How to monitor brand mentions on X (and across the web)

A brand mention is any public reference to your brand — your name, a product, a founder, your @handle — whether or not it tags you or links to you. The ones that tag you are the easy 20%. The mentions that move reputation are the ones where someone discusses you without the @, the link, or the spelling you would have searched for.

That is the whole reason monitoring is harder than it sounds. "Did anyone @ us" is a notification you already get. "Is anyone talking about us right now in a Reddit thread, a quote-post, or a review without using our handle" is the question monitoring exists to answer. Build for the untagged mention and the tagged ones come along for free.

Resist the urge to track everything. Five categories cover almost every real need, and teams that try to monitor more than this usually end up checking none of it.

1. Your brand name and its variants. The exact name, the one-word and two-word spellings ("ReplySocial" and "Reply Social"), the obvious misspellings, and your @handle minus your own posts. This is the highest-volume and noisiest monitor for most brands.

2. Products and features by name. People complain about and recommend specific features without ever naming the company. If a feature has its own name, it needs its own monitor.

3. Founders and public-facing people. For most small brands the founder is mentioned more than the company. If your name is on the homepage, it belongs in a monitor.

4. Competitors. Mirror categories one through three for your top one to three rivals. The payoff is intent — people saying "switching from [competitor]" or "[competitor] alternative" are the highest-converting mentions you will ever see, and they never tag you. There is a full playbook for tracking competitors on X if that is the angle you care about most.

5. Intent phrases. "Looking for a [category] tool," "any recommendations for," "alternative to." These are pure pipeline. They are also where the broader online brand monitoring guide spends most of its time, because they convert better than any branded term.

Here is the part most articles bury under a thousand words of "why mentions matter." The setup is five steps, and you can run all of them today.

1. Build the term list. Write out every term from the five categories above. For each one, note the spelling variants and whether you want exact-phrase matching. This list is the spine of everything else — get it right once and the rest is plumbing.
2. Pick the channels that matter for you. Not all five places are equal for every brand. B2B SaaS lives on X, Reddit, and review sites; a consumer product lives on Instagram, TikTok, and Reddit. Choose the two or three where your customers actually talk before you touch the long tail.
3. Set up a monitor or alert per channel. Google Alerts for news and blogs, an X advanced-search query for social, a Reddit feed for Reddit, manual weekly checks for review sites at low volume. The per-channel playbooks are below.
4. Route the results to where you will see them. A monitor nobody reads is theater. Send alerts to a shared inbox, a Slack channel, or a tool that holds them in one queue. The goal is a single place you triage every morning, not five tabs you forget to open.
5. Filter the noise before you triage. This is the step that decides whether the habit survives past week two. On X especially, raw mention searches are full of bots and scams. Filter them out first — covered in its own section below — or the queue becomes unusable and you quietly stop checking it.

The whole thing is a morning of setup and ten minutes a day after that. The teams that get value are not the ones with the most expensive tool; they are the ones who built step four into a daily habit. Our social media monitoring guide goes deeper on the triage workflow once volume climbs.

X is the channel where the untagged mention problem is worst and the tooling is best, because X advanced search is genuinely powerful if you know the operators. Start here.

Catch tagged mentions, minus your own noise. The base query is @yourhandle -from:yourhandle — every post that tags you, excluding your own replies. That alone beats the native notifications tab, which collapses and hides things.

Catch the untagged ones. Add brand-name variants in an OR group: (@replysocial OR "ReplySocial" OR "Reply Social") -from:replysocial. Now you see the conversations where someone names you without tapping the @ — which is most of them.

Control the volume. Two operators do the heavy lifting: min_faves:5 surfaces only posts that gained some traction, and since:2026-06-01 until:2026-06-30 bounds the window. Wrap multi-word terms in quotes so "product hunt" does not match every post containing both words separately.

Save the query and check it daily, or hand it to a tool that polls it for you. ReplySocial's X mention tracking runs exactly these patterns on a schedule and scores every reply author for bot likelihood, and the keyword monitoring feature handles the intent-phrase queries that have no handle to anchor on.

Reddit is the channel everyone says to monitor and nobody explains how to. It matters more than its size suggests: Reddit threads rank for category queries in Google, and AI engines cite Reddit more than almost any other source — so a Reddit thread about you is read by both prospects and ChatGPT.

The free path is two moves. First, a Google query restricted to the site: site:reddit.com "Brand Name", run weekly or saved as a Google Alert. Second, Reddit's own RSS feeds, which most people do not know exist: append .rss to a search or subreddit URL — for example reddit.com/search.rss?q="your+brand"&sort=new — and pipe it into a reader. Reddit's native search is weak; the RSS-plus-Google-site combination is what actually works for free.

Where the free path breaks is speed and dedup — RSS does not remember what you have already seen, and a complaint thread can climb a subreddit before your weekly check runs. If Reddit is a real channel for you, the Reddit monitoring workflow is worth automating rather than babysitting.

Here is the part the other guides skip, and it is the part that quietly kills most monitoring habits. On X, your mention queue is not just real people. X officially claims under 5% of accounts are spam, but independent 2025 research puts bots at 9–15% of all accounts and 15–44% inside political and entertainment conversations. On a thread that is starting to trend, a real slice of the replies are automated — AI-written flattery, scam DMs in disguise, and rapid-reply farms boosting engagement.

Unfiltered, that noise does two things. It buries the real mentions you needed to see, and it makes the queue exhausting enough that you stop opening it. Checking account age, follower ratio, and bio by hand works for one account; across a full queue it eats 30–40% of your monitoring time, which is the honest reason most teams either stop filtering or stop monitoring.

When we built BotBlock, the scoring engine behind ReplySocial, this was the whole point: score every reply author across 30-plus signals — profile, writing, timing, and scam patterns — and filter the bots out before they ever reach your queue. If you want the manual version, our guide on spotting a Twitter bot breaks down the signals, and the free bot checker scores a single handle in a couple of seconds. The same noise is what turns a normal mention spike into bot-amplified outrage, which is why filtering is a monitoring problem, not just a cleanliness one.

The newest mention surface is not a website at all. When someone asks ChatGPT, Perplexity, Claude, or Google's AI Overviews about your category, the answer is a kind of mention — and it is shaped by the public web, not your homepage. If the engines recommend a competitor and skip you, that is a mention problem you cannot see in any social tool.

Monitor it the same way you monitor anything else: with a repeatable check. Build a short list of buyer-intent prompts — "best [category] tool," "alternative to [competitor]," "is [your brand] worth it" — and run them through each engine on a weekly cadence. Track which brands get named, whether the engine cites your domain, and what context it includes. Tools like Profound, Otterly.AI, and Ahrefs Brand Radar automate this; for a small brand, a weekly manual prompt rotation works. The full setup lives in the brand monitoring guide.

Free monitoring is real and worth starting with. The honest version: it works at low volume and breaks in three specific places.

The free stack. Google Alerts for news and blogs, X advanced search for social, Reddit's RSS feeds for Reddit, Talkwalker Alerts as a broader Google Alerts competitor, and a weekly manual pass on your review sites. For under roughly 100 mentions a month, this genuinely covers it.

Where it breaks. Three things: cross-platform aggregation (you are juggling five tabs instead of one queue), persistence (nothing remembers what you have already triaged, so you re-read the same mentions), and noise filtering (no free tool scores bots out of your X results). Each one costs minutes a day until the minutes add up to a paid tool being cheaper than your time.

The paid options. Broad open-web tools — Brand24, Awario, Mention, BrandMentions — run roughly $99–249/month and cover social, news, and blogs with history and sentiment. Enterprise platforms with full web, news, and broadcast scope (Brandwatch, Meltwater, Talkwalker) start in the four figures monthly. For social-first monitoring with reply velocity and bot filtering built in, ReplySocial Pro is $25/month flat with a real free plan, and a free share of voice calculator if you just want to size the conversation first. Most teams overspend by buying broader scope than they use — see the pricing breakdown before you commit.