Library
32 min read

Design Ticketmaster

A stadium tour goes on sale at 10am. Two million fans want sixty thousand seats, and each fan does not want "a seat", they want G5 in the floor section, next to their friend in G4. Every seat is its own tiny auction, and it closes in milliseconds.

Watch one fan get from the line to a ticket.

FansWaiting roomAPISeat map cacheHold serviceSeat storePayment gatewayOrders and tickets

Line up. At 10am two million fans land on the waiting room. It gives each one a place in line and does nothing else.

The picture is two systems with very different promises. The top half, the seat map, may be a little out of date, and it takes nearly all the reads. The bottom half, the hold, must be exact for every single seat, and it takes very few writes. Most of this page is about keeping those two promises separate.

If you have read IRCTC tatkal, here is the difference that matters. There, seats in a class are interchangeable, so the whole fight is over one number per train. Here the fan chooses seat G5, so every seat is its own row and its own race, and a group of seats has to be won together. The same problem shows up when you book two seats next to each other for a first day, first show at a cinema, or tickets for an IPL final.

The real Ticketmaster

This page builds the core of a ticketing site like Ticketmaster: choose seats from a map, hold them while you pay, and survive an on-sale where demand is many times the supply. The numbers are interview assumptions, not Ticketmaster's. Two real things worth knowing: Ticketmaster puts popular on-sales behind a queue and runs Verified Fan registration to filter out bots before the sale, and it said the Eras Tour presale in November 2022 drew 3.5 billion system requests, four times its previous peak, which it blamed on bots and fans without invite codes. Bot defence is a large real system that this page names but does not build.

A few words used all through this page:

  • A hold takes seats off sale for a few minutes while one fan pays. If they never pay, the seats go back.
  • The seat map is the picture of a section with every seat coloured free, held or sold.
  • The waiting room gives each fan a place in line before they may shop, like the token machine at a bank.
  • An on-sale is the moment a popular event's tickets go on sale, announced in advance.

Step 1: Understand the problem#

Ask about the seats before you ask about the traffic. Whether a fan picks a specific seat or just "two tickets in the upper tier" decides what the contended thing is, and everything else follows from that.

You askThey sayWhat it settles
Do fans pick specific seats, or just a section and a count?Specific seats from a map, with a "best available" button as well.Every seat is its own row and its own race. The contention is spread over sixty thousand rows instead of one counter, but a group of seats now has to be claimed all at once.
How big is the on-sale?About two million fans for about sixty thousand seats, at a time announced weeks ahead.Thirty or more fans per seat, so most fans leave without a ticket. Like tatkal, the front door is a rejection system and the booking path behind it is small.
How long can a fan sit on seats while paying?A few minutes, with a countdown. Assume 8.A hold with an expiry on every seat row, and a rule for when time runs out in the middle of a payment.
Can the seat map be out of date?By a couple of seconds, yes. The hold must never be wrong.Two promises. The map is a cache that can lie a little, and the hold is checked against the seat rows every time. A fan tapping a seat that just went is answered with a 409 and a fresh map.
How many tickets can one order have?Up to 8, and the seats are usually next to each other.All or nothing claims across several rows. A fan who wanted a pair and got one seat has been handed something useless that is now off sale.
What about bots and resellers?A real problem. Assume registration filters most of them before the sale.A per account limit on tickets and on active holds, enforced at the hold. Full bot detection is out of scope, but the design must not reward a script that holds seats it never pays for.

What you are building, and what you cut#

Functional requirements
What the system does.
  • Browse an event and see the seat map by section. Almost all of the reads. Allowed to be a couple of seconds stale.
  • Hold chosen seats, or the best available block, for 8 minutes. The hard part. Exact, per seat, and all or nothing for a group.
  • Pay and turn the hold into tickets. Money moves in another company, which is the payment rail problem again.
  • Survive the on-sale with a waiting room. Thirty fans per seat arrive in the same minute.
Non-functional requirements
How well it has to do it. The numbers you commit to.
  • No seat is ever sold twice.
  • About two million fans in line for a 60,000 seat show, selling out in about ten minutes.
  • A hold or a "that seat just went" answer within about a second.
  • Fair order of admission: joining the line earlier should count.
Out of scope
What you cut, and say so out loud.
  • Bot detection and fan registration. Named as real, and upstream of this design.
  • Resale and ticket transfer. A second marketplace with its own rules.
  • Pricing that changes with demand. It changes the number on the seat, not how the seat is claimed.
  • Venue setup and seat map drawing. Seat maps are created once per venue, long before the sale.

Back of the envelope#

A small version first. A 60,000 seat stadium, four tickets per order, is 15,000 orders. If it sells out in ten minutes, that is 1,500 checkouts a minute, which is 25 a second. That number is small. The people trying to become one of those 25 a second are not. (The back of the envelope chapter covers the method.)

Fans against checkouts
2000 thousand
60000
4
10
Orders in the whole venue60,000 / 4 = 15,000
Fans per order2,000,000 / 15,000 = 133 to 1
Checkouts per second to sell out15,000 / (10 x 60) = 25.0
Fans who leave without tickets1,985,000
25 checkouts a second
is all the booking path has to do

The booking path is tiny: 25 new checkouts a second would sell the venue out in 10 minutes. Everything big on this page exists to keep 2,000,000 people from touching it at once, while the few who are let in see a map that is close enough and a hold that is exact.

Read the first and third rows together. The writes that matter are a few dozen a second. The waiting room and the seat map cache carry everything else, and the hold service never sees the crowd at all.