Why fifty videos about one product look the same, and the fix
Ask any tool for fifty videos about your product and you get fifty topics in one look: the same title card, the same phone in the middle, the same swoosh. A viewer who scrolled past the second one scrolls past the fiftieth without noticing it is new. Topic variety is not variety. The thing the eye registers in the first half second is the mechanic, and that is what has to change.
In Season Reels, a lane is a way of telling. A split-flap departures board that works out monthly versus yearly. A quiz with a wrong answer buzzer. A comic strip. A ledger line that fills in. A stopwatch running over a real screen recording. Topics vary inside a lane, but the mechanic is the lane’s signature, and no two lanes share one. The rule we hold a season to: two videos that look alike are a failed season.
The catalogue
There are twenty lanes. These are the ones a season for an app, a SaaS product or a game draws on, with the angle and a few of the mechanics each lane owns.
| Lane | Angle | Mechanics |
|---|---|---|
| pain | The chaos before the product, told as short stories. The product arrives late, as the relief. Never mocks the viewer. | prescription-pad diagnosis, call log filling up, comic-strip argument, late-night ledger |
| demo | One real feature per video, shown honestly. The only screen-heavy lane, so the phone or browser is staged differently every time. | tap-through with sound, split screen before and after, browser-frame walkthrough |
| numbers | Money and time maths the viewer can check in their head. Every number is an estimate with its assumption on screen. | split-flap board, ledger line, stopwatch |
| personas | One customer type per video, told from inside their day. Scenarios, never testimonials. | day timeline, name-tag cast, diary page |
| versus | The old way against the product, fairly. The old way stays likeable. Never a named competitor. | diary versus phone, comic panels, stopwatch race |
| howto | One use or trick per video, short enough to remember. | three-step card, split before and after |
| hooks | Platform-native formats that invite a comment: lists, POV, rating, quiz, red flags. The product is the punchline. | game-show quiz, “rate my”, red-flag list |
| learn | Useful micro-lessons from the product’s own domain. The brand teaches; the product appears only on the end card. | typewriter notes, flash cards, chalkboard |
| offer | What you pay, what is free, how to start, release dates. Every price and date exact from the facts sheet. | cinema marquee, typewriter letter, receipt, store-page countdown |
| brand | Pure brand craft: each video a different medium, each ending on one product truth. | 3D object, painted signboard, logo construction |
| seasonal | Tie-ins to dates that matter to the audience’s market, posted 10 to 21 days ahead. | calendar flip, lights and patterns, countdown |
| language | Videos for viewers who think in a second language: that script on screen, a native voice, word-timed captions. | the strongest lanes rebuilt natively, not translated |
| gameplay | One mechanic or moment per video in real footage. The clip carries it; graphics only frame it. | blurred-fill clip with a kinetic caption, freeze-frame callout, slow-motion replay |
| clutch | Wins, fails and near misses from real play, cut to land the payoff. | countdown to the payoff, rewind and replay, scoreboard sting |
| challenge | Invite the viewer to play. The challenge must be beatable in the game as it ships. | timer overlay, “only N% can” quiz, ghost-run comparison |
| devlog | How the thing is made. Honest and specific, never a staged behind-the-scenes. | before and after build, notebook page, bug blooper reel |
| lore | The world, its characters and its rules, only as they exist in the game or its published canon. | character card, map reveal, in-world document |
A new season for an app starts from pain, demo, hooks, personas, numbers and offer. A game starts from gameplay, clutch, challenge, hooks, devlog and lore. You can add any lane that fits, and the planner can borrow one from another kind when a concept calls for it.
Rules inside a lane
A lane with five videos can still repeat itself, so each one carries its own rules. Three distinct visual systems per lane, so a lane is never one template with five scripts. At most two phone-led videos per lane, because a phone in the middle of the frame is the default look of every app video ever made. And every agent reads the other lanes’ notes before it plans, so a mechanic claimed by one lane is off limits to the rest.
On the first run that last rule did almost all the work. Eleven lanes produced 55 videos with one near miss: a stopwatch over a screen recording in the demo lane and a stopwatch race in the versus lane. The agents spotted it themselves while reading each other’s notes, and the versus video changed its mechanic.
Keeping it fresh across seasons
The harder problem is month four. The same brand, the same product, and a planner that has already used the split-flap board and the quiz. Without memory, the second season repeats the first.
So every video a brand makes goes into the brand’s log: its id, lane, hook, date, platform and result. The log lives with the brand on the server and syncs to your machine with the facts sheet and the research. When the planner fills an empty day, it reads the log and avoids the mechanics the brand has used recently. Variety stops being something you check for after the fact and becomes an input to the plan.
That is also why lanes are defined by mechanic and not by subject. A subject runs out. A way of telling can carry a new subject every month, and a catalogue of twenty of them, each with several mechanics, gives a brand a long way to go before anything has to come round again.