X Analytics 2026: 14 Metrics That Predict Growth

Photo by dlxmedia.hu on Unsplash

In January 2026, X quietly retired the algorithm that ranked your posts for the last decade.

The legacy "Heavy Ranker" was replaced with a Grok-powered transformer that reads every post, watches every video, and makes roughly 5 billion ranking decisions per day — each in under 1.5 seconds. Two months later, the creator monetization pool was rewritten. The dashboard looks the same. The numbers behind it don't.

Most X analytics guides still teach the 2023 metric stack — impressions, engagement rate, profile visits. That stack was built for a platform that doesn't exist anymore.

X analytics 2026 means tracking a small set of leading indicators — view-to-follower ratio, reply-to-follow conversion, dwell time, and first-30-minute velocity — that predict whether your account grows over the next 30 days, instead of describing what already happened.

This guide walks through the 14 metrics that actually matter, what counts as a good benchmark today, and a 40-minute Sunday review you can run without leaving the dashboard.

Updated May 2026 to reflect the post-Grok ranking model and the late-2025 monetization rewrite.

What's new in X analytics since the Grok rewrite

If you opened your dashboard in February 2026 and felt like the numbers got harder to read, you weren't imagining it. Three things shifted in the last six months — and the old playbooks haven't caught up.

Ranking is semantic, not signal-based. The Grok-powered transformer reads your post in context and asks "is this useful to this user?" — not "did the user follow this account?" or "does it have the right hashtag." PostEverywhere's algorithm review puts the per-decision compute budget at under 1.5 seconds across about 5 billion decisions a day.

Monetization moved off ad share. Social Media Today reported the late-2025 retirement of ad revenue sharing. Creators are now paid from a Premium subscription pool — verified-user impressions count, others don't. Same impressions number on your dashboard, completely different payout math behind it.

Reach is decoupling from follower count. Median engagement on X has collapsed to about 0.10% platform-wide, per Hootsuite's 2026 benchmarks. Premium-boosted accounts run around 0.4%. Sub-10K accounts now routinely outperform 100K accounts because the For You feed rewards dwell time and reply velocity — not historical authority.

If your dashboard is the same one you used in 2024, the numbers haven't changed. What they predict has.

The 3-Layer Growth Stack: 14 metrics that actually predict growth

Every analytics dashboard hands you the same 30 numbers. The 3-Layer Growth Stack is the framework that sorts those numbers by what they predict — not what they describe.

Layer 1 is reach. Distribution metrics tell you whether the algorithm is letting your work get seen at all. They move first.

Layer 2 is engagement. Behavioral metrics show whether the impressions you get are actually catching attention. They move next.

Layer 3 is conversion. Outcome metrics translate engagement into followers, signups, and revenue. They move last — and they're what you actually care about.

Track all 14, but weight them in that order. A spike in Layer 3 with no Layer 1 movement is noise. A Layer 1 spike with no Layer 2 follow-through is a sign that your hooks are weak. Read them in sequence, not in isolation.

Two of those metrics — reply-to-follow conversion and dwell time — are where most accounts under-invest. We'll dig into both later.

For impression mechanics specifically, the deep version of Layer 1 lives in our breakdown of how to get more impressions on X in 2026.

Layer 1: Reach metrics that show distribution health

Reach metrics answer one question: did the algorithm decide your post deserved a seat in the For You feed?

Impressions per post is still the headline number, but read it as a ceiling, not a goal. The average X post earned 2,711 impressions in 2025, down about 5% from 2,864 the year before. If you're below 1,000 impressions on most posts, you're not getting algorithmic lift — you're showing to your followers and stopping there.

Reach (unique users) is the Premium-only cousin of impressions. The same person seeing your post twice counts as two impressions but one reach. A high impressions-to-reach ratio (above 1.4×) means your post is getting served back to people who already saw it — usually because dwell time was high.

View-to-follower ratio is calculated, not displayed. Divide views by follower count. A ratio over 1.0 means your post reached more people than follow you — the For You feed picked it up. Sub-10K accounts should aim for 2–5× on their best posts. If every post lands at 0.3×, the algorithm has decided you're follower-only.

Profile visit ratio is profile visits divided by impressions. Healthy is 0.5–2%. Lower means your post earned attention but your name or hook didn't earn a tap. That's a profile problem, not a content problem.

Layer 2: Engagement metrics the algorithm actually rewards

Engagement metrics are where the new ranking model shows up most clearly. The 2026 Grok-powered model weights certain actions far more than others — and treats some legacy metrics as near-noise.

Weights pulled from Sprout Social's 2026 algorithm breakdown and PostEverywhere's source-code review.

The takeaway: if you're chasing likes, you're chasing the lowest-weighted action on the platform. The deeper context lives in our breakdown of the 2026 X algorithm.

Engagement rate is still worth tracking, but as a relative signal — your account against itself, week over week. Cross-account comparisons are misleading because Premium-boosted distribution moves the denominator.

Reply rate (replies divided by impressions) is the single most actionable engagement metric in 2026. A 1% reply rate on a 5,000-impression post is a stronger growth signal than a 5% like rate. Replies pull the algorithm's biggest weight.

Bookmark rate is the closest thing X has to a save-worthy index. The private "personal hub" rolled out in 2025 made bookmarks more private but kept the 10× weight intact. If your bookmark rate is above 0.5%, your content is being treated as reference material.

Dwell time is the quiet 2026 winner. Posts that hold attention for more than 3 seconds get ranked far above quick-scroll content. This is now visible in Premium analytics.

Share rate (reposts ÷ impressions) is the network-effect metric. Anything over 0.3% pushes your post into second-degree feeds.

Layer 3: Conversion metrics that translate engagement into outcomes

Layer 3 is where strategy meets accounting. These are the metrics that translate Layer 1 and 2 into followers, signups, and revenue.

Reply-to-follow conversion is the most under-tracked metric on X. It measures how many new followers you earn per 100 replies you ship. It's not on the native dashboard — you calculate it weekly. Most accounts that grow systematically run between 1.5 and 4 new followers per 100 replies. Below 1, your reply quality or targeting is off.

Follower growth velocity is followers gained per week, not total followers. A 1,500-follower account adding 50/week is on a steeper trajectory than a 15,000-follower account adding 100/week.

Profile CTR is profile visits divided by impressions — same number as in Layer 1, but read here as a conversion signal. Going up means your name + hook combination is doing the qualification work for free.

Outbound link clicks are still the cleanest funnel handoff. The catch: link-in-body posts lose an estimated 30–50% of initial reach. The 2026 workaround is to put the link in reply #1 of your own post. Track clicks in the destination tool (UTMs in Plausible, Fathom, or GA4), not on X.

Revenue per million impressions (RPM) is the new monetization north star. Creator revenue share now averages about $8.50 per million verified-user impressions, with creators keeping up to 97% of the pool until $50K lifetime, then 90%. RPM is what tells you whether Premium investment pays.

What's a good engagement rate on X in 2026?

Benchmarks shift by follower tier — and the platform median is much lower than most guides admit.

Benchmarks aggregated from WebFX's 2026 X marketing benchmarks, Hootsuite 2026, and Buffer's 2026 social media benchmarks.

The pattern: engagement rate falls as accounts scale. A 0.5% rate on a 100K account is the same predictive signal as a 1.5% rate on a 5K account. Don't compare yourself to creators a tier above or below — and don't trust generic "what's a good engagement rate" articles that quote a single number.

For the formula, when to recalculate, and how to handle reply-heavy accounts where engagement happens off your own post, read the longer X engagement rate guide for 2026.

The reply-to-follow conversion metric nobody tracks

Here's the contrarian take that runs against every X analytics 2026 guide currently ranking on Google: engagement rate is now a backward-looking vanity metric. Reply-to-follow conversion is the leading indicator that actually predicts growth.

The reason is structural. When you reply on a 50,000-follower account's post, your engagement shows up under their impressions — not yours. Your engagement rate stays flat. Your reach explodes. The native dashboard is bad at this exact scenario, which is the most common growth tactic on X in 2026.

Jack Butcher, who built Visualize Value into a 1M+ follower property, summed it up bluntly in a conversation with Nathan Barry:

"I think I'm 10% analytics, 90% gut. A few rough frameworks I use for knowing how viral something is off the bat though: over 20% engagement/impressions on Twitter is going to fly."

Translation: he ignores most of the dashboard and watches the ratio that actually predicts breakout. For reply-led growth, the equivalent ratio is new followers earned per 100 replies shipped.

To track it, you need three numbers each week:

• Total replies posted

• New followers gained

• Replies that came from accounts above your tier (where the algorithm leverage lives)

Divide new followers by replies, multiplied by 100. Healthy reply-to-follow conversion is 1.5–4 new followers per 100 replies. Below 1, you're either replying on accounts too small to lift you or shipping replies that don't earn the tap-through.

This is where reply workflow tools matter. ReachMore's AI Reply and Reply Intents features let you draft on-voice replies in seconds, and the viral probability score on each suggestion is essentially a per-reply prediction of which one will earn the engagement that converts. Track your reply-to-follow conversion weekly, and your reply queue stops being guesswork.

For the upstream half of this loop — picking which tweets are even worth replying to — see how to find tweets to reply to on X.

Dwell time and 30-minute velocity beat follower count

The second 2026 contrarian: follower count is officially decoupled from reach. The variables that actually predict the next 30 days are average dwell seconds per impression and first-30-minute engagement velocity.

Both are weighted higher than follower count in the new ranking model. Dwell time carries a +10 unit weight — equal to a bookmark and half a retweet. A 500-follower account that holds attention for four seconds per impression in the first half-hour can out-reach a 50,000-follower account that gets scroll-past likes.

The clearest public case study is Justin Welsh's "How I Added 44,716 Twitter Followers in 18 Weeks." Welsh credited the delta to two ignored metrics: reply impressions on accounts within 2× his follower count, and thread completion rate (the long-form equivalent of dwell time). His tactic was concrete — 2 original posts per day, 1 thread per week, 45 minutes every morning replying to peer-tier accounts.

Before/after:

• 18 weeks earlier: ~95K followers, stagnant growth

• 18 weeks later: ~140K followers, ~2,500 new followers per week sustained

• Metrics shift: stopped tracking likes, started tracking reply impressions and thread completion

What he didn't change was posting volume. He changed which numbers he watched. The dashboard didn't reward him — the new measurement habit did.

For the velocity side, watch your first-30-minute engagement rate as a separate number. Pull impressions and engagement counts at the 30-minute mark and again at 24 hours. If 30-minute engagement is over 40% of 24-hour engagement, your post hit the For You feed. If it's under 15%, the algorithm capped distribution early and you're showing to followers only.

The 40-minute Sunday X analytics 2026 review (copy-paste checklist)

Most accounts overcomplicate analytics by checking the dashboard daily and chasing noise. A 40-minute weekly review beats a 5-minute daily scroll every time. Copy this checklist into your weekly planner.

Block 1 — Reach (10 minutes)

• Pull impressions per post for the last 7 days. Note the median, not the average — outliers lie.

• Calculate view-to-follower ratio for your top 3 and bottom 3 posts. Flag any post over 2× as a hook to replicate.

• Note profile visit ratio. If below 0.5%, your hook or pinned tweet needs work.

Block 2 — Engagement (10 minutes)

• Reply rate per post (replies ÷ impressions). Anything over 1% is a strong post.

• Bookmark rate on your top 5 posts. Above 0.5% means save-worthy — turn it into a thread.

• Dwell time on your top and bottom posts (Premium). Identify which formats hold attention.

Block 3 — Conversion (10 minutes)

• Calculate reply-to-follow conversion: new followers this week ÷ total replies shipped × 100.

• Follower growth velocity: net new followers / 7. Compare to last 4 weeks' rolling average.

• Outbound link clicks: pull from your destination tool. Cross-check against any link-in-reply posts.

Block 4 — Plan the next week (10 minutes)

• Pick the top-performing format and commit to 2 more posts in it.

• Identify 5 peer-tier or above-tier accounts you'll reply to consistently this week.

• Set a reply target. The math from our how many replies per day on X breakdown holds in 2026.

This is also where reply tooling earns its keep. During the week, ReachMore's AI Reply drafts and viral probability score let you skip the "is this reply good enough" question — the score handles qualification — and the Drafts feature lets you queue replies to peer-tier accounts without context-switching out of work. The Sunday review then becomes a measurement of what shipped, not what you wish you'd shipped.

What to do when X impressions drop suddenly

A sudden 30–50% drop in impressions is the single most common panic moment on X. The dashboard doesn't tell you why. The diagnostic does.

This is also where macro-level metric thinking pays. Rand Fishkin made the broader point about modern measurement in a Skyword interview:

"Traffic is a vanity metric in a zero-click world. When AI overviews and platform algorithms satisfy user intent without a click, optimizing for website visits misses where brand influence is actually built."

The same logic applies to X impressions. A drop in raw impressions matters less than what happened to reply quality, profile CTR, and follower velocity in the same window. If the leading indicators held, the dashboard will catch up.

X analytics 2026 FAQ

How do I check my X analytics for free in 2026? Free accounts still get basic per-post analytics — impressions, engagements, profile visits, link clicks — by tapping the bar-chart icon under any of your own posts. Account-level summary, reach (unique users), and dwell time are Premium-gated. For a free baseline, scrape your own posts weekly into a spreadsheet and calculate the calculated metrics (view-to-follower ratio, reply rate, profile CTR) yourself.

What is a good engagement rate on X in 2026? Platform median sits around 0.10%, per Hootsuite's 2026 benchmarks, but tier matters. Under 1K followers, target 1.5–2% sustained. Between 10K and 100K, target 0.8–1%. Above 100K, anything above 0.3% is above-median. Compare yourself only to your own 4-week rolling average — cross-account comparisons mislead because Premium-boosted distribution shifts the denominator.

Why did my X impressions drop in 2026? Most sudden drops trace to one of six causes: external-link penalty (~30–50% reach reduction), tier-mismatch replies, velocity decay after a quiet week, post-viral throttle, low thread completion, or a video upload issue. Run the diagnostic table above. If all six pass and you've lost ~80%+ of impressions, run a shadowban check.

Does X Premium increase reach? Yes, measurably. Premium accounts show roughly 4× the median engagement of free accounts in Hootsuite's 2026 benchmarks, but the lift is not automatic — Premium amplifies whatever content quality and reply velocity you already have. If your free-tier baseline is weak, Premium won't fix it. Our Is X Premium worth it in 2026 breakdown does the creator math.

Are bookmarks more valuable than likes on X? Yes — by roughly 10×, per the post-Grok ranking weights. A bookmark signals "save-worthy" and predicts future reach because the algorithm reads it as a strong endorsement. Track bookmark rate (bookmarks ÷ impressions) as a primary engagement metric. Above 0.5% means your content is being treated as reference material.

How do I track reply performance on X? The native dashboard shows replies-received but not replies-shipped. Track three numbers manually each week: total replies you sent, new followers gained, and replies that came from accounts within or above your follower tier. Divide new followers by total replies × 100 for your reply-to-follow conversion rate. Healthy is 1.5–4 followers per 100 replies.

What's the difference between views and impressions on X? Views count every time the post appears on a screen — including the same person scrolling past twice. Impressions count the same way for posts but the term is sometimes used interchangeably with "viewed for at least 500ms." The 2026 dashboard standardized on impressions for posts and views for video. Reach (unique users) is the cleaner number, but it's Premium-only.

How many followers do you need to get monetized on X in 2026? The creator revenue share program requires X Premium, 500+ verified followers, and 5 million organic impressions over the past 3 months. Payouts come from a Premium subscription pool at roughly $8.50 per million verified-user impressions, with creators keeping up to 97% of the pool until $50K lifetime earnings, then 90%. Verified-user impressions are the metric that pays — not raw impressions.

The bottom line on X analytics in 2026

Three numbers do most of the work. View-to-follower ratio tells you whether the algorithm is letting you out of your follower bubble. Reply-to-follow conversion tells you whether the replies you ship are earning followers. Dwell time tells you whether the impressions you earn are actually catching attention. Track them weekly, weight them in that order, and most of the rest of the dashboard becomes noise you can safely ignore.

The 14-metric stack above is the full picture. The 3-Layer Growth Stack is the framework for reading them. The Sunday review is the habit that makes both useful. Skip the daily check-in — it's the single biggest time sink in X analytics 2026 and almost none of the resulting decisions are good ones.

If reply-led growth is your main channel, the measurement loop only works if the reply work is fast enough to actually run consistently. Want to turn every reply into reach? Install ReachMore for Chrome →