Summary
GraphComment counts one consumption unit when it loads comments to display.
• In List mode (chronological), 1 load = 1 batch of comments loaded.
• In Bubble Flow mode (relevance sorting), 1 load = 1 discussion loaded (a top-level comment + its full reply tree), regardless of the number of replies.
Definitions
Comment batch (List mode): a group of comments returned in a single response (e.g., 5, 10, 20, 50, 100…), based on your settings in the GraphComment admin.
Discussion (Bubble Flow mode): a top-level comment (a direct reply to the article) and all nested replies under it (the full thread/tree).
List mode (chronological): how loads are counted
In “List” mode, GraphComment loads comments in batches. Each time a batch is requested and returned, it counts as 1 load. A single page can therefore generate multiple loads if the user loads additional batches or refreshes/re-sorts the view.
Examples of actions that count as 1 load (if they trigger a comment load):
• Opening the comments module on a page (first batch).
• Clicking “Load more” / “See more” (next batch).
• Lazy-loading (if enabled) that automatically loads the next batch when the user scrolls down.
• Changing the sort mode (relevance / newest / oldest).
• Refreshing / checking for new comments.
• Displaying the “Top widget” on a page (it triggers a load when it is displayed).
Batch size: we do not differentiate
Whether your batch contains 5, 10, 20, 50, 100 comments (or all comments), we still count 1 load when that batch is loaded. This choice makes consumption easier to understand and better reflects the cost logic: the main unit is the batch load itself (request, processing, moderation rules, delivery/transfer).
You can trade off showing more at once (heavier on low-powered devices) versus showing fewer for faster display, depending on your community’s needs.
Bubble Flow mode: how loads are counted
Bubble Flow organizes the display around discussions (threads): top-level comments (direct replies to the article) and their nested replies.
In this mode, GraphComment counts loads as the number of discussions loaded:
• 1 load = 1 discussion (1 top-level comment + all its replies).
• Replies (level 2, 3, etc.) are not counted individually: they are part of the discussion.
• If an action loads N discussions at once, consumption increases by +N (not +1).
Example
A discussion can contain 500 comments in its tree (1 top-level comment + 499 replies). In Bubble Flow counting, this still counts as 1 load, because it is a single discussion.
Why count by discussion in Bubble Flow?
Counting “per comment” would penalize rich discussions (many replies) and make consumption less predictable. Counting per discussion is more stable and closer to how Bubble Flow is structured.
Why GraphComment is not billed “per comment”
Counting the exact number of displayed comments may sound intuitive, but it creates several issues:
• A very active discussion could make consumption spike without reflecting “abnormal” usage.
• Billing would become unpredictable (a thread can grow quickly).
• It would encourage limiting replies visibility, hurting engagement.
Our goal is a simple, stable unit that correlates with the work required to deliver comments on screen.
Short FAQ
Are my site pageviews equal to GraphComment loads?
No. A load corresponds to loading comments (a batch or a discussion depending on the mode), not a site pageview.Why can a single page generate multiple loads?
Because a user can load multiple batches (e.g., “Load more”, lazy-load), change sorting, or refresh the view.Does the offset/batch size affect counting?
Indirectly, yes: in List mode, a larger batch size usually requires fewer batches to browse the same amount of comments. However, each loaded batch still counts as 1 load, regardless of size.In Bubble Flow, do you count each comment?
No. In Bubble Flow, we count discussions (top-level comments). One discussion and all its replies count as 1 load.
Consumption tracking
You can track your consumption in the “Statistics” section of your GraphComment admin, as well as on your dashboard and invoices. Updates are near real-time (typically within a few minutes).