Search volume for "peloton customer service" runs around 2,900 a month in the United States. The same brand's reviews query does 600. Complaints does not register at all.
That ordering is the problem in miniature. The query a reputation team would watch sits third. The query almost nobody classifies as reputation sits first, by a factor of nearly five.
Branded keywords are the searches that contain a company, product, or executive name. Most of them are not just the name. Ahrefs analyzed roughly 150 million U.S. keywords and found that queries of three or more words account for the largest share of branded search, meaning the typical branded searcher already knows the brand and is looking for something specific about it (Ahrefs, May 2025). The word attached to the name is doing the work. That word is the modifier, and the full set of them is the map.
Six categories, and most teams watch two
Modifiers sort into recognizable groups. The categories matter more than any individual term, because a brand that has never audited its own set will find gaps by category rather than one query at a time.
- Evaluation: reviews, ratings, is it legit, worth it, versus a named competitor, alternatives.
- Operational: customer service, support, login, refund, cancel, not working, phone number, returns.
- Allegation: lawsuit, scam, fraud, complaints, investigation, recall, settlement.
- Employment: careers, Glassdoor, salary, interview, layoffs, culture.
- Community: Reddit, forum, YouTube, Quora, plus the name of any platform where the audience congregates.
- Corporate: stock, revenue, CEO, acquisition, funding, earnings.
Most monitoring covers evaluation and allegation. Those are the categories that feel like reputation. They are rarely the categories with volume.
The operational modifiers carry the traffic
Return to the connected-fitness example, which is worth using precisely because the brand is not in crisis. Customer service pulls 2,900 searches a month. Layoffs pulls 600, matching reviews exactly. Reddit pulls 250. Lawsuit pulls 150. Cancel subscription pulls 40.
A healthy, well-capitalized consumer brand still generates a full modifier spread, and its single largest branded query after the name itself is someone trying to reach a human. That query is usually owned by a support team measured on ticket deflection, not by anyone thinking about what page one communicates. The result is a top-branded search where the ranking page is a help center article, a third-party phone number aggregator, or a forum thread from four years ago.
Nobody files that under reputation. Everybody experiences it as reputation.
THE GAP TO WATCH FOR
When the highest-volume branded query after the brand name is an operational one, the department that owns the answer is almost never the department that owns the search result. Ask who is accountable for what ranks on "brand name plus customer service." In most organizations the honest answer is nobody.
Low volume is not low exposure
Two of the modifiers above report essentially no search volume. Complaints and refund both come back at or near zero. The temptation is to strike them from the map.
That reading confuses a measurement floor with an absence. Keyword tools estimate from sampled clickstreams and stop reporting below a threshold, so a query with a few dozen monthly searches and one with none look identical in the data. The difference between them matters enormously when the searcher is a procurement lead running due diligence, or a journalist checking a tip.
More to the point, a modifier does not need search volume to be displayed. It needs only to exist in the aggregate query data Google draws on, at which point it can surface as a prediction in the search box, in People Also Ask, or in related searches. Autosuggest works from that same well. A term nobody often searches can still be shown to everyone who searches the brand name once.
The same modifier can appear three times on one page
Semrush analyzed more than 10 million keywords using clickstream data from Datos and found that related searches appeared alongside 95.32% of AI Overviews and People Also Ask alongside 90.03% (Semrush, December 2025). Both features are generated from aggregate query behavior, the same source that feeds autosuggest.
So a single search for a company name can render the damaging modifier three separate times: predicted in the box before the search runs, echoed in People Also Ask partway down, and repeated again in related searches at the bottom. The searcher did not type it once. They were shown it three times, and the third time reads as a suggestion.
The mechanics of how those predictions are formed, and what can legitimately be done about them, fall under the autosuggest pillar. What matters here is narrower: the modifier map and the autosuggest set are drawn from the same data, so auditing one gives a substantial preview of the other.
Building the map
The method is unglamorous and takes an afternoon.
- 1Pull every keyword that contains the brand name from a keyword tool, with no volume floor. The zero-volume rows are part of the dataset, not noise to be filtered.
- 2Add the autosuggest predictions for the brand name, then for the brand name followed by each letter of the alphabet. Tools automate this; doing it manually for the first audit is more instructive.
- 3Sort every query into the six categories. Count by category rather than by query.
- 4For each query, record what currently ranks first and whether the brand controls it.
- 5Repeat for each executive name, each product name, and each brand the company operates under.
The output is a spreadsheet, not a dashboard, and it dates quickly. Run it quarterly. The change between runs is more informative than any single snapshot, because new modifiers appearing is an early signal that something is developing before it reaches coverage.
What answers each category
A gap in the map is a content assignment, and the assignment differs by category. Evaluation queries want a comparison page or a review presence the brand has legitimately earned. Operational queries want a genuinely good support page that ranks, which is usually a technical problem rather than a writing one. Employment queries want a careers presence and an honest employer brand, which no amount of search work substitutes for.
Allegation queries are the hardest and the most misunderstood. The instinct is to stay silent so as not to legitimize the term. Silence cedes the page to whoever published it, and on those queries, the publisher is frequently a competitor's affiliate content or a legal marketing site. Review management and Reddit cover the platform-specific mechanics where those conversations actually live.
WHAT THIS IS NOT
Building the map is not a license to create pages targeting damaging modifiers about the company. A page built to rank for a brand-plus-scam query, written to reassure rather than inform, tends to rank and then convince nobody. The map identifies which questions are being asked. Whether a question deserves a dedicated answer, a better existing page, or nothing at all is a judgment call, and the DON'Ts of reputation management covers the promises that should never accompany that work.
The map differs by what is being searched
A brick-and-mortar business, an ecommerce brand, and a named individual generate different modifier sets from the same audit method. Local businesses skew operational and location-bound. Ecommerce skews toward returns, shipping, and legitimacy. Individuals skew toward employment history, net worth, and controversy. Three companion guides work through each case: GBP and branded search for brick-and-mortar, branded search for ecommerce and service-area businesses, and branded search for individuals and professionals.
The Board's View
Branded search reporting that shows a single number for the brand name is reporting on the smallest and safest slice of the surface. The name itself almost always ranks correctly, because the company owns the domain. Everything interesting happens one word later.
The board-level question is whether anyone has ever produced the full map, and whether the categories with the highest volume have an owner. In most organizations, the operational modifiers belong to support, the employment modifiers belong to HR, the allegation modifiers belong to legal, and the search results for all three belong to nobody. That distribution is the structural argument for a single accountable owner.
Board Questions to Ask
- "Show me the full list of queries containing our name, not just the name itself." A team that produces one number has audited the easy part.
- "Which modifier has the most search volume, and what ranks first for it?" If the answer is an operational query owned by support, the reporting line and the search result are in different departments.
- "Which modifiers appeared in the last quarter that were not there before?" New entries are the early warning. They surface before coverage does.
- "What do our executives' names return with a second word attached?" Executive modifier sets are audited far less often than corporate ones and are frequently worse.
- "Where we have chosen not to answer a question, was that a decision or an oversight?" Both look identical on the search results page. Only one is defensible.
Part of the brand reputation management pillar.