Visual Testing FAQ: Answers to the 20 Most Common Questions

Have questions about automated visual testing? Here are the answers to the questions we receive most often, organized by topic.

General Questions

1. What is automated visual testing?

Automated visual testing (also known as visual regression testing) is a technique that automatically compares screenshots of your application to detect unintended visual changes.

The steps of a visual test are as follows:
Simplified process:

1. Capture a reference screenshot (baseline)
2. Take a new screenshot after code changes
3. Pixel-by-pixel comparison
4. Alert if a difference is detected

2. What is the difference between visual tests and E2E tests?

Ideal approach: Combine both for complete coverage.

3. Do I need to know how to code to run visual tests?

The answer depends entirely on the tool you choose. Most solutions on the market require development skills, which creates a significant barrier to entry for QA teams.

Code-based tools (Playwright, Cypress) offer great flexibility, but they require tests to be written, maintained, and debugged by developers. This means QA teams depend on developers to create or modify a scenario, creating a bottleneck in the testing process.

SaaS solutions like Percy or Applitools simplify certain steps but still require integration into the source code and technical configuration.

Delta-QA takes a different approach: its graphical interface allows any team member -- QA, product manager, or designer -- to create, run, and maintain visual tests without writing a single line of code. Scenarios are built visually, baselines are managed in a few clicks, and results are immediately readable. This empowers QA teams to be autonomous and no longer depend on developer schedules for their testing campaigns.

4. How long does it take to set up visual tests?

5. What types of bugs do visual tests detect?

Visual tests can detect:

• Broken layouts (CSS)
• Mispositioned elements
• Truncated or overflowing text
• Incorrect colors
• Missing or improperly sized images
• Responsive design issues
• Unloaded fonts
• Incorrect z-index (overlapping elements)
• Wrong margins/padding
• Regressions after dependency updates

Technical Questions

6. What is a baseline?

The baseline (or reference) is the "correct" screenshot against which future captures will be compared.

The lifecycle of a baseline follows four stages:

1. First run -- the baseline is created automatically
2. Subsequent runs -- each capture is compared to the baseline
3. Intentional change -- the baseline is updated to reflect the new version
4. Bug detected -- the code is fixed, the baseline remains unchanged

7. How do I handle dynamic content?

Dynamic content (dates, avatars, ads) causes false positives. Three main solutions:

• Exclusion zones: Mask dynamic areas (dates, counters) during comparison so they are ignored
• Content mocking: Freeze dynamic data (fixed date, default avatar) to get identical captures on every run
• CSS masking: Hide dynamic elements during capture using an injected stylesheet

8. What tolerance should I use?

9. How do pixel-by-pixel comparisons work?

The basic algorithm works in five steps:

1. Load the baseline image (reference)
2. Load the current image (test)
3. For each pixel, compare the color values (R, G, B, A) and mark as different if the deviation exceeds the threshold
4. Calculate the percentage of different pixels
5. If this percentage exceeds the defined tolerance, the test fails

Advanced algorithms (pHash, SSIM) add perceptual tolerance that more closely approximates human vision.

10. Can I test across multiple browsers?

Yes, most tools allow you to test across multiple browsers simultaneously. You simply need to configure the target browsers (Chrome, Firefox, Safari) in the tool's settings.

Important: Each browser produces slightly different rendering, so you need to maintain separate baselines per browser.

11. How do I test responsive designs?

Simply define the viewports to test and run captures for each one:

Each viewport generates its own baselines, allowing you to detect issues specific to each screen size.

Tool Questions

12. What are the main visual testing tools?

13. How do I choose the right tool?

14. How many visual tests do I need?

Rule of thumb: Start by covering the critical user journeys -- the pages seen by 80% of your users, conversion flows (checkout, signup), and high-business-value features.

15. Who should manage visual tests?

Practical Questions

16. How do I update baselines?

Updating baselines varies by tool:

• With a graphical interface (Delta-QA): Simply click "Accept as new baseline" for each change
• With a command-line tool: A dedicated command lets you regenerate all baselines in a single run

Important: Always review visual changes before accepting a new baseline to avoid inadvertently approving a bug.

17. How do I manage baselines as a team?

Four best practices for managing baselines as a team:

1. Version baselines with your code -- commit them in the same repository to maintain consistency between your code and its visual references
2. Work with branches -- each feature branch has its own baselines, avoiding conflicts with the main branch
3. Review baseline changes -- treat baseline modifications like code: include them in pull requests for validation
4. Assign an owner per area -- designate an owner for each set of tests to avoid update conflicts

18. Can I test applications with authentication?

Yes, several approaches are available:

• Programmatic login -- the test logs in automatically before each capture by filling out the login form
• Saved authentication state -- the session state is recorded once, then reused for all subsequent tests, which significantly speeds up execution
• Staging test token -- in test environments, a dedicated authentication token bypasses login without compromising security

19. Do visual tests replace manual testing?

No, they complement it:

Automated visual tests excel at detecting regressions, comparing against a known reference, and providing fast, repeatable checks. However, they cannot detect user experience issues or verify whether a design matches the designer's "intent."

Manual testing remains necessary for exploring new features, usability testing, complex edge cases, and validating the overall feel of the application.

20. What are the future trends in visual testing?

Artificial intelligence is undeniably the strongest trend in the QA market today. However, an important nuance deserves attention: the non-determinism of AI agents can conflict with the core mission of QA engineers, which is to guarantee with certainty that an application behaves correctly.

A regression test must be reproducible and predictable. If a non-deterministic AI is integrated directly into the visual regression detection process, it introduces a variable of uncertainty precisely where reliability is needed most. Test results could vary from one run to another, which is incompatible with the absolute confidence required by a deployment pipeline.

The real trend, in our view, is to use AI upstream: not to execute tests, but to improve the algorithms of the tools available to engineers. In other words, AI should help build even more deterministic, even more reliable testing software that supports QA teams with the certainty that the deployed software will be high quality.

This is exactly Delta-QA's philosophy. Rather than embedding AI in the test loop, Delta-QA invests in the robustness of its comparison algorithms. Thousands of combinations of HTML page configurations -- complex structures, deep nesting, dynamic content, cross-browser rendering variations -- are systematically tested to strengthen detection reliability. Each version of the algorithm is validated against a stress test matrix covering over 425 real pages, with a clear goal: zero false positives, zero false negatives.

This approach guarantees QA engineers a tool they can rely on for every deployment, with no surprises and no uncertainty.

We build Delta-QA, a visual regression testing tool. Always open to feedback from the community!