Use website activity to inform your account research.

Begin with the event that was recorded: a page view, an interaction, a form submission or an account-level pattern. Retain the period, collection method and relevant limits. “Buyer intent” may be the name of a product or rule rather than a statement the visitor made.

HubSpot’s buyer-intent documentation describes using website activity, IP addresses and other identifiers to associate visits with companies. That explains one provider’s mechanism. It does not turn a company association into proof that a named contact viewed a particular page.

Someone associated with an account can visit a product guide for many reasons: a current customer checking a detail, a colleague supporting a project, a supplier researching the business or someone comparing terminology. The event can be real while its commercial meaning remains unknown. The buying-signals guide helps keep an observed change distinct from fit and action.

Ask what entity the provider matches, what evidence supports it and what confidence means in that system. The company associated with a network connection, the owner of a website domain and the operating site in your CRM may be different entities. Preserve an unresolved group or site distinction rather than selecting the most convenient account record.

A known CRM contact at the matched company is not automatically the visitor. Even when a person has identified themselves through a form, check which events are legitimately and accurately connected to that identity. Do not turn account-level activity into a claim about that person’s private browsing.

Record what remains unavailable. A provider may not match every visit, and collection can be affected by consent choices, settings and technical conditions. “No recorded activity” means no activity in the available observation set; it does not prove that a company has never researched the subject.

Describe which data is collected, which pages it covers, who receives it, how long it is kept and which downstream uses are intended. Include the provider’s sharing settings and the website’s consent behaviour. Installing a tag is a technical action, not a complete decision about whether a particular collection and use is appropriate.

HubSpot’s tracking-code documentation describes data such as IP addresses, timestamps, page paths and visitor identifiers, with behaviour affected by settings and consent. Review the current provider documentation and test your actual configuration, including an external consent tool if you use one.

Keep the applicable legal requirements and platform conditions explicit for each market and purpose. The ICO’s UK advertising guidance addresses advertising-related storage and access technologies. Its scope should not be presented as universal permission or a universal rule for every analytics design.

Inspect URLs and payloads for information that should not reach an observation provider, such as private enquiry text or account-specific tokens. Limit access to the data actually needed for the agreed task. A claim that a tool identifies companies does not mean that its processing contains no personal data.

Include only account information that matters to the review: fit conditions, existing relationships, current opportunities and a relevant owner. The account research brief helps a person see the source, interpretation and next question together.

Keep the proposed action narrow. An observation might help an account owner prepare for an already scheduled conversation or choose which public answer to review. It should not automatically trigger a message implying that the business knows a particular person’s browsing history.

If a person submits an enquiry, follow the inbound qualification process and preserve their actual request. If retargeting is being considered, use a separate paid campaign brief to settle audience eligibility, collection, suppression and activation. An internal account-review rule is not an advertising permission.

Sample the records that reached the team. Check entity matches, event timing, existing-customer treatment and whether the observation added useful information. Record unresolved matches and irrelevant observations rather than hiding them in an engagement total.

Ask the receiving person what decision changed, if any. “Added to the account review” is an operational outcome. “Customer confirmed a requirement” is evidence from a later interaction. Keep the source and date of that later evidence separate so the record does not imply the requirement was known from a page view.

A sale following an observed visit does not establish that the observation or a subsequent campaign caused it. The measurement guide distinguishes recorded history, attribution and evidence of additional outcomes. Small samples are often useful for improving the process while remaining insufficient for strong performance claims.

Walk through three cases: a company-associated visit, an unresolved match and a direct enquiry. Ask the reviewer to state what is known about identity, purpose and permission in each. The interface and CRM note should preserve those distinctions.

Then test withheld consent where it applies, an excluded page, a group with several sites and a known customer with open work. Verify what is collected, what reaches the reviewer and what stops. If the team cannot explain the observation without a provider label, improve the record before expanding use.

A defined research assignment can assess whether these observations answer a useful account question and hand the findings to your team. Within a Build campaign, agreed signals can help us choose research priorities and timing, with contact decisions based on the wider evidence. A standalone observation workflow has its own scope and quote. Agree collection, use and responsibilities around the systems you already have; an intent platform is useful only if it contributes to the task.