How to Use LinkedIn for Competitive Intelligence

Three overlapping records sit behind any company on LinkedIn. The first is what its staff wrote about themselves: roles, tenure, locations, skills and the products they say they work with. The second is what the company wrote about itself through a page administrator: the About tab, the job listings, the product pages, the posts. The third is what the platform computes from the first two and sells back as insights. With over 1.3 billion members and close to every company in the Fortune 500 carrying a page, no other single source covers companies and the people inside them as completely. The provenance is unlike any of them too, because none of it was compiled by anybody neutral.

What it is not is a register. Nobody here filed anything under compulsion, nothing was audited, and almost nothing is dated. A profile shows the present tense and no edit history, so a role that changed in March and a role that changed yesterday look identical. It is also not an archive: a post that stated a strategy plainly is findable while it is recent and effectively gone once it is not. Read it as a large, current, self-reported picture of a company, assembled by people with a career interest in how they appear, and it will not mislead you.

1. 1Set your viewing mode before you look at anything. Open your settings and decide whether a profile visit carries your name and headline, your job title and industry alone, or nothing at all. On a free account, choosing to browse invisibly also switches off your own list of people who viewed you, so the setting costs something in both directions. Make the decision once, as a team, rather than leaving each person to discover it.
2. 2Pin down the right page. Confirm you have the operating entity rather than a dormant duplicate, a country page or a product spin-off. A parent page can carry up to twenty-five showcase pages, each with its own followers and its own analytics, and a rival's sharpest product messaging often sits on one of those rather than on the page everybody checks.
3. 3Record the About tab with the date you read it. Take the declared industry, size band, headquarters, founding year and specialities into your own sheet, and write down the day. None of it is versioned: the previous wording is gone the moment somebody saves a new one, and the specialities list in particular is where a repositioning shows up first.
4. 4Work the People tab by function, not by person. Filter the workforce by job function and compare the shape against your own, in ratios rather than headcounts: engineers to sellers, and whether a function exists this year that did not exist last year. Read the totals and resist opening profiles one by one, because every profile opened from this tab is drawn against a monthly allowance you cannot see.
5. 5Read the Jobs tab in full. This is the one surface with no cap on how much of it you consume, so consume all of it. Take the role, the location, the recruiter and the hiring manager named on the listing, because the manager attached to a vacancy identifies whose remit is expanding, not merely that a seat exists.
6. 6Open the product tab if they sell software. A product page carries price tiers, the integrations they claim, the customers they chose to name and the recommendations they have not hidden. Record all four. Each is the company committing to something in public, on a surface it maintains separately from its own website and updates less often.
7. 7Read the posts, then keep the ones that matter. Work through the company page and the four or five executives who genuinely publish, and save anything that states a plan, names a customer or defends a position under pressure. Copy the wording and the link into your own notes with a date, because a feed is not an archive and nothing here is guaranteed to be findable in six months.
8. 8Turn your own page analytics around on them. Finish where most teams never look: your own page. The follower list names who follows you, where they work and when they arrived, and the competitor view benchmarks your follower growth and engagement against a rival using the platform's own numbers. What you should hold at the end is a dated sheet per competitor covering firmographics, function mix, open roles, product claims and the posts worth quoting.

The constraints come before the surfaces here, rather than after them, because on this platform they decide the method rather than qualifying it. Almost every competitively useful surface is free. Two of the free ones are rationed, one of them is gated behind a company-size threshold, and one of them was quietly narrowed less than a year ago. Plan the work around that and you will not hit a wall halfway through a quarter.

• A signed-in account opens most of it. The About, Posts, Jobs, People, Products and Life tabs of any company page are available to any member, at no cost. The advertising library needs no account at all. Your own page analytics are free to your own administrators. Nothing in the paragraphs below is an argument for buying something; it is a map of where free stops.
• People research is capped by the commercial use limit. A monthly allowance applies once your activity looks like hiring or prospecting, and three things draw against it: searching for profiles, following the people-also-viewed suggestions, and opening member profiles from a company page’s People tab. That is competitive research described almost exactly. The platform states it will not display how much you have left and will not lift the cap on request, and the allowance resets at midnight Pacific time on the first of each calendar month.
• Three activities are explicitly exempt. Searching for a person by name from the box at the top of the page, browsing your own first-degree connections, and anything you do on the Jobs page. The third exemption is the useful one: a competitor’s open roles can be read exhaustively, every month, at no cost in money or quota, which is why the job side of this source is the one to work hardest.
• The insights panel needs thirty employee profiles. The trended view of a company, covering employee count, headcount by function, senior leavers and open roles by seniority, requires a paid subscription and appears only where at least thirty members have attached themselves to the page. Small and early-stage rivals therefore show nothing at all, which is a coverage gap rather than a finding about them.
• Page benchmarking dropped from nine competitors to one. Until 15 October 2025 a free company page could compare its follower growth and organic engagement against nine named rivals. The free allowance is now a single competitor, with nine available on a paid page subscription and trending posts shown for three of them rather than one. Any walkthrough published before that date is describing a product that no longer exists.
• Network distance and privacy settings both narrow what you see. A profile well outside your own network, or one whose owner has restricted visibility, can appear as an unnamed member with parts of the record withheld. A team with no genuine relationships in a competitor’s market therefore sees noticeably less of that competitor’s staff than a team that has been active there for years, which makes network depth a real and rarely noticed input to the quality of this research.
• Nothing belonging to anybody else can be exported. There is no file to download from a rival’s page and no member data you may take. The only export here is of your own page, where an administrator can produce a spreadsheet of visitor, content, follower and competitor analytics. Everything else leaves the platform as your own notes.

Every guide to this subject opens with the same instruction: follow your competitors’ pages and the profiles of their executives. Almost none of them mentions what that instruction does on the other side of the glass. Reading here is an attributable event on somebody else’s system, and three separate mechanisms record it.

The first is the follower list. A page administrator can open a view of every follower that names them, states where they currently work and shows when they followed, ordered with the most recent arrival first. It sits inside the same analytics that export to a spreadsheet. Following a rival from an account carrying your employer therefore places your company in a sortable list on their side, and a cluster of your colleagues arriving in the same fortnight is legible as an event.

The second is the viewer list on individual profiles. Every member chooses how they appear when they visit somebody: name and headline, job title and industry with the name withheld, or nothing but an anonymous marker. The choice is genuinely a trade rather than a free option, because on a free account browsing invisibly also switches off your own list of people who viewed you. You surrender exactly the information you were trying to withhold. A paid account keeps the list while browsing privately, which is the real reason competitive teams end up subscribing.

The third is the one almost nobody accounts for. The platform’s sales product builds an intent score for accounts a subscriber has saved, and the behaviour it draws on is ordinary research behaviour.

None of this is a reason to avoid the source, and the obvious workaround is worse than the problem: an account under a false name breaches the platform’s terms outright and sees almost nothing anyway. The reasonable response is a decision taken once. Agree who does this work, agree what viewing mode they use and what it costs them, and accept that following a rival is a public act. Teams that skip the decision still make it, by accident, usually when half a go-to-market department follows the same competitor inside a fortnight.

The mirror is worth as much as the warning. Your own page analytics name the people following you, where they work and when they arrived, which means competitor employees watching you are identifiable in a tab most teams never open. Read that list quarterly alongside your visitor demographics, filed with the rest of what you hold on each rival. It costs nothing, it arrives without anybody going to look for it, and it is the half of the exchange you control.

The interface presents all of it in one visual register, which is what makes this source quietly dangerous. A number the company chose, a number an individual typed and a number the platform modelled look the same on the screen and carry entirely different weight. Establish the author before the finding.

The employee count deserves its own note, because it is the figure most often quoted from this source and the mechanism behind it is published. Members attach themselves to a page by naming the employer in their own experience section; administrators cannot add or remove anybody; and the associated count can take up to thirty days to reflect an arrival or a departure. So the number lags joiners, lags leavers, and permanently includes anyone who moved on without editing their profile. That makes it a serviceable trend and a poor level. The dated and audited version of a competitor’s headcount comes from filed accounts and national registry disclosures, which nobody volunteered and nobody can quietly restate.

Member profiles, read in aggregate rather than one by one

The temptation is to open profiles individually, and it is the wrong instinct twice over: it is the activity the monthly allowance rations, and one person’s career history rarely decides anything. The value is distributional. How many people at a rival hold a title containing the word you care about, how many arrived in the last year, how many list a particular product as a skill. A count of people naming a tool as a skill is a rough adoption signal for that tool, and one you can sanity-check by looking at where those people work. Read profiles individually only for the handful of people who genuinely matter: the executive who owns the function you compete with, and the two or three hires whose arrival changed something.

The company page tabs anybody can open

The About tab carries declared industry, size band, headquarters, founding year and a specialities list, and the specialities list is the field almost nobody thinks to copy. It is a keyword-shaped statement of what a company believes it sells, rewritten whenever the positioning moves. The Life tab is recruitment marketing, which makes it a statement of who they are trying to attract and what they believe makes them attractive. The Services tab appears only for organisations selling services and is worth a glance for how they package delivery. None of these fields is dated, so copy them with the day you read them or you will never be able to prove they changed.

The People tab, and why it is the expensive one

This is the richest free surface on the platform and the one that consumes your allowance. It breaks a rival’s workforce down by job function, location, school and skill, and shows how you connect to any of them. Worked properly it answers questions no other free source will: whether a rival’s engineering-to-sales ratio is drifting, whether a function appeared this year that did not exist last year, whether the growth is in one country. Take the totals from the filters, which cost you nothing, and open individual profiles sparingly, because those do.

The Jobs tab, the one surface with no meter

Job activity is exempt from the monthly allowance, so this is where to be exhaustive. Each listing carries the role, the location, and frequently the recruiter and the hiring manager attached to it, which names the person the new hire will report to rather than only confirming that a seat has opened. Watch for roles that describe a capability rather than a vacancy, and for the same role reposted in a second region, which is expansion rather than churn. Competitor job postings repay being read as specifications rather than skimmed as headlines, and the requirements list is where a rival describes a capability it does not have yet.

Posts, from the page and from the people

Three registers publish here and they are worth separating. The company page is written by a team, reviewed, and therefore says almost nothing accidental. Executive posts are the interesting layer: a founder or a functional leader arguing a position in public is stating strategy in a way no press release would survive. Ordinary employee posts are the noisiest and the earliest, and a cluster of people in one function suddenly posting about one theme is usually a launch or a reorganisation that has not been announced. Reaction and comment counts are public on all three, which lets you see what landed rather than only what was said.

The advertising library, and what Europe forces into it

Sponsored posts are searchable by advertiser without an account, covering anything served from June 2023 onward and kept until twelve months after an ad stops running. For business software this is frequently the most representative view of a rival’s paid programme, because it is the channel their buyers actually sit in. Anything served to a European audience carries more: an estimated total impression figure, a country breakdown and the top targeting parameters the advertiser selected. Targeting is the unusual disclosure, because it is a declaration by whoever bought the advertising instead of a guess made from the picture. Weigh it against the equivalents on other platforms, since the public ad libraries vary enormously in what they oblige an advertiser to reveal.

The insights panel, and what a subscription adds

On a paid subscription a company page gains a trended view: total employee count over time, headcount distribution and growth by function, total job openings segmented by seniority and function, and former employees who held a title of vice president or above and left within the last twenty-four months. That last one is a maintained senior-departure record with dates on it, which is otherwise a genuinely tedious thing to assemble. All of it needs at least thirty employee profiles attached to the page, and all of it inherits whatever the profiles got wrong.

If a rival sells software, their product tab is the most under-read surface they maintain. It exists to generate demand, which is precisely why it is useful: everything on it is a claim the company decided to make in public, on a property it updates separately from its own website and therefore keeps in step with it imperfectly. Four fields carry real weight.

Plans and pricing. Price tiers and free-trial terms appear here as a distinct copy of what sits on their own site. Diff the two. A tier that exists in one place and not the other, or a trial length that differs by a week, is usually a change that landed on the main site and never propagated, which dates the change for you. Neither copy tells you what anybody paid, which is where reading competitor pricing stops being a matter of finding the page and starts being a matter of finding the discount.

Featured integrations. The products a competitor says theirs works with, curated by them. It is a partnership signal and a roadmap signal at once, because an integration is engineering already spent, and the ordering tells you which relationships they are proudest of. Removals matter more than additions and nobody announces them, so this only works if you kept the previous list.

Featured customers. A company can name up to nine customer organisations on a flagship product page, and a named organisation can ask the platform to remove it. That governance detail is what makes the list worth more than a logo wall on their own site: a name that stays is a customer that has not objected to being publicised. Cross-check each against what those companies say themselves, and treat the set as their strongest references rather than as a representative sample of who buys.

Product recommendations. Members can write about a product, and two facts change how you read what you find. An administrator can hide any individual recommendation, so what is displayed is a permitted subset rather than a corpus. And companies can no longer invite members to write them, which means the volume you see is unprompted rather than campaigned for. A rival with four recommendations and no way to solicit more is not necessarily a rival nobody likes.

The remaining fields are lighter but not empty. The suggested roles a company attaches to a product are it declaring its target buyer in its own words. The call-to-action button reveals whether they are selling a demo, a trial or a download, which is a distribution choice. The media block shows what they think the product looks like at its best. And the count of people who list the product as a skill is a crude adoption proxy that at least has names behind it.

Start with the constraint, because it shapes the artefact. There is no export of anybody else’s page or profile, collecting content with software is prohibited outright, and the platform’s terms go further than most people realise: copying, using, displaying or distributing information taken from the service without the content owner’s consent is restricted regardless of how you obtained it. So the file you build is a record of your observations with sources attached, not a copy of their material. That is a real editorial discipline rather than a legal formality, because a sheet of pasted profile text is unusable anyway and a sheet of dated observations is exactly what a brief needs.

The sheet, one row per competitor per quarter

Date read, competitor, associated employee count, headcount split across the four or five functions you care about, open roles by function, the specialities line verbatim from the About tab, the current price tiers named on the product page, the integrations listed, the featured customers, and a link to each. Add one free-text column for the single most significant thing you saw. Eleven columns holds it, and the value is entirely in having last quarter’s row beside this quarter’s, because none of these fields has a history of its own.

The quotes file, kept separately and kept short

Posts that state a plan, defend a position or name a customer are worth holding, and the feed is not going to hold them for you. Save the wording, the author, the date and the link, and keep it to the few that would actually change a decision. Attribute everything, use it internally, and do not lift somebody’s words into anything you publish without asking them, which is both the platform’s rule and the sensible one.

What can legitimately be exported, and by whom

Your own page is the exception. An administrator can generate a spreadsheet of visitor, content, follower and competitor analytics for a chosen period, and the competitor file contains the public comparison figures the platform already displays. That is a real dataset you own, updated automatically, and it is where the follower-side signal described earlier becomes something you can chart rather than something you scroll. Beyond that, competitive intelligence platforms exist to hold public company and product surfaces under continuous observation and hand back a dated list of what moved, which removes the remembering without removing the reading.

What to read once the sheet exists

Ratios before counts. Engineering to sales, sellers to managers, open roles to current headcount. Then the deltas: a specialities line that gained a phrase, a price tier that disappeared, an integration that was dropped, a function that grew twelve per cent while the company grew two. One quarter of this is a description. Four quarters is a direction, which is the only thing on this platform worth briefing an executive on.

The folklore around this platform is stricter in one direction and far looser in another than the rules themselves. The words competitive, competitor and benchmark do not appear anywhere in the user agreement, so there is no clause aimed at competitive research as an activity. What the agreement restricts is how you access the service and what you do with what you find. None of the following is legal advice.

• Never research under an identity that is not yours. Creating a false identity, misrepresenting who you are, or making a profile for anyone other than yourself is a direct breach. It is also the fact pattern that turns a defensible position into an indefensible one, and it is the shortcut a competitive team is most often tempted by. Research as yourself, at whatever visibility setting you have chosen.
• No software may collect the content for you. The agreement prohibits developing, supporting or using scripts, robots, crawlers or browser add-ons that copy the service and the profiles on it, and separately prohibits automated methods of access. Everything described on this page is a person reading pages on their own account, which is the only method that stays inside the rules.
• The reuse clause is the one that touches ordinary work. Copying, using, displaying or distributing information obtained from the service, whether directly or through a data broker or aggregator, without the content owner’s consent is restricted. Buying the same information from a third party does not launder it. In practice this is the difference between a brief that says what you concluded and a brief that reproduces a rival’s profile text.
• The case everyone cites did not end the way it is remembered. In April 2022 the Ninth Circuit held that collecting public profile data probably falls outside the Computer Fraud and Abuse Act. That November the district court found the same company had breached the user agreement, and in December 2022 the parties entered a stipulated judgment of five hundred thousand dollars, a permanent injunction against further collection, and an order to delete both the member data and the software built from it. A ruling on one criminal statute is not a licence.
• You are handling data about identified people. Everything on a profile concerns a named individual who has privacy rights wherever they live, regardless of the fact they published it. Keep the record to professional context, keep nothing you have no use for, and keep competitive research separate from any file about approaching those people as candidates, which is a different activity with its own obligations.

• Anything financial. No revenue, no growth rate, no margin, no funding detail beyond what somebody chose to post about. Proxy: function-level headcount growth against open roles gives you the shape of the spending, and the direction of that shape is often more decision-relevant than a revenue figure would be.
• Whether customers are leaving. Nothing here reports retention, and the named customers on a product page are a curated list that a departing customer has no obligation to leave. Proxy: a support or customer-success function growing faster than sales, or a featured customer quietly disappearing between two of your own snapshots.
• When anything changed. The structural gap. Profiles, About tabs, taglines, specialities, price tiers and integration lists all display a present state with no history and no date. Proxy: your own snapshots, and archived copies of the pages they mirror on their own website, which at least carry capture dates.
• What it is like inside. Staff write here to be seen by future employers, so the register is uniformly positive and departures are described in the language of gratitude. Proxy: the anonymous employee reviews on Glassdoor, which trade identity for candour and answer the question this platform structurally cannot.
• Anyone who is not on it. Coverage varies enormously by market and by function: engineering and commercial roles in North America and Western Europe are well represented, and plenty of markets and plenty of operational functions are not. Proxy: local job boards and the national registry for the country in question, neither of which depends on individual participation.
• Unannounced product direction with any precision. Job listings hint at capability and executives argue about categories, but neither states a plan. Proxy: documentation and changelog diffs on their own properties, where a new permission scope or an undocumented endpoint appears months before the announcement does.

Split the cadence by surface, because they move at different speeds and cost different amounts to read. Jobs is monthly, draws on no allowance, and is the cheapest early signal available here, so it is the one to never skip. The About tab, the product pages and the function mix are quarterly, which conveniently also spaces out the metered work. Posts are continuous and mostly noise, so watch rather than review them, and capture only the ones where something is argued rather than announced.

Do the metered work in the first week of a calendar month. The allowance resets at midnight Pacific time on the first, so a quarterly people-side pass scheduled for the last week of a quarter is the version most likely to stop halfway through with no warning and no way to restart it. This is a small piece of scheduling that saves an entire wasted afternoon, and it is the kind of thing nobody tells you until it has happened once.

Five events justify an off-cycle look regardless of the calendar: a funding round, an acquisition in either direction, a senior departure, a product launch, and a burst of hiring into a function that was empty last time you looked. Each of those produces a fortnight of unusually informative posting from people who are pleased, worried or leaving. Record the same fields every time, and append them to the competitor profile you already maintain rather than starting a fifth file that one person opens and nobody inherits.

The defining weakness of this source is not accuracy, it is amnesia. A rival’s specialities line, page tagline, product tiers and integration list are fields, and a field is overwritten rather than versioned. The previous value ceases to exist the moment somebody saves a new one, with no date, no notification and no trace. The post where a founder explained the reasoning behind a repositioning is findable for about a fortnight and then effectively is not. So the entire history of a competitor on this platform is whatever you personally happened to write down, and the quarter you were busy is a quarter that no longer exists.

The cost lands later and lands on somebody else. A pricing tier that appeared in February surfaces in the June review, and by then two deals have been argued against a package nobody on your side had prepared for. Closing that gap is what the competitive intelligence software category is for. Flares watches the public surfaces where pricing, product, documentation and messaging changes appear, and puts a date against each one as it lands rather than as somebody eventually notices it, so the series exists whether or not anybody remembered to check. What no product can do is the half of this that requires being signed in as you: nothing can read a rival’s People tab on your behalf, and nothing will tell you why a named person left after four years or what the reorganisation was actually for. That reading is yours, and it is worth the hour a quarter precisely because it cannot be delegated to anything.