> For the complete documentation index, see [llms.txt](https://docs.adpage.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.adpage.io/faq/debugging-monitoring/identifying-and-excluding-bot-traffic.md).

# Identifying and excluding bot traffic

What bots are, why they raise your usage and pollute your data, how to recognise them, how the AdPage Botfilter keeps them out of your platforms, and a plan for your IT or hosting party.

Not all collected data is good data. Bots visit your website every day. Some are useful, such as search engines and the crawlers behind your ads. Others spam your forms, pollute your analytics and raise your monthly usage, without a single extra sale.

This article covers the full picture: what bots are, why they increase your request count, how to recognise them, what you must not block, how the AdPage Botfilter keeps bots out of GA4, Meta and your other platforms, and a plan you can work through with your IT, developer or hosting party.

{% hint style="info" %}
**The key point:** a request to your server container is counted the moment the browser sends it. Filtering in the server container, with the AdPage Botfilter or your own rules, keeps your data clean, but it does **not** lower your usage. Usage only goes down when the bot never loads your tags or never lets them fire. That happens on your website, in your CDN or hosting, or in your GTM web container, before the tags run.
{% endhint %}

## What is bot traffic?

Bot traffic is every non-human visit to your website or landing page. In AdPage it often shows up as:

* **Rising request count:** usage goes up while orders and leads stay flat.
* **Spam submissions:** automatically completed forms.
* **Ghost referrals:** traffic in analytics that never actually visited your page.
* **Ad fraud:** bots that click your ads and waste your budget.
* **Crawlers:** search engines, ad platforms, AI tools and SEO tools reading your pages.
* **Malicious bots:** automated attacks that actively cause harm.

### Useful bots and harmful bots

| ✅ Useful bots (keep them)                                                                                | ⚠️ Harmful bots (data polluters)                                                                    |
| -------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------- |
| **Search engine crawlers** index your pages for Google and Bing.                                         | **Spambots** flood your forms and reviews with fake messages.                                       |
| **Ad platform crawlers** check landing pages, feeds and link previews for Google Ads, Shopping and Meta. | **Scraper bots** copy your content and prices, often for competitors.                               |
| **Performance bots** monitor speed and uptime.                                                           | **Credential-stuffing bots** try stolen logins and can slow down login pages.                       |
| **Link-preview bots** show your page nicely in social posts and chats.                                   | **DDoS bots** overload your website with meaningless traffic.                                       |
|                                                                                                          | **Ad-fraud bots** click your ads to waste budget.                                                   |
|                                                                                                          | **AI training crawlers and headless browsers** load full pages, run your tags and raise your usage. |

## Why bot traffic costs you

Bot traffic is not just a technical issue. It hits you in four places:

* **Higher usage and costs:** every bot that runs your tags adds requests to your monthly usage. See the next section.
* **Polluted algorithms:** Meta and Google optimise for conversions. When bots submit forms or trigger events, the algorithm learns from them and targets more bots instead of real buyers.
* **Damaged email reputation:** fake leads end up in Mailchimp or HubSpot, bounce rates rise and your newsletters may land in spam.
* **Distorted decisions:** your conversion rate looks lower or higher than it really is, so campaigns get judged on the wrong numbers.

### Why bots increase your request count

<figure><img src="https://2963651437-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F5eC7lwaeygB0xx6AOZcE%2Fuploads%2FLKi5g2NF0rHp84gYiBdO%2Fbot-traffic-venn.png?alt=media&amp;token=64711a85-7404-4320-a3d4-ea71b32de877" alt="What happens when a bot loads your page"><figcaption></figcaption></figure>

Every bot that renders a page in a real (headless) browser runs the same code as a visitor:

1. The bot opens a page on your website.
2. The page loads the AdPage container script - **and also your other tracking tools!**
3. The tags fire and send events (page\_view and others) to your tagging subdomain.
4. **AdPage counts the request on arrival** - **just as other tracking tools!**
5. Only then does the server container process the event and forward it to GA4, Meta, Google Ads and other platforms.

Your tagging subdomain points to AdPage, but the bot enters through your website. Anything that happens after step 4 cannot undo the count. That is why the fix for your usage sits before step 1 or 2, and the fix for your data sits in step 5.

| Where you take action                                                    | Lowers usage | Cleans your data |
| ------------------------------------------------------------------------ | ------------ | ---------------- |
| CDN or hosting: block or challenge the bot                               | Yes          | Yes              |
| Website: serve the page without the tracking script to crawlers          | Yes          | Yes              |
| GTM web container: bot exception on your tags                            | Yes          | Yes              |
| Forms: honeypot, reCAPTCHA, validation                                   | No           | Yes (leads)      |
| **AdPage Botfilter** in your server container (Data boosts › Bot filter) | **No**       | **Yes**          |
| Server-side GTM: your own filter on User-Agent, IP or country            | **No**       | Yes              |
| GA4 built-in bot filtering                                               | No           | Partly           |

### What AdPage can and cannot do for you

AdPage does not store raw access logs per request (IP address, User-Agent, country, status code) and cannot export them. Your hosting party or CDN does have this data. They are also the only party that can stop a bot before it loads your page, and so the only party that can lower your usage.

What AdPage does do: it shows you the pattern, and it keeps bots out of your platforms. **Data boosts › Bot filter › What we see** compares the last seven days with the seven before (hits per day, hits per cookie, purchases, the largest User-Agent, traffic at night) and says in one sentence whether it looks like bots. The **Botfilter** itself runs in your server container and stops bots before they reach GA4, Meta, Google Ads and your other platforms. See [The AdPage Botfilter](#id-3.-clean-the-data-that-reaches-your-platforms-the-adpage-botfilter) below.

## How to identify bot traffic

Start with **Data boosts › Bot filter › What we see** in your AdPage dashboard. It needs two weeks of traffic to compare with, and then gives the four numbers that matter and a verdict in one sentence:

<figure><img src="https://2963651437-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F5eC7lwaeygB0xx6AOZcE%2Fuploads%2Fuo4otJ0WIJ38cJV2OryE%2Fbotfilter-pattern.webp?alt=media" alt="Data boosts › Bot filter › What we see: hits per day, hits per cookie, purchases per day and the largest User-Agent, each against the week before"><figcaption><p>What we see: the last seven days against the seven before. Hits that triple while purchases stay flat, and far fewer hits per cookie, are the signature of bots that come in on a fresh cookie every time.</p></figcaption></figure>

Then check these over the last 7 days. One signal on its own proves little. Three or more together make a strong case.

| Check                                                 | Where                                         | Points to bots when                                                                                  |
| ----------------------------------------------------- | --------------------------------------------- | ---------------------------------------------------------------------------------------------------- |
| Hits against the week before                          | AdPage: Bot filter › What we see              | Hits per day double or triple while purchases per day stay flat                                      |
| Hits per cookie                                       | AdPage: Bot filter › What we see              | Far fewer hits per cookie than the week before: every bot visit brings a fresh cookie                |
| Traffic at night                                      | AdPage: Bot filter › What we see, access logs | Between 01:00 and 06:00 traffic stays close to daytime levels                                        |
| One User-Agent that dominates                         | AdPage: Bot filter › What we see              | One exact User-Agent, often an outdated browser version, is a large share of all events              |
| Share of events without an existing GA cookie (`_ga`) | AdPage tracking data                          | Well over half of all events, and even higher at night                                               |
| Page views per purchase                               | GA4 or your shop backend                      | Thousands of page views per order, where a few hundred is normal                                     |
| Spikes in traffic                                     | GA4, AdPage usage                             | Thousands of page views per hour, or a sudden jump compared with your normal trend                   |
| Top networks (ASN)                                    | Access logs, CDN dashboard                    | Large share from cloud providers (AWS, Microsoft, Google Cloud, Hetzner, OVH) or from Meta (AS32934) |
| Top User-Agents                                       | Access logs, CDN dashboard                    | `HeadlessChrome`, empty User-Agents, `python-requests`, AI crawlers, SEO tools                       |
| Unexpected countries                                  | GA4, CDN dashboard                            | Traffic from regions where you do not advertise or deliver                                           |
| Session behaviour                                     | GA4                                           | Sessions of zero seconds or many hours, only new users, Direct or Unassigned traffic                 |
| Fake conversions                                      | Forms, CRM                                    | Random characters, obviously fake email addresses, dozens of leads in minutes                        |

{% hint style="warning" %}
Signals you should **not** rely on:

* **The User-Agent as proof.** Any bot can claim to be Googlebot or Chrome. Only an official IP list or a reverse DNS lookup proves who it is.
* **"It comes from Meta or Google, so it is fine."** The network tells you who sent it, not whether you need it in your tracking.
* **robots.txt as a block.** It is a request that well-behaved bots follow. Some crawlers ignore it on purpose.
* **GA4's own bot filter.** It only removes known bots. Headless browsers with a normal User-Agent are counted as visitors.
  {% endhint %}

Unsure how a normal month should look? See [Monthly Usage](https://docs.adpage.io/accountmanagement/container-management/monthly-usage) for the rule of thumb based on your GA4 sessions.

## What you must not block

Blocking the wrong crawler can stop your ads, product feeds or link previews. Keep these accessible:

| Crawler                                                             | What it does                               | If you block it                                   |
| ------------------------------------------------------------------- | ------------------------------------------ | ------------------------------------------------- |
| Googlebot                                                           | Google Search                              | You disappear from search results                 |
| AdsBot-Google, AdsBot-Google-Mobile                                 | Checks Google Ads landing pages            | Ads can be disapproved, Quality Score drops       |
| Storebot-Google                                                     | Google Shopping and Merchant Center        | Shopping ads and free listings suffer             |
| facebookexternalhit, facebookcatalog                                | Link previews and product catalog for Meta | Bare links, problems with catalog and dynamic ads |
| Bingbot, AdIdxBot                                                   | Bing search and Microsoft Advertising      | Same as Google, for Microsoft                     |
| Pinterestbot, LinkedInBot, Twitterbot, Slackbot, WhatsApp, Applebot | Link previews                              | Links in posts and chats show no preview          |

These are generally safe to block, because they do not affect your ads or feeds: **GPTBot, ClaudeBot, CCBot, Bytespider, meta-externalagent** (AI training) and SEO crawlers such as **AhrefsBot, SemrushBot and MJ12bot** (unless you use those tools yourself). The AdPage Botfilter always lets the crawlers in the table above pass, and stops the AI training and SEO crawlers when you leave the hard rule *Crawlers without value* switched on.

{% hint style="warning" %}
**Do not block Meta's network (AS32934).** Meta reviews landing pages after you create or edit ads, sometimes in a headless browser with a normal Chrome User-Agent. That traffic runs your tags and can make up most of your events. Blocking or challenging it can lead to rejected ads. The right approach is: **let Meta see the page, but without the tracking script.** Real visitors from the Facebook and Instagram apps come in through their own internet provider, not through Meta's network, so you lose no real measurements.
{% endhint %}

## How to exclude bots

The right measure depends on what you want to fix. Measures in the first group lower your usage and clean your data. The other two groups clean your data, but do not lower usage.

### 1. Stop bots before your tags run (lowers usage)

* **Web application firewall (WAF) and bot protection:** your CDN or hosting recognises known bot patterns and blocks or challenges them before they reach your page. Examples are Cloudflare, or the bot protection built into many managed hosts.
* **Block bots without value:** AI training crawlers, SEO crawlers you do not use, empty User-Agents and obvious automation (`HeadlessChrome`, `python-requests`, `curl`).
* **IP rate limiting:** limit the number of requests one IP address can make in a short period. This stops aggressive scrapers and spambots.
* **JavaScript challenges:** for uncertain traffic, such as cloud networks, the browser has to solve a small check first. Most bots fail it. Never challenge the crawlers listed above.
* **Geographic filtering:** only if you do not deliver outside your own markets, and never for verified crawlers or Meta's network.
* **Page without tracking for crawlers:** let useful crawlers and Meta's network see the page, but leave out the GTM snippet for them.
* **Bot exception in your GTM web container:** stops tags from firing in automated browsers. It works on every platform, including Shopify, and catches what slips past the edge.

<details>

<summary>How to set up the bot exception in your GTM web container</summary>

**1. Variable:** type *Custom JavaScript*, name `JS - Is Bot`:

```javascript
function () {
  try {
    if (navigator.webdriver) return true;
    var ua = navigator.userAgent || '';
    return /bot|crawl|spider|slurp|facebookexternalhit|facebookcatalog|meta-external|HeadlessChrome|Lighthouse|PTST|python|curl/i.test(ua);
  } catch (e) { return false; }
}
```

**2. Trigger:** type *Custom Event*, name `EXC - Bot`, event name `.*` with *Use regex matching* on, firing on *Some Custom Events* where `JS - Is Bot` equals `true`.

**3. Exception:** add `EXC - Bot` as an exception to your Google tag, your GA4 event tags and every tag that sends to your server container.

**4. Test:** check in Preview mode that normal visits still fire. The Facebook and Instagram in-app browsers do not contain "bot" in their User-Agent and are not affected.

</details>

### 2. Keep forms and leads clean

* **Honeypot field:** an invisible form field that humans never see. Bots fill in every field, so a filled honeypot exposes them without bothering real users.
* **reCAPTCHA v3:** scores every visitor in the background between 0 (likely a bot) and 1.0 (likely human). Reject submissions below 0.5.
* **Validate email addresses and phone numbers:** enforce correct formats, for example Dutch phone numbers. Many bots enter random strings.
* **Device fingerprinting:** bots change IP addresses, but their device fingerprint often stays the same. Use it to block repeated attacks from one source.

### 3. Clean the data that reaches your platforms: the AdPage Botfilter

The Botfilter is a boost in your AdPage dashboard (**Data boosts › Bot filter**) plus a template in your server container. Every event that arrives in the server container is judged there, with rules that AdPage keeps up to date and that you tune on the page. A bot no longer reaches GA4, Meta CAPI, Google Ads or any other platform your server container sends to, so your algorithms learn from real customers. Remember: the request is already counted at this point, so this does not lower your usage.

{% hint style="info" %}
The Botfilter is being rolled out per account. If you do not see **Bot filter** under Data boosts yet, ask AdPage support to switch it on for you.
{% endhint %}

#### How it decides

Two kinds of rules work together:

* **Hard rules** stop a request on their own, purchases included: automated browsers (HeadlessChrome, Puppeteer, Playwright, python-requests, curl, an empty User-Agent), crawlers without value (AI training crawlers and SEO tools; search engines and ad platforms always pass), and your own lists of countries, IP addresses and User-Agents.
* **Signals** collect points: a browser six or more releases behind the current version (20), a cookie that is seconds old without a referrer or click id (20), a first visit that lands straight on a product page (20), more than twenty add-to-carts an hour from one cookie (40), a visitor outside the markets you sell to (20), one User-Agent that suddenly dominates your traffic (30) and night hours between 01:00 and 06:00 (10). A request is stopped once the points reach the threshold of your strictness: 80 points on *Mild*, 60 on *Normal*, 40 on *Strict*.

One signal is never a reason on its own. A first visit on a product page is simply what a visitor from an ad looks like; it only counts together with other signals. And **a purchase is never stopped on points**, only by a hard rule. Switching the filter on therefore costs no revenue; what a false positive costs is a page\_view or add\_to\_cart of a real visitor.

#### Monitor mode first

A new activation starts in **monitor mode**: the server container measures and flags, but stops nothing. Under *What it did* you see what would have been filtered. After an hour of monitor mode without a flagged purchase you can switch the filter **on**. If monitor mode flagged purchases, make the conditions milder first.

<figure><img src="https://2963651437-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F5eC7lwaeygB0xx6AOZcE%2Fuploads%2FYn5mDmOoRnDUKNtuGote%2Fbotfilter-mode.webp?alt=media" alt="Data boosts › Bot filter › Mode: Off, Monitor and On, with the note that a purchase is never stopped on points"><figcaption><p>Mode: Off, Monitor or On. On becomes available after an hour of monitor mode without a flagged purchase.</p></figcaption></figure>

#### Conditions you set yourself

Under **Conditions** you choose the strictness and switch the hard rules on or off. Then you fill in two lists of countries: countries to block (everything from there is stopped, purchases included) and **your markets** (the countries you sell to; a visitor outside them gets 20 extra points, and leaving it empty switches that signal off). Filling in your markets is the single most useful setting: it catches bots from countries you never deliver to at *Normal* strictness, without touching real visitors. Behind *Advanced* you can block IP addresses and User-Agents; only fill those in when AdPage or your IT partner asks you to, because a real visitor with that browser is then stopped too.

<figure><img src="https://2963651437-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F5eC7lwaeygB0xx6AOZcE%2Fuploads%2F7aXGLmannCgL6qng2WJ6%2Fbotfilter-rules-top.webp?alt=media" alt="Data boosts › Bot filter › Conditions: strictness and the hard rules"><figcaption><p>Conditions: the strictness and the three hard rules.</p></figcaption></figure>

<figure><img src="https://2963651437-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F5eC7lwaeygB0xx6AOZcE%2Fuploads%2FJsPtfsb8fVKhKyMqUo77%2Fbotfilter-rules-countries.webp?alt=media" alt="Data boosts › Bot filter › Conditions: countries to block, your markets and the advanced rules"><figcaption><p>Conditions: countries to block and your markets. Save once with the button at the bottom.</p></figcaption></figure>

#### What it did

**What it did** shows the last seven days from the reports of the template in your server container: how many requests were flagged (monitor mode) or stopped (on), whether any purchase was flagged, and the breakdown per event, per signal and per day. A report counts under every signal it carried, so the *Per signal* column does not add up; the ⓘ behind each signal tells you what it is and how many points it gives.

<figure><img src="https://2963651437-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F5eC7lwaeygB0xx6AOZcE%2Fuploads%2FZBmoBy5mXVz3KvtRc9bZ%2Fbotfilter-effect.webp?alt=media" alt="Data boosts › Bot filter › What it did: flagged, stopped and purchases flagged, with the breakdown per event, per signal and per day"><figcaption><p>What it did: in monitor mode this is what would have been filtered. Purchases flagged should stay at 0.</p></figcaption></figure>

#### Installing the template

The page ends with the installation: your container id, a rules token, the templates and five steps in your server container. The **AdPage Botfilter** variable gives the verdict per event and becomes an exception on every tag that sends to a platform; it also reports every bot to AdPage, which is what *What it did* shows. With a lot of bot traffic you also add the **client** template, which stops a bot before the container even parses the request. Both read the rules from AdPage every five minutes, so a change under Conditions is live within minutes without touching the container again.

<figure><img src="https://2963651437-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F5eC7lwaeygB0xx6AOZcE%2Fuploads%2FFXXYnxbLQhqf0zVDSJcW%2Fbotfilter-install.webp?alt=media" alt="Data boosts › Bot filter › Install in your server container: container id, rules token and the template"><figcaption><p>Install in your server container: the container id, the rules token and the template, followed by the five steps: import the template, create the variable and the exception trigger, optionally the client, test in Preview and publish.</p></figcaption></figure>

#### Doing it yourself instead

You can also build your own filter in the server container, on User-Agent, IP address or country. AdPage provides [GEO server headers](https://help.adpage.io/nl/articles/6552898) you can use for this. You then maintain the rules yourself and see nothing of what they stopped; the Botfilter does both for you.

## The step-by-step plan

Work through these steps in order. Steps 3 to 5 are done by whoever manages your website, CDN or hosting. The section [Talking to your IT or hosting party](#talking-to-your-it-or-hosting-party) tells you what to ask.

{% stepper %}
{% step %}

### Take a baseline (7 days)

Before you change anything, write down:

* The average number of requests per day in your AdPage dashboard.
* Requests per hour at night (01:00–06:00) compared to daytime.
* Page views per purchase and the number of orders.
* The numbers under Data boosts › Bot filter › What we see.
* The top 20 IP addresses, networks (ASN) and User-Agents. Ask your hosting party or check your CDN.
* Which ad channels are active: Google Ads, Shopping, Meta, Pinterest, Microsoft.

Without a baseline you cannot prove later that a measure worked, or that it did not cost you real customers.
{% endstep %}

{% step %}

### Brief your IT or hosting party

Share this article. Agree on who changes what, and on the date the changes go live, so you can compare before and after.
{% endstep %}

{% step %}

### Block bots without value at the edge

In your CDN or hosting, block AI training crawlers, SEO crawlers you do not use, empty User-Agents and obvious automation (`HeadlessChrome`, `python-requests`, `curl`). Add a rate limit per IP address. Challenge (do not block) traffic from cloud networks that is not a verified crawler.
{% endstep %}

{% step %}

### Let useful crawlers in, without tracking

Allow the crawlers from [What you must not block](#what-you-must-not-block). For Meta's network (AS32934) and for crawlers in general, have the website serve the page without the GTM snippet. This is the step with the biggest effect when Meta traffic dominates.
{% endstep %}

{% step %}

### Add a bot exception in your GTM web container

It stops tags from firing when the browser is automated or calls itself a bot. See [How to exclude bots](#id-1.-stop-bots-before-your-tags-run-lowers-usage) for the setup.
{% endstep %}

{% step %}

### Switch on the AdPage Botfilter

While your IT or hosting party works on the edge, keep your platforms clean. Install the template in your server container, fill in your markets under Conditions, run monitor mode for an hour and switch the filter on. See [The AdPage Botfilter](#id-3.-clean-the-data-that-reaches-your-platforms-the-adpage-botfilter).
{% endstep %}

{% step %}

### Protect your forms

If you also see spam leads, add a honeypot field, reCAPTCHA v3 and field validation to your forms.
{% endstep %}

{% step %}

### Allow AdPage's verification check

A stricter firewall can also turn away AdPage's Tagging Pixel check. Add the IP addresses shown in your container dashboard to the allowlist. See [Tagging Pixel verification](https://docs.adpage.io/documentation/getting-started/tagging-pixel-verification).
{% endstep %}

{% step %}

### Measure after 7 days and check your ads

Compare the same metrics as your baseline (see [Measure the effect](#measure-the-effect)). Then check Google Ads, Merchant Center and Meta Ads Manager for new disapprovals, and test a product link in the Meta Sharing Debugger.
{% endstep %}
{% endstepper %}

## Talking to your IT or hosting party

You do not need to know how to write firewall rules. You do need to know what to ask and which decisions are yours.

### Questions to ask

1. Which CDN, WAF or bot protection runs in front of our website, and on which plan?
2. Which tools do we use to manage bot traffic over time?
3. Can you share the top 20 IP addresses, networks (ASN), User-Agents and countries from the last 7 days, and the number of requests per hour?
4. Can you block AI training crawlers and SEO crawlers we do not use, and add a rate limit per IP address?
5. Can verified crawlers (Google, Bing, Meta, Pinterest) stay allowed, while our website leaves out the tracking script for them and for Meta's network (AS32934)?
6. Can you add AdPage's verification IP addresses to the allowlist?
7. When do the changes go live, and how can we roll them back if ads or feeds are affected?

### Decisions you make together

| Decision                       | Recommendation                                                                                                                             |
| ------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------ |
| Block AI crawlers?             | Block training crawlers, for example with your CDN's AI crawler settings. Allow search and agent crawlers if AI visibility matters to you. |
| Restrict countries?            | Only if you do not deliver outside your own markets, and never for verified crawlers or Meta's network.                                    |
| Block or challenge?            | Block bots with no value. Challenge uncertain traffic, such as cloud networks.                                                             |
| Meta traffic                   | Never block. Serve the page without the tracking script.                                                                                   |
| Who checks the ads afterwards? | Your marketing team or agency, within 48 hours of each change.                                                                             |

## Measure the effect

Compare 7 days after the change with your 7-day baseline, on the same weekdays and without large campaign changes in between.

| Metric                                 | What you want to see                                                                     |
| -------------------------------------- | ---------------------------------------------------------------------------------------- |
| Requests per day (AdPage dashboard)    | Back towards the level before the increase                                               |
| Requests per hour at night             | A clear dip at night, following your daytime pattern                                     |
| Bot filter › What we see               | Hits per cookie back to the level of the week before, no User-Agent that dominates       |
| Bot filter › What it did               | Stopped falls once the edge measures are live; purchases flagged stays at 0              |
| Share of events without a `_ga` cookie | A clear drop, especially at night                                                        |
| Page views per purchase                | Back to a few hundred or less                                                            |
| Spam leads                             | Fewer or none                                                                            |
| Orders and revenue                     | **Unchanged.** If orders drop, a rule is too strict and is hitting real customers        |
| Ad and feed status                     | No new disapprovals in Google Ads, Merchant Center or Meta, and link previews still work |

Once usage is back to normal, keep an eye on it. A sudden jump of more than 50% above your monthly average is a reason to repeat the checks in [How to identify bot traffic](#how-to-identify-bot-traffic).

## When to contact AdPage support

Contact support when you have taken the steps above and usage stays high, when you want help reading the pattern in your tracking data, or when you do not see the Bot filter under Data boosts yet. Include your baseline, the date the changes went live and the measures that were taken. That lets us compare quickly instead of starting from scratch.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.adpage.io/faq/debugging-monitoring/identifying-and-excluding-bot-traffic.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
