Overview

Five mutually exclusive populations, the four clocks side by side, and top pages you can scope.

The first two questions, answered side by side: who was here, and how long they actually stayed. Four independent cards, each fetching its own data so the fast ones appear first.

01 · The four cards
CardWhat it asks
01 Who was hereFive mutually exclusive population counts, plus a distinct-visitor estimate and totals for the human half
02 How long they actually stayedThe four clocks as an average and a median over human sessions that have beacon data, and the log-derived figure over all human sessions
03 Sessions over timeOne range facet by hour with the five populations nested inside each bucket
04 Top pagesThe paths, re-run when you switch between humans, everything and bots
Card 02 is the one to look at first

It is the only place you will see the industry number and the honest number next to each other on your own traffic. On most sites the gap is large enough to be worth a second look.

02 · The five populations

Every count on this page splits into five mutually exclusive buckets, and they are the same five everywhere in the panel:

HUMANSand likely humansUNKNOWNnot enough evidenceDECLAREDhonest crawlersAI CRAWLERSalso honestEVASIVE BOTSthe one that mattersThe five colours are a greyscale ramp with the accent reserved for evasive traffic, so the eye goesto the one band that represents a cost you are paying rather than a service you want.

Figure 1 — the population ramp. Declared and AI crawlers are bots and are counted as bots; they are kept visually apart from evasive traffic because conflating the two is what makes other tools useless here.

03 · Every number states its population

Each chart and each stat block carries a visible caption naming what it counts — and it is a paragraph rather than a tooltip, so it survives being screenshotted. An unlabelled number is treated as a lie.

  • Absent stays absent. A metric with no data renders as a dash, never as zero. A session with no beacon shows a log span and three dashes, because nothing measured the other three.
  • The timing card covers a smaller population than the count card, deliberately — only sessions with beacon data can have the three measured clocks — and the caption says so.
  • The distinct-visitor figure is approximate above about a hundred, and its caption says that too.
  • Every timestamp renders as mm/dd/yyyy hh:mm:ss in the timezone you set under Display; Solr stores everything in UTC.

Top pages has a humans/everything/bots toggle for a reason: a top-pages table that silently includes crawler traffic is the most common way an analytics dashboard misleads somebody about what their audience reads.

Clicking any value anywhere scopes the whole dashboard — see filters and facets.

Loghound is open source and MIT licensed. Questions about the Opensolr half — the account, the indexes, the plan — go to opensolr.com/contact; questions about the software itself belong on GitHub.

Loghound Documentation