Benchmark LinkedIn Post Engagement Across Company Pages

Social and content teams get asked the same question every quarter: is our LinkedIn doing well compared with the companies we are measured against? Collecting a competitor’s last 50 posts by hand means scrolling, copying two numbers per post into a sheet, and then realizing the comparison is meaningless because the other page has six times your followers.

Try it live: LinkedIn Company Posts Scraper: No Login, No Cookies. Pay per result delivered. Failed and empty results are never charged.

The LinkedIn Company Posts Scraper pulls up to 70 recent posts per company page, logged out, with exact counts and the page’s follower count on every row. That last field is what turns raw numbers into a fair benchmark.

What you get and who it is for

Field	Role in the benchmark
reactions_count	All reactions on the post, the exact number LinkedIn prints
comments_count	Exact number of comments
company_followers	The page’s follower count at run time, on every row, for scaling
is_repost	True when the company reposted someone else’s post
media_type	none, image, video, document or article, to split results by format
posted_at	Publish time, ISO 8601, to set a fair time window
company_name, post_url	Which page, and a link back to each post for review

Social media managers use this to report share of engagement against named competitors. Agencies use it to set realistic targets for a new client. Content leads use the media_type split to decide whether video or documents deserve more of next quarter’s budget. For a sense of scale beyond followers, a company’s employee list read off its public page gives you headcount to normalise against as well.

The input

{
  "companies": [
    "https://www.linkedin.com/company/stripe/",
    "openai",
    "nvidia",
    "hubspot",
    "zomato"
  ],
  "maxPostsPerCompany": 70,
  "postedWithinDays": 45,
  "includeReposts": false
}

• includeReposts set to false keeps only posts each company published itself. This matters more than it looks: reposts carry the original post’s reactions. In our sample, one repost on Stripe’s feed showed 1,753 reactions, more than any post Stripe published itself in that window (its best had 1,029).
• postedWithinDays puts every company on the same time window. Without it, 70 posts covered 7 weeks for NVIDIA but 11 months for Zomato in our proof run, and you would be comparing different seasons. 45 days fits inside the 70 post ceiling even for a page that posts as often as NVIDIA.
• maxPostsPerCompany at 70 is the most LinkedIn shows a logged-out visitor. The ceiling counts feed items, reposts included, so a page that reposts a lot can return fewer of its own posts.

What the output looks like

[
  {
    "company_name": "OpenAI",
    "company_followers": 11878447,
    "post_url": "https://www.linkedin.com/posts/openai_were-demonstrating-how-frontier-models-have-activity-7508603626362236930-vwKd",
    "is_repost": false,
    "author_type": "company",
    "posted_at": "2026-09-23T19:10:07Z",
    "reactions_count": 1057,
    "comments_count": 257,
    "media_type": "image"
  },
  {
    "company_name": "Stripe",
    "company_followers": 1754153,
    "post_url": "https://www.linkedin.com/posts/stripe_new-stripe-data-shows-fraudulent-actors-disproportionately-activity-7505675379710500864-iMiU",
    "is_repost": false,
    "author_type": "company",
    "posted_at": "2026-09-15T17:14:19Z",
    "reactions_count": 141,
    "comments_count": 27,
    "media_type": "image"
  }
]

These two rows are each page’s median post by reactions in our run on Stripe and OpenAI on October 7, 2026 (20 posts per company, reposts included; we dropped the 6 Stripe and 5 OpenAI reposts before computing). Across their own posts, OpenAI’s median was 1,057 reactions and 140 comments, Stripe’s 141 reactions and 29 comments. That looks like a 7.5 times gap, but OpenAI had 11,878,447 followers against Stripe’s 1,754,153, a 6.8 times gap. Per million followers, the median post drew 89 reactions at OpenAI and 80 at Stripe, nearly level, while Stripe drew more comments relative to its audience (17 per million against 12).

Use medians, not averages. OpenAI’s mean was 2,452 reactions, pulled up by one video with 14,552. Format told a different story at each page too: OpenAI’s 6 videos had a median of 2,355 reactions against 444 for its 5 article posts, while Stripe’s videos (130) did no better than its images (141).

Run it in Apify Console

1. Open https://apify.com/themineworks/linkedin-company-posts-scraper and click Try for free (or Start if you already use it).
2. In Company pages, enter your own page and each competitor, one per line.
3. Set Max posts per company to 70 and Posted within the last N days to 45.
4. Untick Include reposts.
5. Click Start and watch the Output tab.
6. Use the Export button to download CSV, JSON or Excel, then build the per-follower columns in your spreadsheet, or use the script below.

Run it from Python

pip install apify-client

import csv
from collections import defaultdict
from datetime import datetime, timezone
from statistics import median

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("themineworks/linkedin-company-posts-scraper").call(run_input={
    "companies": ["https://www.linkedin.com/company/stripe/", "openai", "nvidia", "hubspot", "zomato"],
    "maxPostsPerCompany": 70,
    "postedWithinDays": 45,
    "includeReposts": False,
})

now = datetime.now(timezone.utc)
posts = defaultdict(list)
followers = {}
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    if item.get("_type") == "info":
        continue  # the run summary row, never billed
    posted = datetime.fromisoformat(item["posted_at"].replace("Z", "+00:00"))
    if (now-posted).days < 7:
        continue  # too young, still collecting reactions
    name = item["company_name"]
    posts[name].append(item)
    followers[name] = item["company_followers"]

rows = []
for name, items in posts.items():
    millions = followers[name] / 1_000_000
    reactions = median(p["reactions_count"] for p in items)
    comments = median(p["comments_count"] for p in items)
    by_format = defaultdict(list)
    for p in items:
        by_format[p["media_type"]].append(p["reactions_count"])
    rows.append({
        "company": name,
        "followers": followers[name],
        "posts": len(items),
        "median_reactions": reactions,
        "median_comments": comments,
        "reactions_per_million_followers": round(reactions / millions, 1),
        "comments_per_million_followers": round(comments / millions, 1),
        "median_reactions_by_format": {fmt: median(v) for fmt, v in by_format.items()},
    })

rows.sort(key=lambda r: r["reactions_per_million_followers"], reverse=True)
with open("linkedin_benchmark.csv", "w", newline="", encoding="utf-8") as f:
    writer = csv.DictWriter(f, fieldnames=list(rows[0]))
    writer.writeheader()
    writer.writerows(rows)

for r in rows:
    print(f"{r['company']:<12} {r['posts']:>3} posts  "
          f"{r['reactions_per_million_followers']:>7} reactions/M  "
          f"{r['comments_per_million_followers']:>6} comments/M")

The script drops posts younger than 7 days, because they are still collecting reactions. In our data, a Stripe image post went from 75 to 92 reactions in its second day. Treat a format with only one or two posts as a hint, not a finding.

Run it every month

Save the input as a task, open Schedules, create a schedule with a cron expression such as 0 7 1 * * (7:00 on the first of each month), and pick the task. Keep onlyNewPosts off for this job: a benchmark needs every post in the window with its current counts, not just the ones you have not seen. Send the dataset to a Google Sheet with Apify’s integration, or run the script from a webhook when the run finishes, and you have a monthly table to chart.

What it costs

You pay per post delivered: $1.49 per 1,000 posts on the Bronze plan, $1.29 on Silver, and $0.99 on Gold, Platinum and Diamond, plus a flat $0.005 run start fee per run. Skipped reposts and posts outside the date window are not charged.

A worked example: five companies at the full 70 posts each is 350 posts, about $0.52 on Bronze plus $0.005 to start the run, so about $0.53 a month on a monthly schedule. Quieter pages return fewer posts in 45 days and cost less.

Limits worth knowing

• About 70 recent feed items per page is the ceiling for a logged-out visitor. This method benchmarks recent activity, not a page’s whole history.
• company_followers is the count at run time, not at the moment each post went out. For a fast-growing page, older posts look slightly weaker per follower than they were.
• There are no impressions, views or repost counts. Reactions and comments are the only engagement numbers LinkedIn shows logged out, and reactions come as one total, not split by type.
• Counts are as of the run. Posts from the last few days are understated, which is why the script leaves them out.

Related

• Monitor Competitor LinkedIn Company Posts on a Schedule
• LinkedIn Company Scraper: Size, Industry, Website, and Followers
• LinkedIn Scraping Tools Compared: Which Actor for Which Job
• LinkedIn Company Posts Scraper