Venbit Intelligence Research

The numbers in our guides come from websites tracked in the Venbit Approved System. This page explains what we count, how each figure is made and where the numbers stop being reliable.

Days in a release
90
Visits from people
229,329
Bot visits set apart
146,638

Which websites are tracked

Every website in the research runs Venbit Intelligence and is approved in the Venbit Approved System before any of its figures are used. They are small and mid-sized business websites in the United States, most of them local service businesses, along with some software, content and business-to-business sites.

All of them are measured by the same tracker with the same definitions, which is the point of the exercise. Two websites counted by two different tools cannot be compared. Websites counted by one tracker in one way can.

What we count

Visit
One stay on a website by one browser, from the first page to the last. A person who returns the next day makes a second visit.
Lead
A form sent, a phone number tapped, or the goal that website set for itself, such as a booking page reached. Leads per 100 visits is leads divided by visits.
Source
Where the visit began: an ad, a search engine, a social network, an AI assistant, an email, a link on another site, or direct when the visit carries no sign of its origin. The groups follow the channel definitions Google publishes.
Runs ads
A website counts as advertising when a fifth or more of its visits in the period were paid clicks.
Bot
A visit made by software. Bots are identified, recorded apart and left out of every figure about people. Bot share is bot visits divided by all recorded visits.
Reached contact
The share of visits that opened a contact page or had a contact form come into view. A phone tap is counted as a lead, not as reaching contact.

How a figure is made

  1. 01Each release covers a fixed 90 days and states its dates. The current one runs from 21 August to 18 November 2026.
  2. 02A website enters a release only after 28 days of tracking inside the period and at least 30 visits from people.
  3. 03Monthly figures are the period's total scaled to 30 days. For a website that joined partway through, the days it was tracked are used.
  4. 04We report the median, which is the middle website when all are lined up. One very busy website moves an average a long way and leaves a median where it was. For all tracked websites together we also give the middle half, from the lower quarter to the upper quarter.
  5. 05Groups by kind of business hold fewer websites than the whole, so they are read as a comparison and not as an industry average. A choice that too few websites match is not described at all.

How bots are kept out

A figure about visitors is only as good as its bot filter. Venbit Intelligence judges every visit on what the software calls itself, where the connection comes from and how the visit behaves, and a visit judged to be automation is recorded apart from people. A suspicion is withdrawn when the visitor does something a person does, such as filling in a form.

We publish what the filter looks at and keep the thresholds private, because a published threshold is a target. The guide on how much website traffic is bots shows the signals on your own visit.

What is never published

No website is named and no figure is published in a form that points to one business. The research holds totals for each website and nothing about any person: no names, no messages, no phone numbers and no addresses. Venbit's own websites are left out of every release.

Limits of the research

The research is a sample, and it leans toward local service businesses. A business that tracks its website closely is more active online than one that does not, so these figures sit above what a census of every small business would show. The smaller groups hold few websites, so read them as a comparison and not as an industry average.

Crawlers that read a page without running its script are invisible to any script-based tracker, ours included, so the bot share we publish is a floor. Towns are estimated from the connection and are approximate. A phone tap shows that someone reached for the phone and says nothing about whether the call connected.

Use these figures

You are welcome to quote the research. Credit it as Venbit Intelligence Research with the dates of the release and link to the guide the figure came from. For a question about the method, get in touch.