Skip to content
Tutorials Center
My learning

Topics

Lesson 9 of 60 ยท Testing Fundamentals

Alpha, Beta and User Acceptance Testing

Understand alpha, beta and user acceptance testing: who runs them, where they happen, UAT, OAT and contract acceptance, with entry and exit criteria.

When the product is nearly done, the question changes. It is no longer "does each function work?" but "will real people be able to use this for real work, and will the people paying for it accept it?" The last testing steps before release answer that: alpha testing, beta testing and acceptance testing. They are the final stage of test levels and are mostly about validation rather than hunting every last code defect.

Alpha testing

Alpha testing is done in-house, at the end of system testing, before the product goes outside the organization. The testers are internal staff who were not on the development team (QA, support, sales, other departments) and sometimes a few invited customers working on the company's premises.

  • Where: a controlled environment, the developer's site or a lab.
  • Who: employees acting as users.
  • Goal: find remaining bugs and usability problems while fixing is still cheap and private.
  • Feedback: direct. Developers can watch and fix quickly, and the product may still have rough edges and crash.

Alpha testing often combines black-box and some white-box activity, and it ends with a bug-fix cycle and a go/no-go to move on to beta.

Beta testing

Beta testing puts a near-final version in the hands of real users in their own environments: real devices, networks and data, with no developer in the room.

  • Where: the users' own sites and devices (a "field test").
  • Who: a limited group of volunteers, customers or the public.
  • Goal: learn how the product behaves in the real world, find compatibility and performance problems that a lab cannot reproduce, and measure satisfaction.
  • Feedback: collected through crash reports, surveys, support tickets, analytics and forums.

Beta programs come in two flavors. In a closed beta a selected group is invited; in an open beta anyone may join. Software published as "preview" or "early access" is a beta by another name.

Because developers do not control the environment, beta findings need good bug reports: what the user did, what device and version, and any log or crash data.

Alpha vs beta at a glance

Alpha Beta
Testers Internal employees External users, customers
Place Lab, company site Real-world environments
Environment control High None
Product stability Rough, may crash Mostly stable
Typical findings Functional bugs, usability, missed requirements Compatibility, real-data, performance, satisfaction
Duration Shorter, controlled Weeks, uncontrolled
Follows System testing Alpha testing
flowchart LR
  ST["System testing"] --> AL["Alpha<br/>internal users"] --> BE["Beta<br/>external users"] --> RC["Release candidate"] --> GA["General availability"]

Each step widens the audience and loosens control.

User acceptance testing (UAT)

User acceptance testing is the formal decision point: do the people who will use or pay for the product accept it? Unlike alpha and beta, which hunt problems broadly, UAT checks the product against business needs and acceptance criteria agreed earlier.

A typical UAT run:

  1. Business users or their representatives take the acceptance criteria and the key business processes.
  2. They design or review realistic scenarios (for example "an accountant closes the month end") using production-like data.
  3. They run them in a UAT environment, recording passes and failures.
  4. Defects are triaged: some must be fixed before acceptance, others become accepted known issues.
  5. The business signs off, or refuses and lists what must change.

QA often supports UAT by preparing environments, data and scenarios, but the verdict belongs to the business. Because the testers are not professional testers, clear steps and quick support matter.

Other flavors of acceptance testing

  • Operational acceptance testing (OAT) is done for the people who will run the system, such as operations and administrators: backup and restore, monitoring, installation, upgrade, failover, security patches and documentation. It complements reliability and recovery testing.
  • Contract acceptance testing checks the product against the criteria written into a contract or statement of work, typically in outsourced projects where payment depends on acceptance.
  • Regulatory (compliance) acceptance testing verifies conformance to laws and standards, for example in finance or health.

Contract and regulatory acceptance testing deserve a little more detail. For contract acceptance, the team lists the acceptance criteria from the contract, designs tests for each criterion (including integration and interoperability with the client's systems), runs them in the client's presence, and records signed results. For regulatory acceptance, testers map each legal or industry requirement (data protection, accessibility, financial reporting, medical device rules) to a test and keep evidence, because auditors may ask for it years later.

Entry and exit criteria

Every acceptance stage needs agreed gates so the team knows when to start and when it is done.

Typical entry criteria: system testing is finished with no open critical defects, the build is stable, the test environment and data are ready, acceptance criteria and scenarios are approved, and testers are trained.

Typical exit criteria: all planned scenarios were run, no open defects above an agreed severity, pass rate meets the target, known issues are documented and accepted, and a sign-off is recorded.

Try it: an acceptance gate

This stand-in evaluates exit criteria for a UAT cycle and prints a go/no-go with reasons. The numbers are invented.

JS A UAT exit-criteria checker

Press Run (or Ctrl + Enter) to see the console output here.

No output.

Fix the data, for example by setting the critical defect to accepted: true (in real life: by fixing it), and watch which rules flip. Notice that a 92.5% pass rate alone already fails the second rule.

Try it: summarize beta feedback

A beta produces a pile of unstructured reports. Tagging and counting them turns noise into priorities.

JS Triage beta feedback by theme and platform

Press Run (or Ctrl + Enter) to see the console output here.

No output.

This is exactly the kind of pattern a lab never shows: one OS version plus one feature.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Who typically performs alpha testing?

    Alpha testing is internal. It happens at the developer's site or in a lab, before the product is released to outside users.
  2. 2. Which finding is a beta test MOST likely to reveal that a lab test would miss?

    Beta testing runs the product on real devices, networks and data, which exposes environment-specific problems.
  3. 3. Who gives the final verdict in user acceptance testing?

    UAT is the business decision to accept. QA supports it, but the business signs off.
  4. 4. What does operational acceptance testing (OAT) cover?

    OAT confirms the system can be operated and maintained in production.

Summary

  • Alpha testing is internal and controlled; beta testing is external and uncontrolled, in real environments.
  • UAT lets the business check the product against acceptance criteria and sign off; OAT checks operability; contract and regulatory acceptance check agreements and laws.
  • Entry and exit criteria make "ready" and "done" explicit.
  • Beta feedback needs structure (tags, counts, platform) to become useful.

Next up: Ad Hoc, Monkey and Gorilla Testing.