# Brand reputation monitoring: what to watch and when Brand reputation monitoring as a program: seven surfaces, free tools, thresholds, owners and a weekly, monthly and quarterly routine. --- Brand reputation monitoring is a standing routine that checks the places where people form an opinion of your company, on a fixed schedule, against thresholds you wrote down in advance, with a named owner for each kind of finding. Seven surfaces matter: branded search results, autocomplete, branded search demand, reviews, news and social mentions, AI answers, and public registers and filings. Most of it can be done with free tools and first-party reports. The usual failure is not a missing tool. It is a feed nobody owns: alerts arrive, a few get read, and the one that matters sits in a shared inbox for weeks. A program fixes this by deciding three things first: what counts as a signal, how soon someone must look, and who decides what happens next. This guide is for the person who runs marketing, communications or operations and needs a routine that fits into a few hours a month. It covers what to watch, how often, when to escalate and what a realistic first quarter looks like. It does not teach alert setup in detail, metrics scorecards, paid software or a full AI answer audit; those are separate jobs. Tool behavior is described as the help pages read for this guide state it, because settings change.

The short version

Write a register of surfaces first: for each one, the source you will check, the cadence, the threshold that triggers action and the person who owns the response.
Use what is free and first-party: Google Alerts for web mentions, Search Console for branded search demand, your Business Profile for reviews, and the public register that holds your company filings.
Check by cadence: weekly for reviews, news and mentions, monthly for the brand search page and autocomplete, quarterly for AI answers, registers and the thresholds themselves.
Write thresholds as observable events, not feelings, and time-box the response to each one.
Expect the first quarter to be about tuning. Noise falls and thresholds get sharper by week eight or so.

## What brand reputation monitoring covers and where it stops Monitoring finds changes sooner than you would by chance. It does not judge them, correct them or decide whether they are true. Those steps belong to the people you name in the register, and they use other routines: a reply to a review, a correction request to a publisher, a legal letter, a better owned page. Two boundaries keep the program small enough to survive. Monitor your brand, not your industry: a competitor's pricing change is market research. And monitor for change, not totals: the new one-star review that cites a safety problem matters, a raw review count does not. Accurate criticism is not a monitoring failure either. If a complaint is true and public, fix the cause and reply honestly. Monitoring that exists to find things to hide tends to produce a second story. ## Seven surfaces and who owns each Start with a register: one row per surface, with the free source, the cadence and an owner. The register is the program. Everything else in this guide fills in its cells.

Surface	What you watch	Free source	Usual owner
Branded search page	The top ten results and any new URL for your exact brand name	Private browser window, saved log	Head of marketing
Autocomplete	Words that appear after your brand name	Search box, logged monthly	Head of marketing
Branded demand	Impressions and clicks for queries containing your brand	Search Console branded filter	Web or SEO lead
Reviews	New ratings, new text, replies, flagged items	Business Profile and each review site you use	Customer lead
News and mentions	New articles, posts and threads naming you	Google Alerts, manual search of social platforms	Communications lead
AI answers	What assistants say about you on a fixed prompt set	The assistants themselves, logged by hand	Web or SEO lead
Registers and filings	Changes to officers, address or filings on the public register	The register's own follow or alert service	Finance or legal lead

Name one person per row, not a team, because a team row is the one nobody checks. The owner watches the surface; a safety allegation still goes to operations and counsel. ## Branded search results and autocomplete The page that appears when someone types your company name is the surface a banker, buyer or journalist sees first, and it usually moves slowly enough for a monthly check to catch the changes that matter. The method is plain. Open a private window, search your exact brand name and your name plus one or two words customers use, and record the top ten URLs with their titles in a spreadsheet. The point of the log is comparison: this month against last month, not a score. Note any new URL, any URL that moved by three or more places, and any result whose title or snippet states something you did not publish. Results differ by place, device and search history, so a private window is only a rough baseline. Google's Search Console Performance report help says as much: search results are specific to the time, place, device and recent history of the person searching, so a query in the report may not show your site when you run it yourself. Treat your log as a sample and run it from the same setup each time. If you want a quick self-check of how a name page looks, our free page one score tool is built for that first look. ### Autocomplete: the words people add to your name Autocomplete shows what people are searching for next to your name. Google's autocomplete help page says predictions are based on real searches that match what you type, taking into account the language, the location, trending interest and your past searches. So what you see is partly yours: a signed-in browser with your history will differ from a stranger's. Log it signed out, in a private window, and note the language and region you used. What you watch is the modifier. A new word after your name, such as a word about refunds, delivery, a lawsuit or a safety issue, tells you what a group of people has begun to ask. It is an early signal about demand, not proof that a claim is true. Check it two or three times over a week before you escalate, because a trending query can fade. The same page says Google does not allow predictions that link named individuals to potentially disparaging or sensitive terms, and that its enforcement teams remove the specific prediction and closely related variations. It points to another help article for reporting, which this guide did not read. The policy is written about named individuals; the page does not say a company name gets the same protection. So for a brand-name prediction, treat reporting as uncertain, read Google's current reporting article before filing, and expect the more durable fix to be a clear owned answer to the question people are asking. ### Branded demand in Search Console If your site is verified in Search Console, its Performance report shows clicks, impressions, click-through rate and average position, and per Google's Performance report help it opens on the past three months. Google's Search Console help on common tasks says you can filter queries to Branded or Non-branded. Branded queries include your brand name, domain or brand-specific products, including common misspellings. The page also says the filter provides data starting from March 11, 2025 and is not available for sub-properties or sites with few impressions. Use it as a trend line. Compare the last four weeks with the previous period, sort by the Difference column and look at what moved. The same page notes that shifts can reflect changes in user sentiment, news events or improvements to competing sites, not only your own changes. A drop in branded impressions after a news item may mean people are searching for you with different words. A rise in a query with a new modifier is the autocomplete check seen from the other side. If you want to group similar phrasings, the page says regular expressions can match several similar queries. ## Reviews: a weekly check on a surface that does not warn you Reviews are the surface where speed matters in days, not weeks, because a rating can affect the next customer's decision at once. Read every review source weekly and have a rule for what happens next. For Google Business Profile, Google's help page on reading and replying to reviews says you must first verify your business, then use Read reviews and the Reply icon. It says replies are checked against content policies and usually take up to 10 minutes to review, but sometimes up to 30 days. It also says reviewers are notified when you reply and can still change their review afterward. The page does not say whether you receive a notification when a new review is posted. So do not rely on an email arriving. Put the profile on the weekly calendar and read it. ### Reply, report or escalate Sort every new review into one of three paths. A genuine complaint gets a reply and, if it exposes a cause, a ticket for operations. A review that appears to break Google's content policies can be reported. A review that makes a serious factual or legal allegation goes to the escalation owner before anyone replies, because a hurried public answer can restate the allegation. For reporting, Google's help page on flagging reviews explains the steps. In the Business Profile you select Read reviews, click Report next to the review, choose a reason and send the report. It says only reviews that violate Google policies can be removed, and that Google will not act on reviews you merely dislike or disagree with. Evaluation typically takes several days. In the Reviews Management Tool the statuses are Decision pending, Report reviewed with no policy violation, and Escalated, which means an appeal is under review. You get one appeal per review, and an appeal can include up to 10 reviews. Build those timings into your thresholds: a reported review is checked again after a week, not the next morning, and an appeal is filed once, with the evidence ready. Your other review sources, from sector sites to app stores, have their own rules, and this guide did not read them. Add each to the register with the same three columns: where, how often, who. ## News, social mentions and filings ### Google Alerts for news and mentions Google Alerts is free and covers most of what a brand needs from news and web mentions. Google's Google Alerts help says you enter a topic, click Create Alert, and can open Show options to set how often you get notifications, the types of sites to include, the language, the region, how many results to see and which account receives the alert. The help page says you receive emails whenever Google finds matching search results. It does not document limits on the number of alerts or detail on which sources are covered beyond the site-type setting, so do not treat alerts as a complete net. Keep the set small and named by purpose: your exact brand name, its common misspellings, your domain, and your brand next to a handful of risk words. The page lists troubleshooting for missing alerts (the account, the alert settings, the spam folder), so test once a quarter that they still arrive. Social platforms' search and notification features differ and change, and this guide did not read them. A practical minimum is a saved search of your brand name on each platform where your customers talk, read weekly. Posts that name you with a verifiable claim belong in the log; general sentiment does not. ### Registers and filings The least watched surface is often the one a due diligence reader checks first. A company register shows the officers, the registered address and what has been filed. A change you did not make can signal an error or a mix-up with a similarly named company. In the United Kingdom, Companies House offers a free Follow service. The GOV.UK announcement says you sign in, search for the company, choose Follow this company, and receive an email alert when a filing is accepted, with a link to the filing history. Follow your own company and check that the alert reaches a shared mailbox, not one person. Registers elsewhere vary, and this guide did not read their alert services, so check whether yours offers a comparable notice. If it does not, put a quarterly manual read of the register entry on the calendar. ## AI answers: a quarterly sample, not a daily feed AI assistants answer questions about companies directly, so the program needs a light check of what they say. Keep it quarterly unless something is happening, because answers vary from run to run. Use a fixed set of five to eight questions a prospective customer or investor would ask, put them to two or three assistants, log the answer and the sources cited, and compare with last quarter. Look for claims that are false, out of date or attributed to the wrong company. Details of cadence, sample size and alert thresholds for this surface are covered in a dedicated guide on monitoring what AI says, so they are not repeated here. If an answer states something false, the owner traces it to the page behind it, because a correction at the source is more durable than a message to the assistant. ## Thresholds and escalation: decide before you look A threshold is a sentence that begins "if we see this" and ends with a person and a time. Without it, every finding is a debate. The table below gives starting points; they are our practitioner judgment from this kind of work, not rules from a platform, so tune them to your size and exposure.

Signal	Threshold to act	Owner	Response window
New review with a serious allegation	Any review citing safety, fraud, discrimination or a legal claim	Customer lead, then counsel	Same business day, no public reply until agreed
New review that looks like a policy breach	Spam, off-topic, a conflict of interest or content aimed at a person	Customer lead	Report within two days; recheck after a week
Rating drop	Average falls by a set amount over four weeks	Customer lead	Cause review at the next monthly meeting
New URL on the brand search page	Any new result in the top ten that makes a claim about the company	Head of marketing	Assessed within two business days
Autocomplete modifier	A new risk word persisting across two checks a week apart	Head of marketing	Owned-content response drafted within two weeks
News item with a factual error	A wrong figure, name, date or attribution about you	Communications lead	Contact the publisher the same day
Register change you did not make	Any filing, officer or address change you cannot explain	Finance or legal lead	Same business day
AI answer states a false fact	False in at least two of three repeat runs of the same question	Web or SEO lead	Source traced within one week
Branded demand shift	Branded impressions move beyond a band you set from three months of data	Web or SEO lead	Check against news and site changes within a week

Each threshold is an event someone can verify, not an opinion like "negative tone", and each response window is a time to decide, not to resolve. Anything involving an allegation about a named person, a possible crime or a legal claim goes to counsel before anyone replies in public. ### The escalation order When a finding meets a threshold, the owner does the same four things in the same order. Capture it with a dated screenshot and the address. Classify it as accurate, inaccurate, outdated or a policy breach. Choose the route that fits that class: a reply, a correction request to the publisher, a platform report, a legal letter or an owned-content response. Then record the decision and the date to look again. ## The weekly, monthly and quarterly routine A routine this size should take under an hour a week. Put it in calendars with names, not in a document that gets read once.

The routine in three cadences

Weekly, 45 minutesRead new reviews on the Business Profile and each other review source. Clear the alerts folder. Run the saved social searches. Log anything that meets a threshold and send it to its owner the same day.
Monthly, 2 hoursCheck the brand search page and autocomplete in a private window and compare with last month. Open Search Console, filter queries to Branded, compare the last four weeks with the previous period and note the biggest differences. Hold a 20-minute review of open findings with the owners.
Quarterly, 4 hoursRun the AI answer sample on the fixed prompt set. Read the register entry and confirm filing alerts arrive. Test that every alert still delivers. Read this quarter's log for patterns, then raise or lower each threshold based on what was noise and what was missed.
After any eventA launch, a funding round, a complaint wave or a news story triggers an extra check of every surface within the next two days, then a return to the normal cadence.

If a threshold fired often and led to nothing, raise it. If something important was found by luck, write the threshold that would have caught it. ## A realistic timeline and what it costs in hours The first quarter is mostly setup and tuning. These are usual patterns, not promises, and they assume one coordinator with owners who answer within a day or two.

When	What happens	What you have after
Week 1	Register written, owners named, baseline logs of the search page, autocomplete and reviews	A reference point to compare against
Week 2	Alerts created, register follow set up, Search Console branded view bookmarked	Collection running
Weeks 3 to 4	First weekly runs, many false positives, first thresholds tested	A noise list and first changes to the alerts
Week 6	First monthly review with owners	Open findings with decisions and dates
Weeks 8 to 9	Thresholds revised after real findings	A feed that is mostly signal
Week 13	First quarterly review, including AI answers and register	A second baseline and a revised register

The chart below is computed from the routine above: 45 minutes weekly over 52 weeks, 2 hours monthly over 12 months and 4 hours quarterly over 4 quarters.

Hours a year by cadenceComputed from the routine above: 39 hours weekly, 24 monthly, 16 quarterly, about 79 in total

Weekly checks39 hours

Monthly checks24 hours

Quarterly checks16 hours

Our own planning assumptions, not a measured figure. Time to respond to findings and any event checks is not included.

## A worked scenario: a retailer with two stores This is an invented scenario with generic parties and stated assumptions. A retailer of outdoor equipment has 35 staff, an online shop and two stores, each with a Business Profile. The marketing lead is the coordinator. Nobody has checked search results for the brand name in over a year. In week one the coordinator writes the register and logs the brand search page: the retailer's own pages take most of the top ten, with one directory listing and one old forum thread. Autocomplete shows delivery and returns. In week two she sets three alerts, follows the company on the register and books the weekly review read. In week five the weekly review turns up a one-star review on one store profile alleging that a product caused an injury. It meets the first threshold. The customer lead does not reply. She sends it to operations and counsel the same day, who check the order records and agree a calm reply inviting the reviewer to contact the store, without repeating the allegation. It is not reported: it describes an experience, and nothing in it plainly breaks the platform's rules. In the second month the monthly check finds a new autocomplete word after the brand name, about a recalled item. A week later it is still there, and the branded filter shows a rise in that query. The recall is real and published by the manufacturer. The response is an owned page stating which items are affected and what customers should do. The persistence rule meant no one acted on a one-day blip. At the first quarterly review one assistant repeats an out-of-date store address. The web lead traces it to an old directory page and updates the listing. Each finding led to a reply, a fix or a clearer owned page. ## Your next step: a one-page register and one calendar invite

If you want help setting this up

Write the seven-row register today, name one owner per row and book the first monthly review. If you would rather have the checks run for you, our monthly reputation monitoring service covers search results and AI answers on a fixed schedule with the thresholds agreed in advance.

Run one full cycle by hand before you spend on anything: a month of the free routine shows where your real exposure sits. Reputation Partners builds these programs for companies that want a short, owned routine, and this register is the starting point we use. ## Questions people ask ### How often should I check my brand reputation? Match the cadence to how fast each surface changes. Reviews and news mentions weekly, the brand search page and autocomplete monthly, and AI answers, register entries and the thresholds themselves quarterly. Add an extra round within two days of a launch or any news about the company. ### Can I monitor my brand reputation for free? Yes, for most of it. Google Alerts covers web mentions, Search Console shows branded search demand if your site is verified, your Business Profile holds your Google reviews, and many public registers offer a free follow service. The cost is your time, roughly an hour a week, plus the discipline to name owners and thresholds. ### What is the most important thing to monitor first? Start with the brand search page and your reviews. A prospective customer sees them first, and they change over days to weeks. Log the top ten results for your exact brand name today and read your latest reviews, and you have a baseline for every later check. ### How do I stop alert fatigue? Keep the alert set small, give each alert a purpose and review noise monthly. Remove alerts that returned only unrelated mentions, add exclusions for common false matches, and send findings to a single weekly read rather than to everyone. If nobody opened the alerts folder for a month, shorten the set before you add anything. ### Who should own brand monitoring in a company? One coordinator, usually in marketing or communications, owns the routine and the log. Each surface has one named owner, who checks it on the agreed cadence and sends findings that meet a threshold onward the same day. Serious allegations, whether about the company or a named person, always route to counsel before any public reply. The coordinator also runs the monthly review and keeps the thresholds current. ### Does monitoring protect me from a reputation crisis? It does not prevent one, but it shortens the time between something appearing and someone acting. A routine with thresholds and owners means the first decision is made within a day or two, with evidence captured, instead of after the story has spread.