The five stages
Tool-seeking. An explicit request for a product or vendor recommendation. The clearest buying signal that exists in public, and the one worth replying to first.
Competitor mention. Someone naming an alternative with a complaint attached. They already have budget and a use case; they are telling you precisely what would make them switch.
Pain venting. A real problem described without an explicit ask. Not a lead today, but this is where tool-seeking comes from three weeks later.
Research. Forming an opinion before knowing what to buy. A genuinely useful answer here makes you the name they remember when they start shopping.
Unrelated. A keyword match that turns out to mean nothing. In real monitoring this is the overwhelming majority of matches, which is the entire argument for scoring.
What this demo cannot do
This page matches a few dozen regular expressions against phrasing. That catches the obvious cases and misses everything else.
It has no idea what your product does, so it cannot judge relevance — only shape. A post can be perfect tool-seeking for a category you have nothing to do with, and this tool will still call it high intent. It cannot read sarcasm, cannot follow a thread's context, and cannot tell that "we solved this internally last year" means the opportunity is gone.
MentionSpot works differently: every candidate match goes through a full language-model pass that reads the whole post against a description of your product, which is why it can correctly classify a post that uses none of the phrases this demo looks for.
Using intent stages in practice
The point of classification is triage, not reporting. A queue sorted by intent stage answers the only question that matters on a Monday morning: which three of these do I reply to before lunch.
A workable policy: reply to every tool-seeking and competitor-mention thread within a few hours, since those decay fast. Reply to pain and research threads selectively, when you have something genuinely useful to add. Let the unrelated bucket accumulate, and review it monthly — if it is growing, your keywords are too broad.
FREQUENTLY ASKED QUESTIONS
Questions about intent classification
Is this the model MentionSpot uses?
No. This is pattern matching, included as a public illustration that needs no sign-up. The product runs a language model over the full text of each match against your product description.
Why was my sentence marked unrelated?
The demo only recognises a limited set of phrasings. A post a human would read as high intent falls through to unrelated if it does not contain one of them, which is exactly the weakness that makes pattern matching unsuitable for real monitoring.
Which stage converts best?
Tool-seeking and competitor mentions convert fastest, because the person is already evaluating. Research and pain signals convert more slowly but are less contested, since most vendors ignore them.
Does engagement matter as well as intent?
It matters for prioritisation rather than qualification. A high-intent thread with recent activity is worth replying to before an equally qualified one that has gone quiet.
SCORE EVERY MATCH, NOT ONE
Real monitoring means thousands of these a month.
MentionSpot reads and scores every Reddit and X match against your actual product, then hands you the short list worth acting on.
Get the 7-day pass