Tracking competitors

What tracking competitors does

The competitors you track are measured alongside your brand in the same answers, so you get a side-by-side view. You can see how often each one is mentioned, where it ranks, and how your share of voice compares.

Rivals and the watchlist

A project holds competitors in two groups, and the difference is how far each one is measured.

Rivals are the brands you are competing with. They are measured everywhere: in the AI answers, in off-site presence and share of voice, in the SEO comparison, in your keyword positions, and in your reports and alerts. This is the group your plan puts a number on.

The watchlist is for brands you want to keep an eye on without committing a rival slot. A watched brand is read out of the same answers your scan already produced, so you get its mention rate, its position and its trend, and none of it costs anything. What you do not get is the rest: no off-site presence, no SEO comparison, no keyword positions, no alerts. Every project can watch up to 20 brands, on any plan.

Moving between the two is one click on the arrow in a competitor's row: promote a watched brand to rival, or move a rival down to the watchlist. Nothing is lost either way, because both groups are read from the same answers. A brand promoted to rival simply starts getting the measurements the watchlist does not buy, from its next scan on.

Which one should a brand be in? Make it a rival if you want to act on it: see where it is cited that you are not, compare its site against yours, or get told when it moves. Watch it if you only want to know whether it keeps showing up.

The Competitors view

Every project has a Competitors section. At the top you get an overview (share of voice and a visibility ranking of your brand versus your rivals) and a head-to-head table. Below that, your rivals and your watchlist, each as its own list.

• Add a competitor from the box above the list (a name and, optionally, its website). As you type, LumaRank checks brands it already knows: if there is a close match, it offers the recognized name and website so you can adopt it in one click and avoid a near-duplicate. New competitors go in as rivals while you have slots free, and onto the watchlist once you do not. The form tells you which before you confirm.
• Open a competitor (click its row) to see its own results in a side panel: visibility and share of voice over time, its on-page SEO score, and where it shows up off-site.

Add the website, not just the name. The name is what we look for in the assistants' answers, so a competitor without a website is still measured there. The website is what unlocks the rest for a rival: it adds that competitor to the SEO comparison, shows its position in your keyword results, and tracks where it ranks when someone searches its own name.

It also keeps share of voice fair. A company's own pages rank for its own name, and we leave yours out of your off-site count so the number reflects places that talk about you rather than pages you publish. Without a website we cannot tell a competitor's own pages apart from third-party ones, so that competitor's share reads higher than it really is. The Competitors page flags any rival that is missing one.

Websites are yours to set. We only keep a competitor's website when we can confirm it is really theirs, so we never guess one that could be wrong. If a rival is missing its website (or has the wrong one), use "Add website" or the edit icon on its row to set it.

What is private and what is not. Your list of competitors is yours: nobody else can see who you track, and none of your visibility numbers ever leave your account, because they are measured against your prompts and would be meaningless anywhere else. Measurements of a public website are a different thing. A company's on-page SEO and where it shows up around the web are the same facts no matter who looks, so we measure them once and reuse them. That is what keeps a rival from costing you anything per scan, and it is why a competitor's off-site numbers can carry a measurement date a few days before your scan.

Instant history. When you track a competitor that already appeared in your past scans, its history shows up right away, so you do not have to wait for a new scan to see how it has been doing. This holds even for competitors we only suggested and you had not tracked yet: they keep building history in the background, so if you adopt one later its past data is already there.

Finding competitors

If you are not sure who to add, LumaRank suggests them for you, and every suggestion says why it is there: how often that brand came up in your answers and across how many scans. You can check any of it against your own dashboard.

Suggestions are ranked by how much a brand behaves like a rival, which is not the same as how well it is doing. A brand named once, in first place, is not a competitor. One that keeps coming up across scans probably is. So the ranking rewards how often a brand appears, how consistently it appears across scans, and whether it is still appearing, and it deliberately ignores what position it lands in.

Automatic on your first scan. The first time you scan a brand, LumaRank looks for its most relevant competitors and starts tracking the ones that fit your plan's rival slots, so your charts have useful comparisons from the start. Any it finds beyond your available slots are kept as suggestions you can track later. This runs once per project, so a second brand in the same company gets its own pass.

From your data (free, in the "Explore competitors" box). Suggestions are scoped to the project you are in: brands seen in this project's scans, competitors you saved on another project measuring the same brand, and brands commonly tracked by companies in your industry. Pick the ones you want and add them; there is no scan cost.

Products are filtered out. AI answers name tools, clouds, model providers and libraries alongside actual companies, and most of them are things you would use, not businesses you lose deals to. LumaRank screens them out of the suggestions, and judges what is left against what your brand actually does, so a supplier or a company in an unrelated market does not get offered as a rival. If something you already track looks more like a product than a competitor, the row says so. It is only a note: nothing is removed or changed for you, because sometimes the odd one really is your competitor and only you can know that.

Explore (a web scan). We scan the web and return likely competitors with their names and websites, and you pick which ones to add. Every project also gets one of these automatically on its first completed scan, so you do not start from an empty list.

Paid plans include a set number of these scans each month (higher tiers include more), and the free plan includes one to try. The competitor exploration add-on raises the monthly number and runs deeper scans that surface more competitors, and it lets you focus a scan on a specific market, segment, or set of keywords. If you run out before the month resets, scan packs top you up: a one-time purchase that does not expire and does not change your subscription. Both are on Plans.

How many you can track

Rivals are counted per project. Your plan sets how many each project can have, and higher tiers allow more (see Plans and what each limit means). Two projects on the same plan each get the full number, so a second brand does not eat into the first one's slots.

The watchlist is free and flat: up to 20 per project, whatever your plan. It is not a plan lever, because watching a brand adds nothing to what a scan has to go and fetch.

When you add a competitor and your rival slots are full, it goes onto the watchlist instead of being refused. You can promote it later by moving another rival down, or by upgrading. If you downgrade, the rivals you already have are kept: you just cannot add new ones beyond the new limit until you have room.

The profile in a competitor's details

Opening a competitor shows a short About block: what the company does, clients it names, and where it operates. It is summarised from that company's own website, and the pages we read are linked underneath so you can check anything that looks off.

Two things worth knowing:

• Clients are only listed when their own site names them. We never add a customer we have not read on their pages, so an empty list means their site does not say, not that they have none.
• Locations marked "estimated" were inferred, usually from the language, currency or contact details on the site. The ones without that mark are stated by the company.

Keeping it current. A summary is dated, and once it is a month old a Refresh button appears next to the About heading: it reads their site again and updates the summary. Until then the button stays hidden, because re-reading a site that almost certainly has not changed adds nothing. Refreshing one competitor updates it for everyone who tracks that company, so a summary you refresh is current for the whole platform.

When their site cannot be read. Sites that build their pages in the browser are handled: if the first read comes back nearly empty, we load the page the way a browser would and read what a visitor would see. What is left over is a site that blocks automated readers outright, or one with almost no text on it. We say so in the About block instead of leaving it blank, and we try that site again every three days: a page that reads as empty after we load it will read the same way an hour later, so retrying sooner just spends your patience. The block tells you when the next attempt opens, and you can trigger it yourself from there once it does. It only affects the summary: every metric on the competitor, from mention rate to share of voice, is measured separately and is unaffected.

If a competitor has no website saved, we may still find one: brands that other parts of LumaRank already know get profiled anyway, and the block then offers to save that address as their website. It is offered rather than applied, because the website decides what we crawl and rank for them, and that is your call to confirm. Saving it also unlocks their SEO comparison and keyword ranks.

For a competitor nobody has a website for, there is nothing to read, and the block asks you to add one.

Where the comparison shows up

• In the AI Visibility results, the competitor comparison shows how you stack up in the assistants' answers (see Comparing against competitors).
• In off-site presence, your share of voice compares where you and your rivals show up across the web (see Share of voice).

Charts show rivals by default. A rival is measured on every scan, so its line runs unbroken and can be read against yours. A watched brand only gets a point on the scans where an answer happened to name it, so its line has gaps that mean "not measured" rather than "not mentioned". On the project overview chart you can switch watched brands on from the Watchlist control next to the legend, with that caveat in mind.

Tracking the right competitors, and keeping the list current, makes every comparison more meaningful.