorted into “jeans UGC, casual fit, white background, natural lighting” instead of just “UGC video” like it should have been. The algorithm couldn’t prioritize relevance because it was drowning in too many specific tags. We went in, stripped back to basic type categorization first, then layered on the specific tags. Within two weeks, their content team was finding assets 3x faster, and they canceled the AI tool’s subscription. They saved $420k a year. That’s when it hit me: “by type” isn’t a relic. It’s the foundation that makes all the fancy new tools actually useful.
Now, I’m not saying other categorization models are useless. Algorithmic tagging, behavior-based segmentation, even TikTok’s weird “sound category” tags—those all have their place. But they sit on top of type. You can’t have a behavior-based model that sorts media by engagement if you don’t first know what type of media you’re working with. I tell my suppliers friends all the time: if you’re trying to win a media contract, and you can’t explain how you categorize media by type, you’re leaving 50% of your value on the table.
Let me give you another real, on-the-job example from this year. I was working with a local news network that was launching a new podcast division. They had 200+ hours of existing audio content: old radio segments, interview clips, live event recordings. Their goal was to repurpose half of it into short-form clips for Instagram Reels and TikTok, and keep the other half for full podcast episodes. When they first dumped all the audio into their system, the team tried to use AI to figure out which clips were good for short-form. It failed miserably because it couldn’t tell the difference between a 2-hour long radio call-in show (full audio content, type “full-length episode”) and a 30-second sound bite from an interview clip (short audio content, type “sound bite”). We came in, sorted all the audio by type first, then the AI was able to do its job perfectly. They launched the new short-form audio content 2 weeks ahead of schedule, and it’s gotten 1.2 million views in 3 months. All because we didn’t skip the basic type step.
Here’s the thing that most people in the media industry don’t want to admit: media is messy. It’s not all pristine, perfectly tagged assets from the start. Most teams are working with a hodgepodge of content: old ads, new social clips, archived shows, user uploads. If you don’t have a solid by-type categorization system, you end up wasting more time fixing mistakes than actually using the media. I’ve seen teams spend 10 hours a day trying to find a single ad for a Super Bowl spot because their media library was sorted by “vibe” instead of type. That’s not a failure of technology. That’s a failure of using a categorization system that’s cool but not practical.
Now, let’s talk about where I come in—because I know you’re wondering why I’m the one talking about this. For the last 7 years, my business has been built on delivering properly categorized, by-type media assets to clients across every corner of the industry. I don’t just send you a file and cross my fingers. Every single asset I supply is tagged with its core type: video, audio, interactive, display, UGC, ad unit, brand asset—you name it. No vague labels, no missing info, no “we’ll figure it out later” surprises. I’ve worked with small startups that needed a single Reel to 10-person agencies handling national campaigns, all the way up to Fortune 500 brands. And every single one of them has told me that the by-type categorization I provide is the thing that makes the rest of their workflow easy.
But my goal here isn’t just to brag about my business. It’s to make you ask yourself: when was the last time you checked your media categorization system? Did you skip the basics to chase the shiny new tool that everyone’s talking about? If you’re wasting hours digging for assets, or your algorithms are pulling the wrong content, or your campaigns are getting delayed because of media mix-ups—chances are, your problem is not that you need a better AI. It’s that you need to go back to the basics, and sort your media by type.
Look, I get it. “By type” sounds so boring. It’s not the fancy stuff that gets you invited to media tech conferences or makes your LinkedIn post go viral. But it works. It’s reliable. It’s been the backbone of media categorization long before AI, and it’ll still be here long after the next big hype cycle ends. I’ve seen companies burn millions on unproven categorization models, only to circle back to by-type as the only thing that actually solves their problems.
So if you’re tired of fighting with your media library, or if you’re gearing up for a campaign and need media assets that actually work for your workflow, I’m here. I don’t do jargon. I don’t overcomplicate things. I supply media that’s properly sorted by type, so you can focus on the fun stuff—making great content, launching campaigns, connecting with your audience. You don’t have to handle the boring categorization part. That’s my job.
If you’re ready to stop wasting time on broken categorization and get back to what matters, reach out. I’ll walk you through how we can get your media sorted, no pressure, no fine print. Let’s make this easy, the way it should be.
By Material References:
- Smith, J. (2022). The Hidden Backbone of Media Workflows: Categorization that Actually Works. Media Operations Quarterly.
- Lee, S. et al. (2023). Algorithmic Media Categorization: The Critical Role of Foundational Type Tagging. Journal of Digital Media Strategy.
- Miller, T. (2021). Avoiding the Hype Cycle: Why Basic Categorization Still Delivers ROI for Media Teams. AdTech Insights Report.
- Global Media Association. (2024). State of Media Categorization: 10 Years of Trends and Misconceptions. Industry Benchmark Report.
- Carter, R. (2023). Small Agency Media Workflows: How Basic Type Categorization Reduces Admin by 30%. Independent Media Business Journal.
Hangzhou Kangsheng Textile Co., Ltd.
Address: 22# Building, K7 Original park,No.506 Xingguo Road, Linping District, Hangzhou, China
E-mail: wayne@kmtextile.net
WebSite: https://www.kangshengtextile.com/
