How to get your first 100 players on a Minecraft server

Your first 100 players come from making the first five minutes good enough that people return, and only then from sending people to the server: a few server lists with vote rewards, a Discord that gives them a reason to come back, and one clear moment to join, such as a launch or a new season. Promotion multiplies whatever your server does with a new player, so if it loses everyone in ten minutes, ads only make it lose them faster.

By 100 players we mean 100 people who come back, not 100 online at the same time. The second number is rarer than most owners think. In the last 7 days, 22 of the 112 playable listings on serverlist.net averaged 100 or more players online, and the median listing averaged 4.9. When we wrote this on 18 October 2026, that was 22 of 115, with a median of about 4.

Before you advertise: what a new player sees in five minutes

Everything below is about getting a stranger to click Join. Whether they stay is decided by what happens next, so walk through it as they would.

1. The server row. Before anyone joins, they see your MOTD, your version range, a ping bar and a player count. "0/500" next to a two-line MOTD full of capital letters is a hard sell. The server description guide covers a MOTD that fits.
2. The connection. A forced resource pack of several hundred megabytes on a phone connection loses people before they see a block. Keep it small, or make it optional.
3. Spawn. Can a new player tell, without reading chat, where to go? A portal, a sign, an NPC or a short path beats a wall of rules text. If your mode needs a wild world, test /rtp or whatever your server uses and make sure it lands somewhere safe.
4. The first goal. Survival servers often give nothing to do in minute two. A starter kit, a first quest, a free claim or a "build your first base here" area gives the player a reason to make something they will want to return to.
5. Chat. An empty server is its own deterrent. Be online yourself at your busiest hours, answer the first "hi", and have staff who say hello to newcomers by name.

The cheapest test there is: ask someone who has never seen your server to join while you watch their screen and say nothing. Every time they stop and look confused, you have found something to fix.

Timing matters as much as the spawn. Players join servers where other players are, so do your launch, your events and your own online hours when people are around. Across every server on our list, the busiest hour right now is 16:00 UTC:

Peak 16:00 UTC

0003060912151821

Times in UTC

Busiest at 16:00 UTC with about 17,845 players, quietest at 01:00 UTC. The busiest hour holds 2.6 times the players of the quietest.

Measured 19 Oct 2026, 22:55 UTCHow we measure

Pick a niche you can win

"Another survival server" is the hardest thing to grow, because the player you want already has one. You need a sentence that makes them curious: "a lifesteal server where hearts reset every 30 days", "an SMP for players over 18 with a strict build-quality whitelist", "vanilla towny on a 1:1000 Earth map". The mode gets them to look, the twist gets them to join.

Our data shows how differently modes behave. SMP is the biggest tag on our list, with 115 listings, of which 68 were playable in the last week, and the median playable SMP averaged 6.8 players. Lifesteal has far fewer listings, 47, of which 18 were playable, but the median playable lifesteal server averaged 47.1. On 18 October 2026 that was 114 SMP listings with 71 playable and a median of about 4.6 players, against 47 lifesteal listings with 19 playable and a median of about 26.

Read that carefully. SMP is easy to start and crowded with small servers, so standing out is the whole game. Lifesteal is a mode where many listings never get going, and the ones that work hold more players, so players compare you with servers that are already busy. Neither is "better". Some smaller modes on the ledger below are dominated by one or two servers, which means you are not competing with a crowd but with a single established server, and you need a reason to be the alternative.

Mode	Listed	Playable now	Players now	Players, 7 day average	Busiest hour	Votes, 30 days
PvP	104	50	2,466	3,544	15:00 UTC	12,990
SMP	115	58	2,390	3,517	15:00 UTC	14,150
Survival	70	40	1,854	2,346	15:00 UTC	12,638
LifeSteal	47	17	1,498	2,255	15:00 UTC	11,292
Towny	12	6	680	1,358	13:00 UTC	4,300
RPG	10	6	1,117	1,290	17:00 UTC	680
PvE	26	17	517	994	15:00 UTC	1,295
Raiding	13	6	739	979	14:00 UTC	9,071
Anarchy	29	15	667	974	15:00 UTC	3,532
Skyblock	15	7	751	935	15:00 UTC	2,907
Factions	12	6	735	849	16:00 UTC	3,841
OneBlock	9	4	608	762	15:00 UTC	3,170
Prison	9	4	515	704	15:00 UTC	2,815
KitPvP	14	7	483	665	14:00 UTC	3,508
Bedwars	11	6	213	639	10:00 UTC	174
Creative	9	6	472	620	15:00 UTC	3,317
Vanilla	22	11	156	559	10:00 UTC	7,143
Earth	6	3	352	371	17:00 UTC	45
BoxPvP	7	4	143	234	10:00 UTC	175
Gens	5	2	83	116	20:00 UTC	0
Mini Games	20	12	32	54	14:00 UTC	980
Skywars	4	3	1	18	14:00 UTC	11
Modded	8	4	5	8	17:00 UTC	110
Cobblemon	3	2	5	8	17:00 UTC	109

A server tagged with three modes counts in all three, so the rows overlap and do not add up to the network total.

Measured 19 Oct 2026, 22:55 UTCHow we measure

Small modes have small samples, so treat a single week as a hint, not a verdict. Before you commit, look at the hub for your mode, for example the SMP servers, lifesteal servers or survival servers, and check three things on the busiest listings: what they do that you would do differently, which hours they are busy, and whether they are really up. The game modes by the numbers report has the monthly view.

Server lists: what moves you up

Server lists are where many players go looking for a new server, and every list ranks differently. Here is how ours works, since it is the one we can describe exactly.

Votes and rewards

Players vote when the server gives them something for it, and that needs Votifier, which passes each vote to your reward plugin. In the last 30 days, listings with Votifier configured received 99.7% of all votes on serverlist.net, although only 95 of 216 listings have it set up (99.8% of votes and 97 of 218 listings on 18 October 2026). Listings without it collect almost nothing, however good the server is.

Set up NuVotifier on day one with the NuVotifier guide, check it with the Votifier tester, then build the loop:

• a /vote command that prints every list's vote address,
• a small reward per vote that a new player feels on their first day (a key, some currency, a cosmetic),
• a reminder when a player's vote cooldown has passed,
• a streak or monthly reward, because most voters stop early. In the last 30 days, 61% of voter and server pairs voted only once (61% on 18 October 2026).

Bedrock players cannot click links in chat, so if you accept Bedrock players, print the address as text. The Geyser guide covers how their names reach your vote plugin.

Uptime and online-first ranking

Our default order puts every listing that was playable in the last 24 hours above every listing that was not, and only then sorts by votes in the last 30 days. A server that is down for a day drops below all running servers, whatever its vote count. "Playable" is strict: a free host's sleeping placeholder or a "maintenance" MOTD answers our ping but does not count. The methodology page has the full definition.

If you host on a plan that sleeps or restarts often, fix that before anything else in this guide. A player who finds you offline once rarely tries a second time.

Your listing page

Your listing is the page a curious player reads before copying your address. A banner, a description written for that list, and filled sections for rules, economy and vote rewards do more than any paid slot. The server description guide goes through it.

Paid placement exists too. On serverlist.net it is called a boost, it is our own product, and it puts a listing in a labeled sponsored slot for one to four weeks through the advertising page. A boost buys attention, not retention, so it works best when the first five minutes are fixed and you have a launch or a season start to point at.

Discord as your home base

Players drift between servers, but a Discord keeps them in reach when they are offline. 171 of 216 listings on serverlist.net link one (173 of 218 on 18 October 2026), so a Discord is expected, not a bonus. Use it for the things that bring people back: announcements of events and resets, a support channel where staff answer within hours, and a place for the community to show off builds. Link game chat with DiscordSRV so the server feels alive even to people who are not logged in.

A serverlist.net link pasted into Discord unfurls into a card with your players online, your votes and buttons to view the server and vote, so post your listing link rather than a bare IP. The Discord guide covers channels, roles, onboarding and moderation in detail.

Creators, Reddit and short video: what works for small servers

Large creators are out of reach for a new server. Small ones are not. A creator with a few thousand subscribers who already plays your mode is more likely to answer, more likely to bring viewers who join, and more likely to come back. Offer something that makes good video: a build event, a first-week tournament, an early look at a new season. If you pay or give a creator anything, say so openly: Mojang's Usage Guidelines require servers to make creators who promote them disclose the relationship and follow the US Federal Trade Commission's guidelines.

Reddit and similar communities have places for server ads, each with its own rules on how often you may post. Read them first. A post with one real screenshot or a 20-second clip of something that happened on your server beats a list of features, because every server claims "custom plugins" and "friendly staff".

For short video, show one mechanic in the first two seconds: a heart being stolen, a cannon going off, a town from above. Then put the address on screen. Nobody watches a short to the end to read an IP.

Seasons and resets as growth events

A new season is the one moment when nobody on your server has an advantage, and that makes it the easiest time to invite strangers. 24 game modes on our list declare a season length, and 12 of them run 30-day seasons. When we wrote this, those numbers were 24 and 12.

Mode	Declared	Median	Lengths
SMP	9	67 days	30 days4 times6 months90 days67 days120 days18 weeks
PvP	4	45 days	30 days2 times2 months30 months
LifeSteal	3	60 days	60 days90 days30 days
Anarchy	3	30 days	30 days2 times90 days
Skyblock	2	32 days	5 weeks30 days
Factions	1	90 days	90 days
Prison	1	30 days	30 days
Mini Games	1	30 days	30 days

Season lengths come from what owners write on their listing. Ask in the server's Discord when the current season started.

Measured 19 Oct 2026, 22:55 UTCHow we measure

• Announce the date early. Two weeks ahead on your listing, your Discord and your MOTD gives creators and returning players time to plan.
• Start at your region's peak. A Saturday at the busiest hour beats a Tuesday morning launch.
• Keep something across seasons. A hall of fame, cosmetics or a map archive gives veterans a reason to return and talk about it.
• Put the schedule on your listing. The "Seasons and resets" section exists for that, and it answers the question new players ask before they invest a week.

How long it takes: what our listing data shows

Growing a server is slow, and many listings stop before they find out whether it would have worked. Of the 119 listings first published on serverlist.net at least six months ago and still public, 59 were playable at all in the last 30 days, and 43 were playable at least 90% of the time. On 18 October 2026 that was 61 and 42 out of 117. Listings that their owners deleted before our current system are missing from these numbers, so the real share that stops is higher.

Playable 90% or morePlayable at least onceNot playable

1. Nov 202512/26
2. Dec 202518/33
3. Jan 20266/18
4. Feb 20263/11
5. Mar 20265/12
6. Apr 20268/10
7. May 20264/9
8. Jun 20269/14
9. Jul 202614/22
10. Aug 202613/16
11. Sep 202614/15
12. Oct 202612/12

Of 119 listings that are six months or older, 59 were playable at least once in the last 30 days (50%).

Measured 19 Oct 2026, 22:55 UTCHow we measure

New listings start quiet. Of the 36 listings published in the last three months that were playable last week, the median averaged 0.6 players, and 22 averaged fewer than one. On 18 October 2026 that was a median of 0.8, with 20 of 38 under one player. Older listings have higher medians, but be careful with that comparison: they include servers that were already established when they listed with us, and the ones that failed are no longer there.

The list as a whole is very top-heavy. The 10 busiest listings hold 70% of all players on the list (67% on 18 October 2026), and most servers sit in the lowest buckets:

1. 60
2. 38Under 1
3. 131 to 5
4. 185 to 20
5. 1520 to 100
6. 17100 to 500
7. 5500 or more

Median server

4.9

Five biggest

54%

Ten biggest

70%

Each column counts servers by their average players. Half of the 112 servers average 4.9 players or fewer.

Measured 19 Oct 2026, 22:55 UTCHow we measure

What that means in practice: expect the first three months to look like failure even when they are not. Judge progress by returning players and by your peak hour climbing week over week, not by the overall average. The players per server report and the listing survival report go deeper.

Measure it

Most owners have no idea which list sent their best players, and that is fixable in an afternoon.

Per-list hostnames. Create one hostname per place you advertise, all pointing at the same server: sl.yourserver.net for serverlist.net, yt.yourserver.net for a creator's video, and so on. If you use SRV records, give each hostname its own SRV record and its own target name, so the names still differ when they reach your server. Java clients send the hostname they connected with as part of the handshake, Velocity exposes it to plugins as the virtual host, and analytics plugins such as Plan (Player Analytics) can group players by join address and show retention per group. After a month you know which source brings players who stay, not just players who click.

Owner analytics. Your serverlist.net dashboard shows views of your listing, address copies, votes and where visitors came from. Compare address copies from our listing with joins on sl.yourserver.net: the gap between the two is roughly how many people copied your address and then decided against you, which usually points back at the first five minutes.

What not to do

• Fake player counts. Plugins that inflate the number in your server list ping are easy to spot from outside. Our pinger records your count every minute, and a real server's curve follows the clock, quiet at night and busy in the evening. A padded count does not, and players who join an "80 online" server with three people in chat do not come back. A list that notices can remove the listing.
• Bought votes and auto-voters. They break list rules, ours included, and they fill your vote count with names that never play. Our rolling 24-hour cooldown checks the IP address, a browser cookie and the username, and every vote needs a captcha. How voting works explains why.
• Pay to win to fund ads. Selling a competitive advantage runs against Mojang's Usage Guidelines and against what players say in reviews. Making money within Mojang's rules sorts out what is allowed.
• Advertising on other people's servers. Posting your IP in another server's chat or Discord gets you banned there, and owners talk to each other.

When you are ready to list, add your server. The address step checks Java and Bedrock live, and the share kit at the end gives you vote links, status banners and in-game vote text to paste.