How long does a coin stay trending, and why it falls off so fast
Why trending positions decay in minutes rather than hours, how recency weighting works, and what makes a coin hold a board or flash across it.
Creators get surprised twice. First that their coin appeared on a board at all, then that it was gone twenty minutes later. Both reactions come from the same misunderstanding.
A ranking isn't a position, it's a recalculation
There's no list you get added to and removed from. Every discovery surface recomputes its ordering continuously, from on-chain data, over a short trailing window.
So you never hold a position. You re-earn it every time the ranking recalculates, and if the activity that earned it has aged out of the window, you're not there any more. Nothing was taken from you. The clock moved.
How short is short
Different surfaces use different windows and none of them publish the weighting. What's consistent is the shape:
- The last few minutes count enormously
- The last hour counts somewhat
- Anything older counts very little
- Lifetime totals count for almost nothing
A coin with 400 SOL of lifetime volume and no trades today ends up invisible, while a coin minted 40 minutes ago with steady flow sits near the top. Every one of these surfaces answers "what's happening now", never "what has happened".
The spike problem
Now the most expensive mistake in this category.
Someone spends a budget on activity, it all lands inside twenty minutes, the coin appears on a board, and then the window rolls forward and the coin disappears. Money gone. Exposure lasted exactly as long as the spike.
Everyone browsing during those twenty minutes saw the coin. Everybody else, which is almost everybody, didn't. And because it then vanished, the people who did see it have no reason to look again.
Now spread the same budget over twelve hours instead. The coin keeps re-entering the recency window, gets ranked repeatedly, and reaches a different set of browsers each time. Same spend, far more separate people. Duration is a first-class setting for exactly this reason rather than an afterthought, and how trending works covers it from the ranking side.
What keeps a coin on a board
Four things, roughly in order.
Continuing trades matter more than trade size. A steady stream of small swaps keeps you inside the window, while a pause of even 10 minutes during a busy stretch can be enough to drop you out of it entirely and hand your slot to whatever minted after you.
Distinct wallets over time matter next. A ranking that watches the same handful of addresses trade back and forth reads that differently from one watching new participants arrive. Fresh wallets entering is a stronger signal than old wallets cycling.
Trades from real arrivals matter most of all. Once a board sends people to your coin, some fraction of them trade, and those trades feed back into the ranking. Compounding is what everyone actually wants here, and it only starts if the page they land on gives them a reason to stay.
And then: no negative shock. A large sell, a liquidity move, or a visible dump from a creator wallet doesn't just hurt sentiment. It changes the data the ranking reads.
Why coins fall off
| What happened | What the ranking saw |
|---|---|
| Activity ended | Window rolled past, nothing recent |
| Activity was concentrated | One participant, not a crowd |
| Bigger launches appeared | Relative sort, you got pushed down |
| Traffic arrived and left | Visits without trades produce no data |
| Creator sold | Curve progress reversed, holders reacted |
Row three sits outside your control, and it's worth sitting with. Rankings are relative, so a quiet Tuesday and a busy Friday evening need completely different amounts of activity for the same slot. Trending requirements explains why no fixed threshold exists.
The compounding case, honestly
Everything above sounds bleak, so here is the version that actually works. A coin gets ranked, people arrive, some fraction of them buy, those buys feed straight back into the same recency window that produced the ranking, and within an hour the coin is holding its position on genuine flow rather than on anything you paid for. Launches take off through that loop, not through any single push.
The condition is the middle step. What happens if people arrive at a coin with a moving chart, an empty comment section, a description that says nothing, and a creator profile that looks abandoned? They leave without trading, the loop never starts, and the ranking delivered a visit that the page then wasted.
Social presence on the page isn't decoration. It's what converts delivered attention into the trades that keep you ranked.
What to do with this
- Buy duration before you buy size. Twelve hours of moderate flow beats one hour of heavy flow for exposure
- Don't aim at a single peak. Peaks photograph well and reach one slice of browsers
- Make the landing page worth arriving at before you pay for arrivals
- Expect decay. Falling off a board is normal, not a failure of the campaign
- Watch trade count, not just price, because that's what the ranking reads
None of this manufactures demand. Staying on a board longer puts your coin in front of more people. It doesn't make them want it. Creators who treat a ranking as the finish line usually end up disappointed, and the ones who treat it as the start of a conversation tend to do better, because they prepared something for the arrivals to find.
For the arithmetic of picking that window, how long a campaign should run works through the trade-off. To size activity against a window rather than a spike, the console prices duration explicitly before anything is signed.
Frequently asked
Only if it keeps producing activity for a whole day. Rankings recompute against short windows, so a position is never held, it's re-earned every few minutes. Coins that appear to sit on a board all day are being traded all day.
Generally no. Most discovery surfaces sort on recent activity rather than size, which is why small coins with heavy trading outrank larger ones that have gone quiet. Size affects which boards you qualify for, not how long you stay.
Several times, almost always. Each appearance reaches whoever is browsing at that moment. One strong spike reaches the people watching during those minutes and nobody else.
Everything in this guide describes mechanics that settle on Solana mainnet, so you can check any of it yourself. Mint addresses, swap signatures, curve progress and wallet counts are all public, and an explorer will disagree with us if we are wrong.
Curvegrad sells a volume service, and guides that touch on what a campaign does say so in the text rather than in a footnote.
- Solana mainnet transaction history, read through the standard RPC methods
- Pump.fun bonding curve and PumpSwap pool accounts as they appear on-chain
- Public block explorers, where every claim here can be verified against a real mint
- Campaigns run through our own engine, which is where the operational detail comes from