Plausible export to dashboard
Plausible's tidy CSV, turned into a dashboard in a second.
Export your stats from Plausible, take visitors.csv out of the zip, and drop it here. The columns are lowercase and unambiguous, which is refreshing, but bounce_rate is a whole percent rather than a fraction, visit_duration is in seconds, views_per_visit is a ratio that cannot be averaged, and visitors is a daily unique count that must not be added up. This page handles each one.
Prefer the full explorer? Open the app, or start from the generic dashboard builder.
The cleanest analytics export there is, with three catches
Plausible does not track people, so its export has no user identifiers, no cookies and nothing that would make a privacy officer nervous. What you get is seven columns of daily aggregates in snake case, no metadata preamble, no formatted percentages and no currency symbols. After a week of wrestling with other analytics files I find this one almost relaxing to open.
The catches are all about aggregation. visitors is a count of unique visitors for that day, so summing thirty rows gives you a number bigger than your monthly unique count. views_per_visit is already pageviews divided by visits for that day, and averaging thirty such ratios weights a quiet Sunday the same as a launch Tuesday. Both problems disappear if you recompute from the raw counts, which is what happens here.
How to export from Plausible
- Open your Plausible dashboard Sign in at plausible.io, or at your own domain if you self-host, and select the site from the list. Self-hosted instances behave identically here, since the export format is the same code.
- Set the date range Use the period selector at the top right. Custom range lets you pick anything, and unlike most analytics tools there is no sampling to worry about at longer ranges, so a full year is a reasonable thing to ask for.
- Click the export icon The download arrow sits near the top of the dashboard. It exports whatever period and filters are currently applied, so a filter you forgot about will quietly shape the file.
- Unzip the download Plausible sends a zip containing several CSVs: visitors, pages, sources, countries, devices and others. Each one is a different breakdown of the same period, and visitors.csv is the daily time series.
- Drop visitors.csv on this page The daily trend is drawn immediately, with bounce rate and visit duration shown in readable units and the ratio columns recomputed rather than averaged.
The Stats API returns the same figures as JSON if you would rather automate it. Field names match the CSV columns exactly, so a flattened response lands on the same panels without any renaming.
Try with sample Plausible data (the same columns, 91 rows, no upload needed).
What the dashboard shows
Every panel below is built from a column that is actually in the export, named the way Plausible names it.
- Visitors and pageviews over time. date as the axis with both series, and visitors flagged as a daily unique count so nobody totals it by accident.
- Visits against visitors. The gap between the two is your returning traffic within a day, which is a surprisingly good early signal that content is being shared.
- Bounce rate trend. bounce_rate as a line rather than a single headline figure, since the number only means something when you can see whether it is moving.
- Time on site. visit_duration converted from seconds to minutes and seconds, weighted by visits when you aggregate up to a week or a month.
- Depth of visit. views_per_visit recomputed as total pageviews over total visits for whatever range you have selected, not averaged from the daily column.
- Day of week pattern. The same data folded by weekday, which for most B2B sites shows a shape you can plan a publishing schedule around.
date,visitors,pageviews,visits,views_per_visit,bounce_rate,visit_durationThree columns to read carefully, and a zip
- The export is a zip, not a CSV. It contains a file per breakdown. visitors.csv is the daily series; pages.csv, sources.csv and countries.csv are separate dimensions of the same period rather than extra columns.
- bounce_rate is a whole percent. A value of 47 means forty-seven percent, not four thousand seven hundred. Treat it as a fraction and every chart is off by a factor of a hundred in the other direction from the usual mistake.
- visit_duration is in seconds. A value of 138 is two minutes and eighteen seconds. It is the average for that day, so aggregating it over a week means weighting by visits rather than taking a plain mean.
- visitors is a daily unique count. Someone who reads you every morning is thirty visitors across a month of rows and one visitor in the monthly total. Sum the column and you have counted your most loyal readers the most times.
- views_per_visit is already a ratio. Averaging thirty daily ratios gives every day equal weight regardless of traffic. Recompute it as summed pageviews over summed visits for the range you care about.
- Filters travel with the export. The download honors whatever is applied on the dashboard, including a page filter you set an hour ago. There is nothing in the CSV to tell you a filter was on.
Three unit conventions and one aggregation rule, and that is the whole file. Plausible's export is what an analytics CSV should look like, and the only reason this page exists is that even a clean file has ratios in it that people keep averaging.
Frequently asked questions
Why does my Plausible download come as a zip?
Because the dashboard has several breakdowns and each becomes its own CSV. You get visitors over time, top pages, sources, countries, devices, browsers and operating systems, all for the same period. Unzip it and load visitors.csv here for the daily trend. The other files work too and give you the corresponding breakdown, they just do not carry a date column.
Is bounce_rate a percentage or a fraction?
A whole percent. Plausible writes 47 for forty-seven percent, unlike GA4 which writes 0.47 for the same thing. That difference has caught out more than one person moving between the two tools, because the resulting chart looks plausible in both directions until you check the axis. This page reads the Plausible convention and labels the axis accordingly.
How do I get monthly unique visitors from daily rows?
You cannot, not from this file. Each day's visitors figure is deduplicated within that day only, so summing counts anyone who returned more than once. To get a genuine monthly unique count, set the dashboard date range to that month and export again, letting Plausible do the deduplication over the whole period. Sessions and pageviews are additive; visitors never are.
What is the difference between visits and visitors?
A visit is a session, and visitors is the count of distinct people behind those sessions for the day. If the same person comes back in the afternoon that is two visits and one visitor. The ratio between the two is a decent daily engagement signal, and it spikes on days when something you published is being passed around rather than merely found.
Does self-hosted Plausible export the same format?
Yes. The community edition and the hosted service share the export code, so the column names, the units and the zip structure are identical. The only practical difference is where you sign in. If you have migrated between the two, files from either side merge cleanly and the trend will run straight through the switchover date without any renaming.
Is there any personal data in this file?
None. Plausible is cookieless and does not store identifiers, IP addresses or anything that would let you single out a person, which is the whole point of the product. The export is aggregate counts only. The file is still parsed entirely in your browser here, because there is no good reason to upload data anywhere when the computation can happen locally.
Related tools
Load a Plausible export and see the ratios done properly
91 daily rows with bounce rate as a whole percent, visit duration in seconds and a traffic spike partway through that shows why averaging ratios goes wrong.
Open the sample Plausible dashboard