liveBook · Manning

Cold start problem

As explained in Fast by Default

The elapsed time required to start an application or service when it is not already initialized in memory.

A cold start is the time required for an application that is not in memory to fully initialize and become usable. It is the most demanding startup path, unlike a warm start from an inactive in-memory app or a hot start from a recently foregrounded app. Because it shapes first impressions, cold-start delay can affect acquisition, engagement, retention, and app-store reviews. section 3.2.3 section 6.2.6

Treat cold start as a user-centered performance measure rather than a single universal number. Simple mobile apps can target under 2 seconds, while complex apps with significant data loading can allow under 4 seconds. A target should also name the user segment and conditions it represents, such as under 2 seconds for 90% of users on mid-range devices. section 3.2.3 section 3.4

Set different cold-start expectations for the environments users actually have:

  • High-end mobile devices can target under 1.5 seconds; mid-range devices under 2.5 seconds; low-end devices under 4 seconds, with lower frame-rate expectations on low-end hardware. section 3.4.2
  • Desktop applications can target launch to the first interactive screen in 3 seconds under fresh conditions, or 5 seconds during heavy multitasking while maintaining acceptable responsiveness. section 3.4.2 section 6.2.6
  • Measure on representative supported hardware, especially slower devices, rather than relying on development machines whose performance can hide startup problems. section 6.2.6

Turning a slow mobile launch into a budget

A fitness tracking app supports both flagship phones and budget Android devices. On a representative mid-tier device with 4GB of RAM, its cold start takes 4 seconds. The team sets a 2.5-second budget, profiles startup, and identifies initialization work that can be delayed or run in parallel. The budget becomes an early-warning threshold: if the user-facing goal is 2 seconds, a 1.8-second budget leaves room to detect regression before users feel it. section 6.2.6 section 6.8

Budget the startup path in visible phases instead of only timing the total. Track when the splash screen appears, when the user can first interact, and when data becomes visible. This separates an apparently slow launch into measurable work, makes tradeoffs explicit, and helps teams assign limits before a regression reaches users. section 6.2.6

Use a repeatable cold-start budgeting workflow:

  • Choose cold start when first impressions, user complaints, or business impact make it the highest-priority mobile performance goal. section 6.8
  • Measure a realistic baseline at least three times and use the median result; use platform tooling on a mid-tier device representing typical users. section 6.8
  • Set a budget below the user-facing goal, or use phased budgets when the current baseline is far from the desired target. section 6.8

How Fast by Default covers it

How deep it goes
Defines startup states, supplies contextual thresholds, and details baseline, phase, and budget-setting procedures without platform-specific implementation internals.
Examples
Uses a fitness tracking app on a mid-tier Android device and desktop launch under fresh and multitasking conditions.
What you can do after
Operate/configure cold-start goals and budgets using representative-device baselines, phased measurements, and regression buffers.
How it connects
Connects cold start to warm starts, hot starts, frame rate, performance budgets, profiling, and user-centered targets.

Also explained in

Compare how 7 books explain cold start problem →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Fast by Default. The section links point to the parts of the book it draws on.