FEBRUARY 17, 2026Production Feature

Sentiment Attribution Engine: Know Exactly Why Your Reputation Changed

Sentiment Attribution shows which forums, topics, author groups, and time periods drove a reputation shift so teams can act on the real cause.

Know the cause, not just the change

A sentiment drop is only half the story. This engine is built for the next question: what actually caused the movement, and where should the team look first?

Net sentiment shift

-15%

current-period change versus the prior comparison window

Top source

Rollitup

responsible for 60% of the total decline

Top issue

LED drivers

topic cluster driving 25% of the negative change

Positive offset

+0.02

THCFarmer partially counteracting the drop

Four ways to find the why

Attribution works because it lets you decompose the same shift from multiple angles instead of forcing every problem into one view.

Forums

Which communities moved the score

See whether the change came from one forum, several communities, or a spread that signals a wider reputation shift.

Topics

Which issues carried weight

Break sentiment change down by subject so you can tell whether product quality, support, shipping, or some other topic is driving it.

Author cohorts

Which audiences reacted

Separate new growers, returning voices, and high-volume posters so you know whose opinion is actually moving the brand.

Timeline

When the shift happened

Follow the curve over time to see whether the change arrived suddenly or built gradually across the reporting window.

What appears on the page

The analysis flow is designed to move from headline movement to supporting detail quickly, so you can stop at the top factor or keep drilling as needed.

Period comparison

The page starts by comparing current and previous sentiment so you can see the size of the move before reading any explanation.

Contribution view

Each segment gets a contribution score showing how much it pushed sentiment up or down rather than just how often it appeared.

Top factors

The three biggest drivers are surfaced in plain language so you do not need to interpret the whole table before acting.

Drilldown table

A sortable detail layer shows contribution share, volume context, and segment performance so the explanation stays defensible.

Evidence context

Post counts and linked source views keep the analysis tied to real conversation instead of turning attribution into a black box.

How to work the breakdown

The best use of attribution is disciplined and quick: find the biggest contributor, validate it, then move to the underlying evidence.

01

Compare the current period against the previous one so the overall shift is explicit.

02

Switch between forums, topics, author cohorts, and timeline to see which dimension explains the movement best.

03

Read the top drivers and drilldown rows to find the specific segment pulling the score up or down.

04

Jump into Mentions or related dashboards to validate the explanation and decide what action comes next.

How teams use it day to day

Attribution is most useful when it shortens the path from “something changed” to “here is the exact thread, topic, or audience we need to examine.”

When sentiment drops

Find the forum, topic, or author group carrying the biggest share of the decline before you start triaging blindly.
Use the explanation to separate isolated noise from a real pattern that deserves support, product, or community follow-up.
Move directly from attribution to raw mentions so the team can confirm what the model is saying with source evidence.

When sentiment improves

See which communities and topics are contributing positively so you know where brand strength is actually coming from.
Validate whether a launch, support change, or new message is showing up in the conversation as intended.
Use attribution as the layer that turns good news into something repeatable instead of accidental.

Open Analytics

Sentiment Attribution is available on the Analytics page when there is enough comparison data to support a real reading. If the sample is too thin, VueLeaf tells you that directly instead of pretending certainty.

Use this view when the score moves enough to raise a question. The bigger the change, the more valuable it becomes to know exactly which slice of the conversation caused it.
Go to Analytics