More than half the web
isn’t human anymore.
In 2024, automated traffic crossed 51% of all web traffic for the first time in a decade. Your reports still count most of it as people, because the platform selling you the traffic is the one that decided what a person was.
51% automated, 37% bad bots: Imperva 2025 Bad Bot Report (2024 data). Roughly 30% automated, with Google’s own crawler the single largest source: Cloudflare Radar 2025.
They can’t see it. The decision was already made.
- ×Applied upstream, once, before anyone watched the visitor
- ×Decided by the platform that profits from the traffic
- ×Counted into your conversion rate as a person
- ×Never audited, because the fraction’s bottom is invisible
- ✓Taken from how the visitor actually moved and read
- ✓Sampled densely, in real time, on the page itself
- ✓Counted separately from the bots, then compared
- ✓Reported to the dealer, the one party with no reason to lie
A label is not a measurement. Behavior is.
Every bot click spends real money and buys nothing.
A bot that clicks your ad costs you exactly what a buyer’s click costs. It just never buys. On a web that has crossed half automated, a meaningful share of the clicks you are billed for were never going to convert, because nobody was ever there.
The deeper cost is the one you can’t see. Those wasted clicks still count as visits, so they inflate the traffic number your campaign is judged by. A campaign that sent you bots can look as busy as a campaign that sent you buyers. The rate hides the rot.
The only way to know your number is to count the people separately. That is the one thing your stack will not do for you.
Bots on the web, answered straight.
How many bots are on the web?
In 2024, automated traffic crossed half of all web traffic for the first time in a decade. The Imperva 2025 Bad Bot Report put automated traffic at 51% of the total, with bad bots alone at 37%. Cloudflare Radar, measuring differently, reports roughly 30% of traffic is automated and notes that Google’s own crawler is the single largest source of it. Either way, a large and growing share of the visitors counted in your reports were never people.
Why does my analytics not show this?
Because your analytics decided who was a person before it did the math. The same platform that sells the traffic also defines what counts as a real session, and it has no incentive to shrink the number it bills against. A conversion rate is a fraction, and almost nobody audits the bottom of it. The bot share that slips through is counted as human, quietly, and folded into the rate you trust.
What are bots costing my ad budget?
Every bot that clicks a paid ad spends real money and produces nothing. On paid campaigns we measure, it is common to find that a meaningful share of the clicks billed as visits never came from a person, and that the humans who did arrive behaved completely differently from the bots. The cost is not just the wasted clicks. It is that the wasted clicks inflate the traffic number your campaign is judged by, so a failing campaign can look healthy. The only way to know your figure is to count the people separately.
Is Google wrong about what a bot is?
Google is the default authority on what counts as a bot at the reporting layer, and that authority is a single judgment made once, upstream, by the party that profits from the traffic. It is not measurement of the visitor’s actual behavior. It is a label applied before anyone watched what the visitor did. A label is not a measurement. Behavior is the measurement, and behavior frequently disagrees with the label.
How does Seer tell a bot from a person?
Seer does not trust a declared label. It reads behavior directly: how a visitor moves, reads, hesitates, and acts on the page, sampled densely in real time. A person and a script leave different traces, and those traces do not lie the way a self-reported source does. Seer counts the humans from the way they behave, not from what the ad platform claimed they were, and reports the difference to the dealer every morning.
Where can I see the bot rate on my own traffic?
Inside the product. Every Seer tenant gets their own floor: the humans and the bots counted separately, per campaign, per day, with the paid sources scored on whether they sent buyers or bots. It arrives as the Daily Seer, a short morning read sent to the principal. This page shows the industry picture. The product shows yours.
Ask Seer how many of your visitors were real.
This page is the industry picture. Your picture is one floor down: your humans and your bots, counted separately, per campaign, every morning. No form. Just ask it.
We don’t use forms. We don’t believe in them. Schedule with Seer.