Twitter (X) engagement rate: the formula, real benchmarks, and what's actually good

Hootsuite puts the average X (Twitter) engagement rate at 1.8%. RivalIQ's 2024 Social Media Industry Benchmark Report, drawn from thousands of professional brand accounts, puts the median at 0.029%. Neither is wrong. They are measuring different posts, on different accounts, with different math.

That gap is the whole problem with the question "what is a good engagement rate." A single percentage is meaningless until you pin down two things: which formula produced it, and how big the account is. A 1% rate is below average for an account under 1,000 followers and a top-10% result for one over 100,000. Same number, opposite verdict.

The rest of this guide is the honest version: the formulas that produce these numbers, real benchmarks broken out by follower count, why the rate X's own Analytics shows you looks so low, and how fake followers quietly drag the whole thing down.

There is no single formula. There are three in common use, and they divide by different things, so the same post produces three different rates.

1. By followers (the benchmark default). Engagements divided by your follower count. This is what RivalIQ, most agency reports, and our own free X engagement rate calculator use:

(likes + reposts + replies + bookmarks) / followers × 100

2. By impressions (what X Analytics reports). Engagements divided by how many times the post was seen. This is the number inside your native X Analytics, and it runs far lower than the follower-based version because a viral post collects impressions from people who do not follow you:

engagements / impressions × 100

3. By reach. Engagements divided by unique accounts that saw the post. X does not expose reach to most users, so this one is mostly theoretical unless you have a third-party analytics suite. Treat it as a more flattering cousin of the impression formula.

A worked example. A post gets 200 likes, 30 reposts, 20 replies, and 50 bookmarks — 300 engagements — from an account with 10,000 followers, and it pulled 40,000 impressions. The follower-based rate is 300 / 10,000 = 3.0%. The impression-based rate is 300 / 40,000 = 0.75%. Both describe the same post. Quote whichever one flatters you and nobody can technically call it wrong, which is exactly why benchmarks are such a mess.

The numerator matters as much as the denominator. X counts more interactions as engagement than most people put in the formula, and the cheap ones inflate your number without meaning much.

The four that signal real attention are likes, reposts, replies, and bookmarks. Bookmarks are the one most calculators forget — they quietly became one of the strongest intent signals on the platform, because a bookmark means someone wanted to come back to your post. We include all four in the engagement rate our calculator reports.

Two of those four are worth isolating, because they tell you different things:

• Amplification rate (reposts / followers × 100) — how far your audience pushes a post past its own walls. High amplification is how small accounts reach people who do not follow them.
• Conversation rate (replies / followers × 100) — how much a post pulls people into talking back. This is the metric that actually predicts replies you can answer, which is the part of X that turns into pipeline.

If you want to pull these numbers for any public handle without exporting a spreadsheet, our free X analytics tool reads them straight off a profile.

Here is the table almost nobody publishes, and the reason the 1.8%-vs-0.029% fight exists. Engagement rate does not hold steady as an account grows — it collapses. Follower-based rates, by tier:

The average rate drops 20x from a sub-1K account (1.8%) to a 500K+ account (0.09%). That single fact reconciles the two benchmarks from the intro. Hootsuite's 1.8% is roughly the small-account average; RivalIQ's 0.029% median is dragged down by the large brand accounts that dominate its dataset. Both are real readings of different parts of the same curve.

For context on where the bottom sits: RivalIQ's report puts the cross-industry X median at 0.029% with a posting cadence of 3.31 posts per week, and even the busiest category — sports teams at roughly 41 posts a week — averages only 0.07%. Engagement on X is scarce by design, and the headline 1-5% numbers you see quoted are almost always small accounts on the follower-based formula.

The useful answer is a comparison against your own tier, not a universal threshold. Find your follower row in the table above and read across. Beat the average and you are ahead of the median account your size; clear the top-25% column and you are in good company; reach top-10% and the post genuinely outperformed.

A few honest reference points. If you are under 1,000 followers, anything above 1.2% is respectable and the bar for "good" is roughly 3%. In the 10K-50K range, the median is 0.3% — so a 1% post that would look mediocre on a tiny account is a top-10% result here. Over 100K, sustained rates above 0.35% put you ahead of most peers. The number that means "good" gets smaller as you grow, and that is normal, not decline.

The trap is benchmarking against creators who are not your size. A 50K-follower brand comparing itself to a 3K-follower creator's 4% will always feel like it is losing. They are running different races.

The single most common engagement-rate panic comes from comparing two formulas as if they were one. Someone reads that a "good" rate is 1-3%, opens their native X Analytics, sees 0.6%, and concludes the account is dead.

It is not. X Analytics reports the impression-based rate — engagements divided by views — while almost every benchmark you have read is follower-based. Those are different denominators, and the impression number is structurally smaller because a healthy post is seen by far more people than follow you. Comparing your 0.6% impression rate to a 2% follower benchmark is comparing a temperature in Celsius to one in Fahrenheit and worrying that the numbers do not match.

The fix is to compare like with like. Pick one formula and stick to it. If you want to measure against the published benchmarks in the table above, calculate the follower-based rate yourself from likes, reposts, replies, and bookmarks — the engagement calculator does it and tells you which tier percentile you landed in. If you only care about improvement over time on your own posts, X Analytics' impression rate is fine — just never hold it up against a follower-based number.

The follower-based formula has a denominator you do not fully control: your follower count. Every bot, dead account, and bought follower sits in that denominator and drags your rate down without ever engaging. An account with 10,000 followers where 2,000 are bots is really a 8,000-follower account doing 10,000-follower math, and its engagement rate is understated by a quarter.

It cuts the other way too. Bought engagement — fake likes and reply spam — inflates the numerator and produces a rate that looks healthy on a post no real person cared about. Either way, the metric is only as honest as the audience underneath it.

This is why we built bot detection into the product in the first place. When we score reply authors for bot likelihood with BotBlock, a meaningful share of the engagement on viral threads comes back flagged as automated. Before you read too much into a rate, it is worth knowing how much of the audience is real. You can spot-check any handle with the free bot checker, and the 30 signals that flag a bot cover what to look for manually.

Most "boost your engagement" advice is noise. Four things reliably move a follower-based rate, in rough order of impact.

1. Write for replies, not likes. Conversation rate is the metric that compounds — replies signal the algorithm that a post is worth surfacing, and they are the part of X you can answer to turn a viewer into a customer. Asking a real question beats posting a hot take that gets liked and scrolled past.

2. Tighten the post. The highest-engagement length on X sits around 240-259 characters — long enough for a complete thought, short enough to read in one glance. Padding to fill the limit lowers engagement. The mechanics of trimming without flattening the message are in writing tweets that fit.

3. Post when your audience is actually online. Impressions are the denominator on X Analytics and the ceiling on every other rate — a great post seen by nobody engages nobody. Our best time to post tool reads the windows where a handle's audience is most active.

4. Clean the denominator. Pruning or filtering bot followers raises a follower-based rate honestly, because it stops dead accounts from diluting it. The same monitoring habit that catches mentions worth replying to — covered in our social media monitoring guide — is what surfaces the conversations where engagement actually turns into pipeline.