The five metrics to actually track on X (and the ones to ignore)
Impressions are a vanity metric without context. Likes are noisy. Reply count includes spam. Here are the five numbers we track per slot every week, why they matter, and how to read them as a system rather than in isolation.
1. Follower growth quality (not just count)
Net followers per week is the headline number, but the real signal is the composition of those follows. A slot that grew by 200 followers and 195 of them are 0-post 0-follower accounts is going backwards. Track:
- Net followers per week.
- Of new followers, what % have more than 50 followers themselves and more than 10 posts. We call this the "active" ratio. Healthy is above 60%.
- Unfollows per week. A creeping rise in unfollows usually precedes reach drop.
If active ratio falls below 40% for two weeks running, the engagement strategy is attracting the wrong audience. Tighten author filters, narrow keywords.
2. Engagement ratio per post
Engagement ratio = (likes + reposts + replies) / impressions, per post. Track the median, not the mean — outliers from viral posts wreck the mean.
Reference numbers (median engagement ratio):
- Below 1.0% — content is not landing. Likely the topic or the format is wrong.
- 1.0%–2.5% — typical for a working account.
- Above 2.5% — strong. Scale this content shape.
If this drops sharply week over week without changes in posting, the slot may be soft-throttled. Pause for 48 hours and check.
3. Reply performance
For accounts running Reply (Search) or Reply (List), the second-order metric is replies that get engagement back. Track:
- Average likes per reply over the last 7 days. Should be at least 1.0 on a healthy account.
- % of replies that get any like at all. Above 30% is good.
- Click-through to your profile from the reply context. The X analytics panel shows this; it is the truest measure of whether a reply earned a new follower.
The trap: a 200-reply day with 0.1 likes per reply is worse than a 50-reply day with 2.0. Cap the volume to whatever your last 7 days can sustain at above-1.0.
4. DM open and reply rate
For slots running Welcome DM, the meaningful numbers are:
- Open rate — X does not expose this directly, but inbound replies are a strong proxy. A welcome DM sequence with a reply rate above 10% is doing well; above 20% is exceptional.
- Spam-mark rate — implied from delivery failures. If a slot's outbound DMs start landing in "message requests" and never moving out, the template is being filter-flagged.
- Time-to-reply distribution — if most replies come within 24 hours, the messaging is hitting active followers. Long tails suggest cold-side noise.
5. Outbound CTR
If the account links to anything (homepage, signup, blog post, calendar), outbound CTR is the only metric that proves the funnel works. Two ways to get it:
- UTM-tagged URLs in posts and bio. Read in your analytics tool, not on X.
- Per-link click counts in your own backend if the URLs are self-hosted.
Healthy outbound CTR depends entirely on the niche; a B2B SaaS account sending traffic to a pricing page might see 0.3% per impression and be very healthy. A content account sending traffic to a free guide might see 3%. What matters is the trend, not the absolute.
Metrics to ignore
These look like signal and are mostly not:
- Raw impressions. Without engagement ratio, impressions tell you whether X served you, not whether anyone cared.
- Total likes. Dominated by the one or two viral posts in the window.
- Profile visits as reported by X. The denominator is opaque and the number is noisy.
- "Reach" in X analytics. Same problem — calculated against a base that includes scroll-pasts.
Reading them as a system
The five metrics are not independent. A typical healthy pattern looks like: follower growth is positive, active ratio above 60%, engagement ratio in the 1.5–2.5% range, reply success rate above 30%, DM reply rate above 10%, outbound CTR steady or rising.
When one of those flips, look at the others. Engagement ratio drop with stable follower growth means the new followers are not engaging — usually an audience-mismatch problem. Follower drop with stable engagement ratio means the existing audience likes you but a recent post lost a chunk of them — that one is fine if intentional.
Track these in a weekly spreadsheet. The dashboard view is for daily sanity-checks; the spreadsheet is for noticing the shape of the trend.
Where to go next
For the configuration side of replies, see reply automation safety. For the posting side, see posting cadence for X.