# How to Filter Low-Intent Website Visitor Alerts

A playbook for filtering low-intent website visitor alerts with page context, account fit, repeat visits and suppression, set up in Snitcher, Leadinfo, Warmly, HubSpot and Make.

Visitor alerts turn noisy when every identified company pages a rep. The fix is a short set of filters that decide which visits become an alert and which wait in a digest.

- Filter low-intent visitor alerts by passing only visits where page context, account fit and repeat behaviour agree.
- Let pricing, demo and comparison visits alert, and send blog, careers and docs visits to a digest.
- Filter by industry, employee range and country before any page rule runs.
- Require a second session or minimum time on a high-intent page before alerting.
- Suppress customers, open deals, competitors and your own team at the source.

This applies to teams that already identify visiting companies, send alerts to Slack, email or a CRM, and can edit segments, triggers or workflows in their tool. Filters only sort what the tool has matched. They cannot make a company-level match name a person, and the thresholds need a few weeks of tuning against your own traffic.

## What Makes A Website Visitor Alert Low Intent?

A low-intent visitor alert is one that names a company with no fit to your market, no visit to a buying page or no return visit, and with [bots now accounting for over 53% of all web traffic](https://community-cybersec.thalesgroup.com/blogs/tim-chang/2026/05/05/the-2026-bad-bot-report-is-here), a single pageview is weak evidence on its own. Thales reported that figure in its 2026 Bad Bot Report.

A company you already serve, or one that competes with you, is low intent for sales too, however many pages it reads. Each of these alerts costs a rep the time to open it, judge it and dismiss it. After enough of them, reps stop opening the channel, and the good alerts get ignored with the bad.

The alerts that matter are few, and they arrive early. In 6sense's 2025 Buyer Experience Report, based on more than 4,000 buyers, [the balance between independent research and seller engagement shifted from a 70/30 split to 60/40](https://www.businesswire.com/news/home/20251112018032/en). Buyers still do most of the deciding before they talk to anyone.

The same 6sense survey found that the vendor buyers prefer before they engage sellers wins 80% of deals, and that 94% of buying groups ranked their preferred vendors before first contact. An alert that reaches a rep during that research window is worth far more than ten alerts about companies that will never buy.

Filtering sits inside [website visitor identification](https://growthtribune.com/website-conversion/visitor-identification/) as the step between a match and a sales action. The wider argument for why [a website visit does not always mean buying intent](https://growthtribune.com/why-a-website-visit-does-not-always-mean-buying-intent/) has its own guide. This playbook covers the rules and the settings.

> **Low-intent alert:** A visitor alert that fails at least one test of fit, page context or repeat behaviour, or that names a customer, competitor, partner or your own company.

## The Filters That Decide Which Visits Become Alerts

A visitor alert should reach a rep only when the company fits your market, the visit touched a buying page such as pricing or a demo request, and the account has shown repeat or engaged behaviour, while everything else belongs in a daily digest. Suppression then removes the companies that should never alert at all.

Snitcher's help centre gives the same advice in its own words. It recommends combining good-fit leads with behaviour such as visiting pricing pages, rather than alerting on either alone.

**The alert filter model**

| Filter | Question it answers | Evidence it reads | Example rule |
| --- | --- | --- | --- |
| Page context | Which pages? | URL path, time on page | Path starts with /pricing |
| Account fit | Which companies? | Industry, size, country | 51-200 employees, software |
| Repeat behaviour | How much, how often? | Sessions, pageviews, time on site | Two or more sessions |
| Suppression | Who never alerts? | CRM stage, tags, domain lists | Lifecycle stage is customer |

The rule that ties them together is simple. An alert fires only when fit and page context both pass and behaviour clears the threshold. A visit that fails one test goes to a digest, not to the bin, because a weak signal today can be the first visit of a real evaluation next week.

Order matters too. Run suppression and fit first, since they cut the most volume with the least risk. Page context and repeat behaviour then sort what is left into alert and digest.

## How To Use Page Context To Filter Visitor Alerts

Page context filters an alert by the URL path a visitor reached, and Snitcher, Warmly, Leadinfo and HubSpot all let a rule match on the page URL, such as a path containing /pricing, so a team can alert on buying pages and route blog, careers and support visits to a digest.

Start by sorting your site into three tiers:

- **High intent:** Pricing, demo request, comparison, integrations and security pages.
- **Middle intent:** Product and solution pages.
- **Low intent:** Blog, careers, docs, support and login pages.

The tools express page rules in three ways:

- **Path operators.** Snitcher's `page` and `url` fields accept contains, not\_contains, starts\_with and doesnt\_start\_with. HubSpot's buyer intent criteria use equal to, contains and starts with.
- **Page logic.** Warmly segments filter pages in or out with any-of all pages (the visitor must see every page), any-of at least one page, and none-of all pages.
- **Time on a page.** Leadinfo's Visit time on page URL rule can require, for example, at least 30 seconds on a URL containing "pricing".

The decision rule is to tier pages, not delete them. A rule that excludes every blog session also drops a target account that read a comparison post on the blog. Let low-tier pages lower an alert to the digest, and keep hard exclusions for pages that never signal buying.

Time on page separates a reader from a bounce. A visit that spends half a minute on pricing is a reader. A visit that bounces off it in a second may be a crawler or a misclick.

> **Tip:** Put careers and login paths in a hard exclusion. HubSpot's own buyer intent settings use careers as the example of a page to exclude. Job seekers and logged-in customers should not page sales, so keep those paths out of every alert rule.

## How To Filter Visitor Alerts By Account Fit

Account fit filters an alert by who the company is rather than what it read, and in Snitcher that means industry, company size range and country conditions on a segment, while HubSpot calls the same idea a target market built from industry, location and employee count.

Turn your ICP into four alert-side fields:

- **Industry.** The sectors you sell to, matched against the tool's reference values.
- **Employee range.** The company sizes where you win deals.
- **Company country.** The markets you can sell and serve in.
- **Visitor country.** Useful when a global company's local office is the buyer.

Each tool reads these fields its own way. Snitcher filters `company.size` by predefined ranges such as 51-200 employees, not by a raw number, and checks `company.industry` and `company.country` against reference values. Leadinfo's Branche rule matches the industry listed on the company's LinkedIn profile. HubSpot target markets combine industry, location and employee count.

The decision rule is that fit runs before page rules. A pricing visit from a company outside your market is still noise, so fit should cut it before anything else looks at it.

The risk runs the other way too. A fit filter that is too narrow stops alerts without any error. Snitcher's Slack troubleshooting list names an overly restrictive segment filter as one of the reasons alerts never arrive. Start wider than your ideal profile, watch what lands in the digest for two weeks, then tighten.

Alert filters can only be as precise as the profile behind them and the firmographic fields a tool returns. The comparison of the [best company-level visitor identification tools](https://growthtribune.com/best-company-level-visitor-identification-tools/) sets out how much firmographic depth each one offers.

## How Many Repeat Visits Should Trigger An Alert?

A repeat-visit threshold stops one-off visits from paging a rep, and HubSpot lets each intent criterion require a minimum number of visits and unique visitors within a window of up to 90 days, while Snitcher segments can require more than a set number of sessions or seconds on site. For net-new accounts, a second session is a sensible bar.

The right threshold depends on what you already know about the account. A net-new company needs more evidence than a named account or a deal you lost last year. For net-new fit accounts, alert on a second session within the window, or on one engaged session on a high-intent page. For named or closed-lost accounts, alert on the first high-intent visit.

**Example thresholds by account type**

| Account type | Alert when | Otherwise |
| --- | --- | --- |
| Net-new, good fit | Second session or long pricing visit | Daily digest |
| Named target account | First high-intent visit | Daily digest |
| Closed-lost account | First high-intent visit | Weekly digest |
| Net-new, weak fit | Never in real time | Weekly digest |

The levers each tool offers:

- **Snitcher.** Numeric `sessions`, `pageviews` and `time_on_site` fields (in seconds), with equal, greater\_than, less\_than and between.
- **HubSpot.** Minimum visits and minimum unique visitors per intent criterion, over a time period of up to the last 90 days.
- **Leadinfo.** Leadinfo's trigger rules do not include a visit count. Its repeat filter is the Visit time on page URL rule plus the Slack setting for how often a message is sent per unique company.

An engaged session is the practical middle ground. Define it as one visit that clears a time-on-site or pageview bar, and treat it as equal to a second session. That keeps a serious first-time evaluator from waiting in the digest.

## How To Suppress Customers, Competitors And Internal Traffic

Suppression removes companies that should never page a salesperson, and Snitcher can do it inside the alert segment by excluding HubSpot lifecycle stages such as customer, while HubSpot's own buyer intent tool can exclude up to 100 domains. Every alert stack needs a never-alert list before it needs a threshold.

Build the list from five groups:

- **Current customers.** Send their visits to customer success, not sales.
- **Open opportunities.** Send them to the deal owner. The ownership rules to [route identified accounts to sales](https://growthtribune.com/how-to-route-identified-accounts-to-sales/) cover that step.
- **Competitors.** Their visits, pricing pages included, should never page a rep.
- **Your own company and agencies.** Internal staff and contractors inflate every page count.
- **Job seekers.** Careers and about pages are their trail.

Each tool handles suppression differently:

- **Snitcher.** Segments can filter on `hubspot.company_life_cycle_stage`, `hubspot.deal_stage`, `hubspot.company_owner`, `hubspot.company_type` and tags. Without a native CRM integration, Snitcher's help centre recommends Lists for lost deals and active opportunities.
- **HubSpot.** Buyer intent accepts up to 100 excluded domains and filters by CRM status, including lifecycle stage, deal stage and owner.
- **Warmly.** Sessions from blocklisted email domains and from segment exclusions are filtered out and never alert.

> **Warning:** Suppress companies, not pages. Excluding the pricing page to stop customers alerting also hides every prospect who reads it. Put customers, competitors and your own domains on a company-level list, and leave page rules to sort intent.

## How To Set Up Alert Filters In Snitcher

Snitcher filters Slack alerts with segments, so a team builds one segment that combines company size, industry, a pricing-page condition and a minimum session count, then attaches it to an automation whose action is Send to Slack. The segment does the filtering, and the automation only delivers.

- **Where alerts are built:** Automations. Each one pairs a trigger, such as New company identified or Company visit, with a segment filter and the Send to Slack action. Snitcher's own segment examples are ICP Match, Named Accounts and 500+ employees.
- **Page-context filter:** The `page` and `url` fields with contains, not\_contains, starts\_with and doesnt\_start\_with, so `url` starts\_with /pricing marks a buying page.
- **Account-fit filter:** `company.industry`, `company.size` in predefined ranges, `company.country` and `visitor.country`.
- **Repeat-visit filter:** `sessions`, `pageviews` and `time_on_site`, with greater\_than or between for a minimum.
- **Suppression:** HubSpot lifecycle stage, deal stage, company owner and company type fields, plus tags, or Lists when no CRM is connected.
- **Permissions and setup:** Settings > Integrations > Slack > Connect Slack, then choose the workspace and a default channel. A private channel needs `/invite @Snitcher` before messages arrive.
- **Limitation:** Snitcher's filter fields do not include a lead score, so scoring logic has to be rebuilt as segment conditions. Data passed on through the Zapier trigger Organisation Enters Segment can take a few minutes to sync.

**Segment-filtered Slack automations** (documented): Each Snitcher Slack message carries the company, industry, employee count, location, pages viewed and session duration, so a rep can judge the visit without opening the app. The help centre suggests separate channels for new high-intent leads, per-rep leads, lost or old deals and customers, and warns against sending too many notifications.

## How To Set Up Alert Filters In Leadinfo

Leadinfo filters alerts through Triggers, where a rule set such as a pricing URL visited for at least 30 seconds by a company in a chosen branch fires an action such as a Slack message sent once per unique company. Every condition lives inside the trigger itself.

- **Where alerts are built:** Triggers. Name the trigger, select the website, set the rules, choose at least one action, then enable and save. Actions include Tag, Email, CRM, Autopilot and LinkedIn Audience, plus Slack once it is connected.
- **Page-context filter:** Rules on the URL of the current session and the referrer, and the Visit time on page URL rule.
- **Account-fit filter:** The Branche rule, which reads the industry on the company's LinkedIn profile, or a SIC code. Rules combine with "all of the following" or "any of the following".
- **Repeat-visit filter:** The Slack action's setting for how often a message is sent per unique company, paired with time on a URL.
- **Suppression:** Use the Tag action to label customers, partners and competitors as they appear, and keep those companies out of the rules of the trigger that posts to the sales channel.
- **Permissions and setup:** Settings > Integrations > Slack, then grant Leadinfo access to the workspace. In the trigger, pick Slack as the action, the channel or users, and the message language.
- **Limitation:** The number of triggers depends on the package, from three on the entry package to ten on the middle one and no cap on the top one. A tiered alert design uses several triggers, so the cap shapes the design.

**Slack action with per-company frequency** (documented): Leadinfo's Slack action controls how often a message is sent for each unique company, which is the closest Leadinfo setting to a repeat-visit rule. The Columns setting applies to every Slack message, so a high-intent alert and a digest-style alert show the same fields.

## How To Set Up Alert Filters In Warmly

Warmly filters each Slack alert at three levels, the trigger type, real-time conditions on page path, UTM or location, and optional segments, and the person creating it needs Slack admin rights or app-install permission. Each alert also posts to one public channel, so the channel choice is part of the filter.

- **Where alerts are built:** Add Alert in the Slack settings. The trigger type is page visit or form entry, which fire in real time, or full identification, which fires once Warmly has found both the company and the person. The audience is identified company or all visitors.
- **Page-context filter:** Conditions on domain, page path, UTM and visit location, joined with all or any. Segments 2.0 page-visit signals filter pages in or out with any-of and none-of logic, and Warmly's help suggests filtering in pricing or demo pages for Slack notifications.
- **Account-fit filter:** Fit lives in segments rather than in the alert's own conditions. Attach an ICP segment and set the audience to identified company, so unmatched visitors never alert.
- **Suppression:** Sessions from blocklisted email domains and from segment exclusions are dropped automatically. Add your own domain and your agencies to the blocklist.
- **Permissions and setup:** The installer must be a Slack admin or able to install apps, or an Unauthorized error appears. Each alert can @mention no one, the lead-routing owners or named users.
- **Limitation:** One public channel per alert, and full-identification alerts can take from minutes to an hour, so treat them as near-real-time at best.

Warmly announced on 30 June 2026 that it has agreed to be acquired by HubSpot, and says existing contracts, account teams, product and integrations stay unchanged for now.

**Slack alert triggers, conditions and segments** (documented): Warmly's optional segments, such as SDR-owned accounts or closed-lost deals, let one alert serve a narrow purpose. A closed-lost segment on a page-visit trigger for pricing gives a re-engagement alert that fires on the first serious visit.

## How To Set Up Alert Filters In HubSpot

HubSpot filters website visitors with buyer intent, a feature it labels beta, where each of up to 10 intent criteria combines page paths, a minimum visit count and a window of up to 90 days, and a Professional or Enterprise workflow then sends the Slack notification. Buyer intent decides who qualifies, and the workflow delivers.

- **Where alerts are built:** A workflow with the Send a Slack notification action, which can notify users, channels or record owners. Companies from buyer intent can be enrolled in that workflow.
- **Page-context filter:** Page path operators equal to, contains and starts with, plus page exclusions such as careers.
- **Account-fit filter:** Target markets by industry, location and employee count, and visitor country on each criterion.
- **Repeat-visit filter:** Minimum visits and minimum unique visitors per criterion, over a period of up to the last 90 days.
- **Suppression:** Up to 100 excluded domains, and filters on CRM status such as lifecycle stage, deal stage and owner.
- **Permissions and setup:** Configuring buyer intent needs the Buyer Intent permission and the HubSpot tracking code, and may need a cookie consent banner. A Super Admin assigns the Buyer Intent and Data Enrichment permissions. Private Slack channels appear only after the HubSpot app is added, and only users who mapped their email addresses can be notified.
- **Limitation:** Companies added from buyer intent arrive with an empty owner field, and custom intent properties are not updated by saved views. Subdomains roll up to the root domain.

The running cost sits in HubSpot Credits. Adding and tracking new companies uses credits, and so do backfills and Recent visits digests, so set a monthly credit limit before widening any criterion. Because HubSpot labels buyer intent beta, and Warmly is joining HubSpot, re-check a filter stack built on either after each product change.

**Buyer intent visitor criteria** (documented): HubSpot buyer intent combines up to 10 visitor intent criteria with target markets, so one configuration can hold page, fit and repeat-visit rules together. It is labelled beta in HubSpot's documentation, which means fields and limits can change.

## How To Add A Filter Layer With Make

Make filters visitor alerts between modules, so a scenario can receive each identified visit, drop bundles that fail fit or page conditions, and send the rest through a router whose fallback route catches every visit no other route matched. It suits rules the source tool cannot hold.

Typical cases are a check against a spreadsheet of competitors, or a combined rule that a tool's own trigger cannot express. The scenario receives the visit from a webhook or an app trigger, then filters and routes it.

How the pieces work:

- **Filter.** A filter sits between two modules. Bundles that meet its conditions pass, and the rest stop there. Operators include exists, not exists, and text, numeric, date and array comparisons.
- **Router.** A router splits the scenario into routes, each with its own filter. Routes run one after another, not in parallel, and the fallback route runs last for data that matched no other route.
- **Credits.** Each action a scenario performs uses credits, most use one, and reading data from an app or webhook counts. The Router module and error handlers use none. When credits run out, scenarios stop and webhooks queue up to the queue limit. Extra credits come in bundles of 1,000 or 10,000.

The decision rule is to filter in the source tool first and use Make only for what the tool cannot do. Every bundle that passes a Make filter spends credits in the modules after it, so a loose source filter becomes a monthly bill.

- [ ] Webhook or app trigger that receives each identified visit
- [ ] Filter that drops suppressed domains and poor-fit companies
- [ ] Router with a high-intent route to the sales channel
- [ ] Digest route that collects middle-tier visits
- [ ] Fallback route that logs everything else
- [ ] Credit alert or limit check before go-live

## How To Test A Visitor Alert Before Reps See It

Test a visitor alert before reps see it by loading a high-intent page from an incognito window and confirming the session appears in the tool and the message lands in a private test channel, which is the routine Warmly documents with its w\_email URL parameter.

Warmly's test adds `w_email` with a business email to the page URL, then checks the Warm Calls view (on paid plans) for the session before checking Slack. Snitcher's troubleshooting list covers the cases where nothing arrives.

- [ ] Point the new alert at a private test channel first
- [ ] Load a high-intent page from an incognito window
- [ ] Confirm the session appears in the tool's session view
- [ ] Confirm the message lands in the test channel
- [ ] If nothing arrives, check the integration shows Connected
- [ ] Check the automation is active and has the Slack action
- [ ] Check the segment filter is not too restrictive
- [ ] Check the bot has been invited to the channel
- [ ] Move the alert to the sales channel only after a clean pass

## How To Measure Whether Your Alert Filters Work

Alert filters work when reps act on most of what reaches them and few converting accounts were sitting in the digest, so measure the share of alerts acted on and the share of later opportunities that never alerted, not the number of alerts sent.

Alert count is the wrong target. Most of the decision happens before a buyer contacts a seller, and the preferred vendor usually wins, as the 6sense figures above show. The measure that matters is whether alerted accounts were contacted early in that window.

**Alert filter success measures**

| Measure | What it shows | Watch for |
| --- | --- | --- |
| Alerts per rep per day | Load on each rep | Rising with no more meetings |
| Share of alerts acted on | Whether reps trust alerts | A steady fall |
| Alerts to meetings or opportunities | Whether filters pick buyers | Flat while volume grows |
| Digest accounts that later converted | Missed signal | Any repeat pattern |
| False-positive reasons | What to tune next | One reason dominating |

Log a reason for every dismissed alert, such as wrong fit, customer, competitor or job seeker. Review the reasons weekly for the first month, tighten the filter that causes the most noise, and check the digest for accounts that should have alerted. After the first month, a monthly review is usually enough.

## Common Mistakes When Filtering Visitor Alerts

The most common filtering mistake is sending every identified visit to one shared Slack channel, because reps stop reading it, and the second is a fit filter so narrow that alerts stop without warning.

- **One channel for everything.** Split high-intent alerts, digests, customers and lost deals into separate channels.
- **Deleting page groups.** Excluding the whole blog drops target accounts that read comparison posts. Tier pages instead.
- **A fit filter that is too narrow.** Snitcher lists an overly restrictive segment as a reason alerts never arrive.
- **Forgetting the meters.** Leadinfo caps triggers by package, HubSpot buyer intent spends credits on new companies, and every Make action spends credits.
- **Treating delayed identification as real time.** Warmly full-identification alerts can take from minutes to an hour.

> **Warning:** A filter that silently blocks alerts looks exactly like a quiet week. Check the digest and the tool's session view whenever alert volume drops sharply.

## Frequently Asked Questions

Send no more alerts a day than reps can act on the same day, and move everything below that bar into a daily or weekly digest.

**How many visitor alerts per day is too many?**

Too many is any number reps cannot act on the same day. There is no universal figure, because it depends on rep count, deal size and how much research each alert needs. If alerts sit unopened by the end of the day, raise the fit or repeat-visit threshold and move the rest to a digest.

**Should a company-only match ever trigger a real-time alert?**

Only when fit and page context are both strong, such as a named target account on the pricing page. Otherwise send company-only matches to a digest. The difference between [company-level and person-level identification](https://growthtribune.com/company-level-vs-person-level-visitor-identification/) decides how much a rep can do with the alert, since a company match names no person.

**Do alert filters remove bot traffic?**

No. Alert filters reduce noise, but they are not bot detection. Fit and engagement rules narrow which identified visits reach a rep, and that is all they do. Treat filters as noise reduction, keep a time-on-page or repeat-visit threshold in every alert, and watch for sudden spikes in alerts from one company or country.

**Does HubSpot have built-in visitor intent filtering?**

Yes. HubSpot buyer intent, which HubSpot labels beta, filters website visitors with target markets built from industry, location and employee count, plus up to 10 visitor intent criteria. Each criterion sets page paths, minimum visits and a window of up to 90 days. Adding and tracking companies uses HubSpot Credits.

## Sources

- [The 2026 Bad Bot Report is Here!](https://community-cybersec.thalesgroup.com/blogs/tim-chang/2026/05/05/the-2026-bad-bot-report-is-here), community-cybersec.thalesgroup.com
- [The Timeline for Influencing B2B Buyers Is Shrinking: Insights From 6sense's 2025 Buyer Experience Report](https://www.businesswire.com/news/home/20251112018032/en), businesswire.com

---
Source: https://growthtribune.com/how-to-filter-low-intent-website-visitor-alerts/
