"Find a few competitor accounts to model after first" — every account-growth guide opens with this. There's even a standard way to do it: search your niche keywords, like and follow whatever looks good, and after enough scrolling the algorithm figures out your taste and starts feeding similar accounts and industry content into your stream. This isn't useless — your feed really does fill up with peers. The problem is it's passive. You can only receive whoever the algorithm hands you, and the stream is a jumble of big accounts, small accounts, and off-topic posts. Who's actually worth modeling after versus who just happened to get pushed your way — you can't tell them apart.
The more reliable approach runs the other direction: instead of starting from "who should I follow," start from "which content has already been proven to go viral," then aggregate the accounts behind those viral posts. An account that consistently produces hits has, by definition, had its topic choices and playbook validated by the algorithm — so a competitor surfaced from data has a real string of viral posts standing behind it, far more reliable than whatever the algorithm casually feeds you. Pull enough of these viral-post accounts together and they naturally sort into tiers: the top players, the mid-tier, and the low-follower hits at roughly your own size — each serving a different purpose. This post is about using data to reverse-engineer all three tiers.
Tier your competitors first: top accounts for direction, mid-tier for playbook, low-follower for copyability
Competitor research breaks into at least three tiers, and each answers a different question. Top accounts show you the ceiling of the niche and where the mainstream direction is heading. Mid-tier accounts let you see mature content playbooks and how series and columns are structured. Low-follower accounts at roughly your own size give you samples you can directly copy.
Studying only the top accounts usually backfires — they have teams, brand backing, and years of accumulated weight you can't replicate. Studying only low-follower accounts pulls your direction off course. Use all three together and your competitor pool won't be skewed by a single point of view. The good news: you don't have to assemble these three tiers in three separate passes — one move reverse-engineers all of them from viral posts at once. Here's how it works.
Search one niche keyword and let the system reverse-engineer competitors from viral posts
The logic is straightforward: take a niche keyword, pull the top few dozen viral posts from the last six months, sorted by likes (say 60 to 80 of them), gather this batch of algorithm-validated content together, then pick out the accounts that appear more than once.
Why this logic? Because an account that shows up again and again across a batch of top-liked viral posts is not getting there on luck — it means the account has a systematic content layout and has hit the algorithm time after time. Accounts that surface with one fluke post get filtered out naturally at the "appears more than once" step. Accounts coasting on old content that have actually stopped posting don't make it in either, leaving only the people who are still consistently producing hits right now.
What makes it even easier: this single move delivers all three tiers at once. The accounts that appear repeatedly with large follower counts are your top and mid-tier benchmarks. The system also specifically flags the low-follower high-performance signal — accounts with modest follower counts that still show up again and again in this batch of top-liked viral posts. Those are the hidden gems at roughly your size, the ones you most want to copy directly. For every candidate, the sample appearance count, follower tier, peak and average likes, and signature posts are laid out together, with a one-line AI recommendation reason attached. Who's the ceiling, whose playbook is worth opening the profile to double-check, who's a hidden gem you can learn from directly — all on one screen. You don't have to set follower and like thresholds yourself to filter again.
Once you've locked onto an account, check three things before deciding whether to dig deeper: how long it's been running, how many posts it's published in total, and roughly what its hit rate is. For example, a 500-follower account that has posted 30 times with 3 of them passing a thousand likes — about a 10% hit rate — tells you its topic direction is stable and effective, worth scraping the entire profile to break down.
This habit of using one keyword to map out the whole niche before making a call is the same logic as the fastest way to grow on Xiaohongshu is searching one keyword to run topic analysis — see the full landscape first, then decide who to learn from. If you'd rather skip the accounts entirely and manually dig out individual low-follower viral posts by your own follower and like thresholds, that's a different job — how to find viral content from small accounts on Xiaohongshu and Douyin covers it in more detail.
Automating the grunt work of paging and aggregating is exactly what MediaClaw does — a ready-to-use, no-code Chrome extension. "Find Competitor Accounts" is part of the Niche Strategy toolkit on its search page: search one niche keyword and it automatically aggregates the accounts that recur across the last six months of top-liked viral posts, flags the low-follower high-performance signal, attaches a one-line AI recommendation reason to each account, and lets you add a locked account to monitoring with one click. This step targets the exact pain point of "how do I actually find competitor accounts," sparing you the whole process of scrolling through 100 search results and opening each profile to check follower counts one by one.
Building the competitor pool isn't the finish line. Once you've locked onto these accounts, either scrape their full profiles for a deep breakdown, or set up a competitor monitoring system for them so they push to you the moment they post — no more manually refreshing your feed every day.
Finding benchmarks is only the first step. Continue with the Xiaohongshu Account Analyzer to organize public profiles, historical posts, recurring themes, and engagement evidence before deciding which patterns to learn from and which creators to monitor.
FAQ
Everyone says to find competitor accounts first when starting on Xiaohongshu — where exactly do I find them?
Don't just search a keyword, like and follow, and wait for the algorithm to feed them to you slowly. The more reliable move is to reverse-engineer with "Find Competitor Accounts": the system pulls the top-liked viral posts under your niche keyword from the last six months, picks out the accounts that recur in them, and in one move sorts out the top, mid-tier, and low-follower hidden-gem tiers — each with a line on why it's worth studying. More accurate than passively waiting for recommendations, and far less work than scrolling 100 results by hand.
How do I find Xiaohongshu creators at roughly my own size that are actually worth imitating?
The easiest way is "Find Competitor Accounts," which flags the low-follower high-performance signal as it aggregates candidates — accounts with modest follower counts whose per-post likes far exceed their size get picked out. You can also do it fully manually: when scraping search results, include each creator's follower count, export, then filter again for low followers and high per-post likes — say, keeping content under 5,000 followers but over 500 likes. These accounts prove their topic choices have been validated at a small scale, which makes them more useful to learn from than just studying big accounts.
Can I use the same method to find competitor accounts on Douyin?
Yes. "Find Competitor Accounts" and the underlying search-result scraping work the same on Douyin — same logic of reverse-engineering recurring accounts from the top-liked viral posts under a niche keyword, identical to Xiaohongshu, with no need to switch tools or change your approach.
How many competitor accounts should I track?
Keeping the pool to 5–6 is ideal, and no more than 10. Pick too many and you study none of them deeply. Keep 2–3 each across the top, mid-tier, and low-follower hidden-gem tiers — fewer but sharper makes ongoing breakdowns easier.
