It's Monday, there are five slots on the calendar, and the ideas list is either empty or full of things you've already tried. Ideas aren't really the problem: a trend page or a chatbot will hand you a list in seconds. The problem is knowing which five are worth a shoot for your account. No single source of ideas answers that. A mix does: proven shapes from your own posts, gaps nobody has filmed yet, and a riskier idea tested where it can't hurt you. Here's what each source of ideas is good for, and where it lets you down.
Where next week's ideas usually come from
The team brainstorm
Everyone brings two ideas to the planning meeting and the room picks.
Good for: ideas that fit the brand's voice, and a team that feels ownership of the plan. Where it breaks: the room picks what it likes, which isn't always what the audience responds to, and the same people tend to bring the same kinds of ideas. Nothing in the meeting says which idea has evidence behind it.
Trend pages
TikTok's Creative Center shows trending hashtags you can filter by industry and time frame (TikTok); its Top Ads page shows "high-performing auction ads", so paid content only (TikTok Creative Center). Instagram lets you "bookmark trending songs & audio clips" and shows curated suggested reels so you can see what's trending (Instagram).
Good for: timing. A sound or format on the way up gets you in front of people who are already watching that kind of video. Where it breaks: anyone in your market can see the same list, and a trend says nothing about whether it suits what you sell. By the time a format is on every list, it may already be worn out.
An AI assistant
ChatGPT will give you a list of ideas from a two-line brief. YouTube now has Ask Studio, which can "brainstorm video ideas, get channel insights, and analyze your performance" and "only uses data from your channel and audience" (YouTube Help).
Good for: volume and first drafts: angles, titles, caption variants. Where it breaks: a general chatbot works from what you give it: a brief, a pasted export. Unless you feed it your numbers, its ideas fit any account like yours. Ask Studio, by YouTube's own description, uses only your channel's data, so it covers one platform and only your side of it. We compare this route in detail in why not just ask a chatbot.
Your own best posts
Go back through your last few months and find the posts that held people past the opening. Take the shape (the opening, the format, the length) and put a new topic in it.
Good for: this is the most reliable source you have, because it already worked on your audience. Where it breaks: do only this and the account narrows. Every post starts to look the same, and you never find out what else your audience would watch.
What you've never filmed
List what you sell or do, then list what has actually been on camera in the last three months. The difference is a coverage gap: products, services, or parts of the process customers never see. The questions in your comments are the same kind of gap, written by your audience.
Good for: ideas nobody else can copy, because they're about your business. Where it breaks: a gap tells you what to film, not how. A video about an unfilmed product still needs an opening that holds, which is why this source works best paired with the previous one.
Watching competitors
Tools like Hootsuite let you "monitor up to 20 competitors simultaneously" and track "their posts, posting frequency, follower growth, engagement rate, and more" (Hootsuite); Sprout Social's competitor reports cover audience growth, engagement and post volume, with the most popular posts and hashtags on Instagram (Sprout Social). Some teams skip the tools and keep a shared folder of competitor posts they liked.
Good for: seeing which of their posts drew a response, and noticing when a rival changes course. Where it breaks: the numbers say which posts worked, not why. The reason sits in the video (the opening, the format, what's on screen), so someone still has to watch them. We cover how to read rivals without copying them in reading the accounts around you.
A research tool
Rankfield is one: it audits your account, watches your videos and your competitors', and finds accounts like yours to learn from.
Good for: putting the sources above (your best shapes, your gaps, the field) into one ranked plan, with the evidence under each pick. Where it breaks: it works from public posts and what you tell it, not your private analytics, and the first plan knows your brand only from your brief and your site. It gets closer to your voice as you rate what it gives you.
Match the idea to what's actually wrong
Which source to lean on depends on what your numbers are telling you. A quick read of the last month against your usual numbers usually points one way:
| What you're seeing | What tends to help this week |
|---|---|
| Fewer people reaching your posts than usual | A sharper opening, in a format that has worked for you before |
| People watch, then nothing happens | A post that asks one specific question or shows a real decision |
| Posts in bursts, then silence | A smaller format you can repeat on a fixed day |
| Every post looks the same | One format you haven't used recently |
| Things you sell never appear on camera | Film the offer that has no video yet |
Fix one of these at a time. If you change the format, the opening, the day and the topic all in one week, you won't know which change did anything.
A plan that combines them
For a week of five posts, here's the split I'd start with:
- Two from your own best shapes, with new topics. This is the floor of the week.
- Two from gaps: an offer you've never filmed, or a question from your comments, each built on one of those proven shapes.
- One experiment from a trend or an AI brainstorm, posted where a miss costs little. Instagram's trial reels are "shown to non-followers first", and Instagram decides whether to share one with your followers based on the views it gets in its first 72 hours (Instagram).
Then check what each one did against your usual numbers before you plan the next week. That review is the other half of the job, covered in is your content working?
Questions
How do I test a riskier idea without hurting the account?
On Instagram, post it as a trial reel: Instagram shows it to people who do not follow you first, and only shares it with your followers if it performs well.