Key takeaways
- Exploratory testing is learning, test design and test execution at the same time. Each result shapes your next test.
- It's not random clicking. Good exploration has a goal (a charter), a time limit and notes.
- It finds the problems scripts miss, because nobody can script what they haven't imagined.
What exploratory testing is
In scripted testing, you design test cases first and run them later. In exploratory testing, you do both at once.
You try something, watch what happens, and let that guide your next test. "That was slow, what if I add 100 items?" "The error message changed, what if I go back and resubmit?"
Testing expert Cem Kaner coined the term in the 1980s. James Bach later gave the short version still quoted today: simultaneous learning, test design and test execution.
Why it finds different bugs
Scripted tests check what someone expected. They're great for confirming known behaviour and for regression.
But many serious bugs come from combinations nobody predicted: unusual sequences, interruptions, odd data, real-world habits. Exploration is how you find those, because you follow what the software shows you.
In January 2019, 14-year-old Grant Thompson was setting up a group FaceTime call with friends. While a call was still ringing, he added another participant, and found he could hear the other person's microphone before they had answered.
His attempts to report the problem didn't get quick attention. Once videos showing the steps spread online, Apple quickly disabled Group FaceTime. A fix shipped on 7 February 2019, and Apple's security notes credited Grant Thompson and another reporter, Daven Morris.
Lesson: the bug appeared only with an unusual sequence of actions during a ringing call. That's hard to predict in a script and natural to find when you ask "what if I do this while that is happening?"
How to explore with focus
1. Write a charter
A charter is a one-line mission for your session. Elisabeth Hendrickson, in Explore It!, suggests this template:
Explore (target) with (resources) to discover (information).
For example:
- Explore the checkout with expired and declined test cards to discover how payment errors are handled.
- Explore the profile page with very long names and non-Latin characters to discover display and saving problems.
2. Set a time box
Explore in sessions of about 60 to 90 minutes, without interruptions. This approach, called session-based test management, was developed by Jonathan and James Bach. See Session-based test management for how teams use it at scale.
A fixed time stops you from wandering off forever, and makes exploration easy to plan and report.
3. Take notes as you go
Write down what you tried, what you noticed and what you want to try later. Short notes are enough:
- Tried 3 declined cards: same generic error each time. Question: should users learn why?
- Back button after payment shows the form again. Pressed Pay again: second order created! Bug.
4. Debrief
At the end, summarise: what you covered, the bugs found, open questions and ideas for the next session. Share it with the team in five minutes.
Ideas to explore with
When you don't know what to try next, these questions help:
- Interruptions: what if I lose the connection, lock the phone or open a second tab halfway through?
- Sequences: what if I do the steps in a different order, go back, or repeat one?
- Data: empty, very long, special characters, zero, negative, a date in the past?
- Users: what would an impatient user do? A first-time user? Someone trying to cheat?
- Timing: double-click, slow network, two people editing the same thing?
Do
Combine both approaches. Use scripted or automated tests for known, important behaviour, and exploratory sessions for new features and risky areas.
Don't
Treat exploratory testing as "what we do when there's time left over". Some of the most valuable bugs are found there.
Check yourself
Write a charter for exploring a new password reset feature.
Show answer
One example: "Explore the password reset flow with expired links, reused links and multiple reset requests to discover security and usability problems." A good charter names the target, the resources or angles to use and what you want to learn.
