The last 30 days, one row per UTC day, newest first.
Use this to watch the whole acquisition path day by day: site visitors, download clicks, real server downloads, new installs, week-one retention, signups, and purchases, all lined up so you can spot where a day fell off. All days are UTC, and today's row is partial (it's still going). A dash means we couldn't get that cell (not a zero). Click a day to filter the sections below to it.
| Day | Visitors | Download clicks | Server downloads | New installs | D7 retained | Signups | Purchases |
|---|---|---|---|---|---|---|---|
| 2026-07-18(today, partial) | 1 | 0 | 20 | 0 | – | 0 | 0 |
| 2026-07-17 | 15 | 3 | 67 | 1 | – | 0 | 0 |
| 2026-07-16 | 23 | 0 | 67 | 28 | – | 0 | 0 |
| 2026-07-15 | 6 | 2 | 63 | 20 | – | 0 | 0 |
| 2026-07-14 | 6 | 0 | 61 | 4 | – | 0 | 0 |
| 2026-07-13 | 10 | 4 | 44 | 34 | – | 0 | 0 |
| 2026-07-12 | 15 | 2 | 34 | 11 | – | 0 | 0 |
| 2026-07-11 | 12 | 4 | 39 | 6 | – | 0 | 0 |
| 2026-07-10 | 9 | 4 | 38 | 78 | 1%(1) | 0 | 0 |
| 2026-07-09 | 10 | 2 | 48 | 80 | 0%(0) | 0 | 0 |
| 2026-07-08 | 4 | 1 | 27 | 14 | 0%(0) | 0 | 0 |
| 2026-07-07 | 11 | 4 | 53 | 9 | 0%(0) | 0 | 0 |
| 2026-07-06 | 6 | 2 | 45 | 1 | 0%(0) | 0 | 0 |
| 2026-07-05 | 6 | 2 | 44 | 47 | 0%(0) | 0 | 0 |
| 2026-07-04 | 11 | 3 | 54 | 146 | 0%(0) | 0 | 0 |
| 2026-07-03 | 6 | 6 | 28 | 8 | 0%(0) | 0 | 0 |
| 2026-07-02 | 8 | 3 | 76 | 4 | 0%(0) | 0 | 0 |
| 2026-07-01 | 18 | 11 | 86 | 3 | 0%(0) | 0 | 0 |
| 2026-06-30 | 13 | 11 | 105 | 32 | 0%(0) | 0 | 0 |
| 2026-06-29 | 8 | 1 | 35 | 4 | 0%(0) | 0 | 0 |
| 2026-06-28 | 11 | 3 | 77 | 27 | 7%(2) | 0 | 0 |
| 2026-06-27 | 5 | 1 | 32 | 15 | 7%(1) | 0 | 0 |
| 2026-06-26 | 3 | 0 | 50 | 3 | 0%(0) | 0 | 0 |
| 2026-06-25 | 8 | 0 | 43 | 12 | 0%(0) | 0 | 0 |
| 2026-06-24 | 10 | 1 | 22 | 2 | 0%(0) | 0 | 0 |
| 2026-06-23 | 14 | 3 | 15 | 6 | 17%(1) | 0 | 0 |
| 2026-06-22 | 10 | 4 | 25 | 6 | 0%(0) | 0 | 0 |
| 2026-06-21 | 10 | 5 | 20 | 29 | 3%(1) | 0 | 0 |
| 2026-06-20 | 15 | 3 | 21 | 12 | 0%(0) | 0 | 0 |
| 2026-06-19 | 16 | 4 | 27 | 8 | 0%(0) | 0 | 0 |
Visitors and download clicks come from Umami (cookieless, in-browser). Server downloads, new installs, DAU, and D7 come from the app's own telemetry (D1); signups from Listmonk; purchases from Paddle. Clicks and server downloads won't match: server downloads also include Homebrew, direct links, and GitHub-page traffic, and bot user agents are filtered but imperfectly. D7 needs a cohort at least 8 days old, so recent rows show a dash there.
Use this to see which channels drove downloads: a download's ref is the channel the visitor first arrived from (a UTM source or campaign, else the referring site). Ref is first-touch per browser visit and comes from the URL only, so return visits and cross-device journeys (read on the phone, download on the Mac) carry no ref and land in "(none)". Homebrew and direct links have none too, and rows before 2026-06-12 predate the column. So treat "(none)" as "channel unknown", not "direct". All days UTC.
| Channel | Downloads |
|---|---|
| (none / unknown) | 1,277 |
| alternativeto.net | 39 |
| blog-smb2 | 18 |
| reddit.com | 9 |
| blog-mtp-mount | 7 |
| yandex.com | 7 |
| google.com | 5 |
| old.reddit.com | 1 |
| duckduckgo.com | 1 |
| github.com | 1 |
| youtube.com | 1 |
The Referer host of each /download request. Unlike the first-touch channel above (set only by the website button), this is captured on every hit, so it shows where the direct, no-ref downloads came from: a link on AlternativeTo, a directory, GitHub, Reddit, or a forum. "(none)" means no usable referer: a typed URL, a privacy browser, a referrer-policy strip, Homebrew/curl, or rows before 2026-06-25 that predate the column. A download can appear both here (by referer) and above (by ref); the two breakdowns count the same downloads differently. All days UTC.
| Referrer | Downloads |
|---|---|
| (none / unknown) | 1,270 |
| getcmdr.com | 68 |
| license.getcmdr.com | 28 |
Cmdr is macOS-only, so a non-macOS client downloading the .dmg literally cannot install it. "Human installs" is the raw server downloads minus those provably-impossible bot hits, so the headline stops reading as half noise. It keeps every ambiguous download (we only drop the clearly-fake ones). The scraper spoofs Mac browser UAs too (many from China), so "human" means "could be a real install", not proof of one, and is NOT a clean count. "unknown" (no or unrecognized UA, including rows before UA capture) stays counted because we can't tell. Only the bot row is the high-confidence exclusion. All days UTC.
Human installs: 873 of 1,366 raw server downloads
| Client | Downloads |
|---|---|
| human (Mac, Homebrew, curl/wget) | 617 |
| bot / impossible (Windows, Android, Linux, X11) | 493 |
| unknown (no / unrecognized UA) | 256 |
How many people see Cmdr content?
Use this to gauge top-of-funnel reach across the three sites, and which days or sites are growing. Umami is cookieless and proxied, so it dodges most adblockers, but it still undercounts by a few percent, and one person on two devices counts twice.
Total page views
205
-8.1%veszelovszki.com views
93
getcmdr.com views
111
getprvw.com views
1
| Source | Views |
|---|---|
| alternativeto.net | 6 |
| reddit.com | 6 |
| google.com | 3 |
| github.com | 2 |
| duckduckgo.com | 1 |
| chatgpt.com | 1 |
| youtube.com | 1 |
| Source | Views |
|---|---|
| chatgpt.com | 1 |
How many engage with the product page?
Use this to see getcmdr.com engagement over time, and to cross-check the two trackers against each other. Umami is cookieless and undercounts a bit; PostHog needs its own client-side script to load, so adblockers trim it more. Treat the two as independent estimates, not one exact number.
getcmdr.com page views
111
-19.0%Unique visitors
84
| Event | Count |
|---|---|
| download | 14 |
| Page | Views |
|---|---|
| / | 52 |
| /features/ | 27 |
| /pricing/ | 7 |
| /blog/ | 6 |
| /roadmap/ | 3 |
| /changelog/ | 3 |
| /blog/kindle-support/ | 2 |
| /refund/ | 2 |
| /blog/total-commander-for-macos/ | 2 |
| /blog/35-years-of-file-managers/ | 1 |
How many actually download?
Use this to see real new installs by source (website, Homebrew, other) and which release and platform people are grabbing. These are server downloads from the app's own endpoint, deduped per day by a daily-rotating hashed IP, with bot user agents dropped (imperfectly). It won't match the Umami download clicks above, which fire in-browser. Rows before 2026-06-11 have no source attribution.
New installs (deduped)
390
Download requests (raw)
395
Counts downloads of the macOS DMG through getcmdr.com (download endpoint), deduplicated to distinct people per day via a daily-rotating hashed IP, with bot and link-preview hits dropped by user agent. Website = the getcmdr.com download button, Homebrew = `brew install --cask cmdr`, Direct / other = links shared elsewhere. In-app auto-updates never count here (they fetch from GitHub, not this endpoint). Hover a bar for exact numbers.
Scroll to zoom timeline
| Architecture | Downloads |
|---|---|
| universal | 138 |
| aarch64 | 136 |
| x86_64 | 121 |
| Country | Downloads |
|---|---|
| Vietnam (VN) | 49 |
| Brazil (BR) | 33 |
| Argentina (AR) | 21 |
| Pakistan (PK) | 19 |
| United States (US) | 19 |
| Bangladesh (BD) | 18 |
| Iraq (IQ) | 13 |
| Sweden (SE) | 13 |