Which object properties should I choose ?

Hi ​@Augustin,

This is a common challenge when working with A360 and web automation in general. The short answer is there’s no universal best property, because stability depends on how the website is built. However, there are some general guidelines and a rough ranking based on experience:

1. Element ID (id)

• If the site uses unique, consistent IDs, this is usually the most reliable.
• Many modern frameworks generate dynamic IDs, though, so check if they change on reload.

2. Name Attribute (name)

• Often stable for form fields and buttons.
• ARIA Labels / Accessibility Attributes
• These tend to remain stable because they’re tied to accessibility compliance.

3. Inner Text / Visible Label

• Works well for static text elements (buttons, links), but breaks if text changes (e.g., localization or minor wording changes).

4. CSS Class (with caution)

• Classes can be stable if they’re semantic (e.g., btn-primary), but many frameworks use dynamic class names.

5. XPath / DOM Path

• Least stable in most cases because even minor layout changes break it.
• Avoid absolute XPaths; use relative ones with stable anchors if you must.

Many websites use dynamic DOM structures (especially SPAs built on React, Angular, or Vue).
Paths often change when developers add wrappers, ads, or restructure HTML—even if the visual design looks the same.

Best Practices

• Prefer semantic attributes (ID, name, aria-label) over structural ones (XPath).
• Use multiple properties for resilience (AA allows combining attributes).
• Fuzzy matching or wildcards for dynamic parts (e.g., id="btn_*").
• If the site is highly dynamic, consider Computer Vision + Anchor-based approach or API integration instead of DOM selectors.
• Avoid using the property Path in any case

Yes. Some sites have stable IDs; others don’t. For enterprise apps, IDs are usually stable. For consumer-facing sites, expect dynamic attributes.

Hello ​@Augustin ,

Like ​@Padmakumar, I almost always rely on the ID and Name attributes first. My next choice is a relative XPath, and combining these—ID/Name plus relative XPath—has consistently given me the strongest stability. I always inspect the element to confirm the correct XPath and that the Name attribute is unique before finalizing the capture.

One of my most stable bots uses only these two or three properties across all its capture actions. I also find Inner Text to be reliable.

The property I always uncheck is the Path, since it tends to change frequently - as you have now also experienced, unfortunately.