How I Research Trends Before Editing a Video or Creating Content
Editing starts before the timeline. Here's the actual research stack I use to figure out what's worth making in the first place.
Most video and content work doesn't fail in the edit. It fails before the timeline even opens - because nobody checked whether the idea was worth making in the first place. Before I cut anything, I run it through a small research stack: what's actually moving on the platform right now, what people are searching for around the topic, and what an AI tool can help me see faster that I might otherwise miss. None of these tools decide the content on their own. They narrow the field so the editing time goes toward something with a real shot at landing.
Start with what's already moving on the platform
Instagram and TikTok both publish their own trend surfaces, and I check both before planning a short-form piece. Instagram's current Reels trends page shows which audio, formats, and edit styles are actually being used right now on the platform - not guessed at, not from a third-party blog post recycling last month's list. TikTok's Creative Center hashtag trends does the same thing from the advertiser side, with a rolling 7-day window so I'm not planning around something that already peaked.
The point of checking both isn't to chase every trend. It's to know what's live before deciding whether a format fits the actual message. A trending edit style that doesn't fit the subject just looks like noise. A trending style that does fit is free distribution.
Cross-check with what people are actually searching
Platform trends tell me what's being watched. They don't tell me what's being searched for, which matters just as much for anything meant to have a life outside the algorithm feed - a YouTube video, a website video, a piece meant to answer a real question. For that I use Google Trends to check whether interest in a topic is rising, flat, or already fading, and AnswerThePublic to see the actual questions and phrasing people use around a topic, not the phrasing I'd assume they use.
This step catches a specific mistake: making content around what's interesting to make, instead of what's actually being looked for. The two overlap less often than you'd think.
AI tools narrow the list, they don't make the call
I use ChatGPT, Claude, and Perplexity in this process, but not to generate ideas from nothing. Once I have real trend data and real search phrasing in front of me, AI tools are useful for synthesizing a pile of raw signals into a shortlist faster than doing it by hand - grouping similar angles, spotting an obvious gap, or checking whether an angle has already been done to death. Perplexity specifically is useful for a fast sanity check on whether a topic is already saturated, since it pulls current sources rather than relying on stale training data alone. The judgment call on what actually gets made is still mine. AI narrows the pile; it doesn't pick the winner.
How it comes together before an edit starts
This is the same honesty standard as the rest of this site: I'm not claiming a formula that guarantees a viral result, because nobody can honestly promise that. What this process does is remove the most common failure mode - spending real editing time on something nobody was ever going to look for. If you want to see the editing side of this once the research is done, the Video Editor service page has real edited samples, not a stock reel.
FAQ
No, and I won't claim it does. What it does is remove the most common failure mode - making something nobody was searching for or watching for in the first place. The performance still depends on execution, timing, and the platform's own distribution, none of which any research tool can fully control.
The two platforms don't always trend the same formats or audio at the same time, and a piece built for one doesn't always translate directly to the other. Checking both before planning avoids building around a trend that's already specific to a different platform's audience.
Speed on synthesis, not the creative decision itself. Once real trend and search data is gathered, tools like ChatGPT, Claude, and Perplexity are useful for grouping similar angles and flagging oversaturated topics faster than doing it manually - the final call on what gets made is still a human judgment call.