Overview -> Traffic

Traffic answers who came to your pages over a stretch of time you choose, where they came from and what they did once they were there. Under the figures stands what they are made of, and a click on any part of it narrows the whole screen to that part. Further down is what visitors did and how fast your pages answered, and on the right who is on your pages right now.

Everything is measured on our own servers, not with a script in the browser. Nothing is blocked by an ad blocker, no cookie is set, and nobody is followed from one day to the next, so there is no consent banner between you and your numbers. Visits by Propper staff to your environment are never counted.

The period

The buttons in the header pick the period: today, the last 7, 30 or 90 days, or the last 12 months. Custom opens two dates. Both days count, in the timezone of your firm. Every period is compared with the stretch of the same length right before it, and the line under the buttons says which dates that is. Today is compared with yesterday up to the same hour.

Every page view is kept one by one for 90 days. Older days are kept as a summary per day and per page. A period that reaches past those 90 days says so in the header, and on it only the pages break the visitors down (see What older periods can show).

The period, the groups and every filter are in the address, so a reload or a link to a colleague opens on the same figures. The screen also remembers the last period you looked at.

Whose visits

Filter in the header picks whose visits count: people who are signed out (prospects, and traders before they sign in), traders, and your team. Your team is left out unless you pick it, so your own work in the admin panel never counts as traffic. Anyone with a role on your team counts as team on every page they open, not only in the admin panel.

The same menu lists the filters that narrow the screen, each with a way to take it off, and Clear filters puts everything back.

Visitors

Four figures sit above one chart, and clicking one draws it:

Visitors: different people who opened a page. A person is recognised within one day only, so somebody who comes back tomorrow counts again. That is what keeps the pages free of cookies, and it means a visitor here is a visitor-day.
Page views: every page that was opened, reloads included, with the pages per visitor and the number of sessions under the chart.
Sign-ups: accounts made in the period. A sign-up is tied to the visit it came from, so a filter on a source or a country also narrows the sign-ups. Sign-ups that cannot be tied to a visit count in the total and are named under the chart.
Conversion: sign-ups as a share of the visitors who arrived signed out, since somebody who was already signed in cannot sign up.

Each figure shows its change against the period before. Only visitors, sign-ups and conversion are coloured, green when they go up. More page views is neither good nor bad, so it stays in the colour of the text. The dashed line is the period before.

What they are made of

Under the chart, one breakdown at a time:

Sources: where a visitor came from on their first page of the day. The utm_source of the link when there is one, otherwise the site that linked to you (www. and the mobile m. left off), otherwise Direct.
Campaigns: the utm_campaign of the link a visitor arrived through.
Pages: how many visitors saw each page. A visitor who saw three pages counts on each, so the pages add up to more than the total. When two rows share a name and an address, usually one page recorded under an older and a newer name in the product, each says which part it belonged to: the public site, the trader app or the admin panel.
Countries, Devices, Browsers and Systems: where the visitor was, from their connection, and what they arrived on.

Each row shows its visitors, page views, sign-ups and conversion, and its change in visitors against the period before. The list is sorted by the figure picked above the chart and can be sorted from its column heads and searched. It holds the 500 largest rows.

A click on a row narrows every figure on the screen to it: only visitors from Google, only visitors who saw the pricing page, only phones. A source, campaign, country, device, browser or system is read from the visitor's first page of the day, so "from Google" means visitors who arrived from Google, and everything they did afterwards counts with them. A page narrows to the visitors who saw it. Click the row again, or the cross beside the filter in the header, to show everybody again. A right click narrows as well, opens the page or the site in a new tab, or copies the name or the address.

What visitors did

The actions people took, beside the pages they looked at. Key actions shows the ten that matter most, each always listed even when it did not happen: signed up, signed in, bought a challenge, was given a challenge, requested a payout, opened a ticket, reset their password, failed to sign in, was locked out and added a passkey. All shows every action the activity log knows that happened in the period, by its code.

Each row says how often it happened, by how many different people, and the change against the period before. A person is their account, or the visitor for an action taken signed out, such as signing up. Actions by your team only count when your team is picked, with one exception: a challenge your team gives a trader always counts, and the person behind it is the trader who got it. A filter on visits narrows the actions to the visitors it leaves in, as long as the period is within the last 90 days. A row opens the activity log on that action and those dates.

Speed and errors

How your pages held up for the visitors above, as four figures over one chart:

Average time: how long the server took to answer a page, from the request to the answer. What the browser does after that is not measured.
Slowest 5%: one page view in twenty took longer than this. It shows the slow moments an average hides.
Refused: the share of page views answered with not found, not allowed or an expired session. This is often an old link or a visitor whose session ran out, rather than a fault.
Failed: the share of page views that failed on our side. Each one is a fault.

For these four a rise is bad news, so up is drawn in red and down in green.

Under the chart is every page with its views, both times and both kinds of error. Errors rank by the share of a page's views that failed, not by their count, so a quiet page that fails every time is not hidden behind a busy one with a handful of errors. Hovering an error share shows the count behind it. A click narrows the whole screen to that page.

Right now

The column on the right does not move with the period. It shows the people on your pages in the last 30 minutes, minute by minute, and lists the pages they are on, where they came from, and the key actions of the last 24 hours with who they were about: who signed in or signed up, or the trader your team gave a challenge to. When a page failed in the last hour, it says how often and on which page. It follows the groups picked in the header, and no other filter.

Each list has a search and a pager, and holds the 200 busiest pages or sources, or the 200 newest actions, with a line saying so when there were more. It refreshes every minute. Page views reach it in batches, so it can run about a minute behind. J and K walk its list and Enter opens the lit row: a page or a source narrows the screen, an action opens its entry in the activity log.

What older periods can show

Every page view of the last 90 days is kept, so everything on this screen works there. Before that, a summary per day and per page is all that is kept, which changes what a period that reaches back further can show:

The visitors break down by page only. Sources, campaigns, countries, devices, browsers and systems cannot be read, and a filter on one of them is set aside with a note in the header and a way to take it off.
Visitors per page add up the visitors of each day, which is the same count as on the rest of the screen.
Under a filter on a page, sign-ups and conversion cannot be tied to that page and show no value.
The slowest 5% is estimated from each page's daily figure, weighed by its views, and per page refused and failed are one number.

Actions are kept one by one for two years, so they follow the period exactly. When a period reaches before that, people are counted once per day.

Export

Export downloads the period, the groups and the filters on screen as a CSV file, in one of four cuts: every figure with a line per point of the chart, the breakdown under the chart with every row, every page with its times and errors, or every action. Every download is written to the activity log.

Who sees what

Opening Traffic takes the permission to view analytics, which also opens Revenue. The header names the tracking tools your firm connected, with a link to the tracking settings for a role that may open them. In the right column, a trader is named only for a role that may view traders and a member of your team only for a role that may view the team, and an action only opens the activity log for a role that may view it.

In the older designs

A firm on the OG or Tactical design keeps the page it had: page views, visitors, sessions, the average load time and the error rate against the period before, a chart of views and visitors, the split between signed out visitors and traders, the most opened pages, the failing pages ranked by their error rate, the actions, and over shorter ranges the referrers, campaigns, devices, browsers and countries. It has no custom dates, filters, conversion, slowest 5%, right now column or export. Those are only in the new Default.