Buyer intent data is too old for an action when its timing, identity, or current relevance no longer supports that action. Check when the underlying event happened, what it actually meant, and what has changed since; the date a record arrived in your system does not make the behavior recent.

In his guidance for this article, Mike Mangione, founder of The Mangione Group, recommends:

If the identity, meaning, or current relevance is weak, we should reduce our confidence accordingly.

Mike Mangione, founder of The Mangione Group, recommends checking the origin, age, identity, and current meaning of an intent signal before acting. His submitted guidance for this article distinguishes an old signal that warrants fresh research from evidence that someone is ready to buy today.

Key Facts at a Glance

  • Event time and import time answer different questions and should be recorded separately.
  • An old research event can remain useful background without supporting immediate prospecting.
  • Company-level activity does not establish a particular employee's current buying intention.
  • HubSpot documents event-based score decay as one way to reduce older activity's contribution. Overview of the lead scoring tool
  • Its intent-signal feature can backfill earlier events when tracking begins. Use intent signals
  • Source history helps a reviewer assess the reliability of data and its interpretation. W3C provenance overview
  • A fresh request to stop takes priority over an old positive signal when deciding whether to continue contact.

What does stale buyer intent actually mean?

Stale buyer intent is evidence whose age or changed context makes it inadequate for the decision you want to make now. Staleness is relative to the action, not simply a fixed number of days. A signal can remain useful for understanding an account while being too weak to justify treating a named person as an active buyer.

Consider two hypothetical situations. A facilities manager requests help for a specific near-term disruption, while a purchasing team explores options for a replacement planned far ahead. An unanswered request in the first situation can lose usefulness quickly because the problem may already have been resolved. The second situation can remain relevant for research much longer. Neither example establishes a universal expiration rule. Each tells you to connect the signal to the decision and the likely life of the underlying need.

Separate three questions: did the event happen, does it still describe the situation, and does it justify the proposed action? A genuine page visit can answer the first question while leaving the other two open. A reliable source does not prevent circumstances from changing after collection. Likewise, a recent timestamp does not repair a vague event definition. The word “fresh” should therefore describe a checked relationship between evidence and use, rather than act as a general badge of quality on a purchased file.

The buyer intent data guide covers the broader task of understanding signal sources and limitations. This article focuses on what happens after a signal exists: preserving its age, deciding when to revalidate it, and preventing old context from becoming a new sales claim. That narrower process matters whenever records wait in queues, move between systems, or are reused for another campaign.

Which dates must you separate before judging freshness?

Separate the event date, collection date, provider update date, import date, and your own verification date. These timestamps describe different stages in the data's journey. Use the event date to assess behavioral age, the verification date to assess what your team checked, and the remaining dates to identify delays or transformations that could affect interpretation.

A file delivered this morning might contain activity collected earlier. A contact record updated this afternoon might have changed because a company name was standardized, not because the person expressed renewed interest. An enrichment process can improve an address without refreshing an old purchase signal. Ask exactly what caused each timestamp to change. If the provider only supplies a generic “last updated” value, do not relabel it as the last time the buyer researched your service.

HubSpot's intent-signal documentation describes a backfill of earlier signals when tracking begins. Use intent signals That is a concrete example of why a newly visible event need not be a newly occurring event. The practical safeguard is to preserve occurrence time alongside arrival time. When the actual occurrence time is unavailable, label it unknown and limit the action accordingly. It is better to research an uncertain account than to send an apparently timely message built on a timestamp that means something else.

Dates that should not be treated as interchangeable
DateQuestion answeredWhat it does not prove
EventWhen did the behavior occur?Current need
CollectionWhen was it recorded?Immediate delivery
Provider updateWhen did a provider change the record?New buying activity
ImportWhen did your system receive it?Recent behavior
VerificationWhen did your team check specified facts?That every field was checked

Use a consistent timezone and keep the original timestamp where practical. The purpose is not to make ordinary prospect review technically elaborate. It is to prevent a record from becoming younger each time it is copied. If a spreadsheet export drops the event date, the downstream campaign cannot apply the original freshness rule correctly. Include the field in the transfer contract, test a known record across the systems, and flag a missing date before the campaign takes an action that depends on recency.

How do you verify what the original signal represented?

Read the event definition and source record before interpreting age. Identify whether the signal describes a person, company, topic, conversation, or model-generated assessment. Then ask what behavior was actually observed and what alternative explanations remain. A precise timestamp attached to an ambiguous signal does not establish a precise purchasing intention.

HubSpot's buyer-intent documentation describes company-level website activity with fields such as visits and last visit. How to Use Buyer Intent Data to Identify Companies Ready to Buy Its Company Surge documentation describes changes in company research activity on topics. Use Company Surge signals, powered by Bombora These observations can inform investigation, but they are different from a person requesting a sales conversation. Preserve the original category in your record. Do not translate every category into “ready to buy,” then attempt to solve the resulting uncertainty merely by shortening the allowed age.

Ask what makes the event relevant to your actual offer. Research into a broad operational subject could concern training, an existing product, a competitor review, or a planned purchase. If the source does not resolve those possibilities, your next step should not pretend it does. A hypothetical business reading about office relocation might be moving, advising a customer, or publishing an article. Your message should be grounded in a relevant business problem you can help with, not an unsupported claim that you know the reader's situation.

W3C's provenance overview describes information about how data was produced as useful for assessing reliability. W3C provenance overview Apply that idea by retaining the source, the event definition, any transformation, and the interpretation your team added. When a reviewer asks why a record entered the campaign, the answer should lead back to evidence rather than a chain of summaries. If the chain ends at an unexplained score, lower confidence and request better documentation before using the score for a more intrusive action.

A woman takes notes during a phone call beside an open laptop.

Why does the identity behind a signal need rechecking?

Identity needs rechecking because the event may concern an organization rather than the named contact, and the relationship between the person, role, and need may have changed. Verify the entity the source identifies and the person relevant to the current decision. Do not transfer old company activity to a particular employee without supporting evidence.

Mike's written guidance cautions that activity associated with a company does not establish that a particular employee wants to buy. That recommendation remains relevant even when the company match appears reliable. A large organization can have several locations, business units, and purchasing projects. The person you intend to contact may have no involvement with the observed topic. An accurate company record is useful context, but the appropriate next action may be account research rather than personal outreach based on supposed browsing behavior.

Check the role at the time of the event and the role now where the distinction matters. A contact may have changed responsibilities or left the organization. A former project sponsor may now advise a different team. Preserve old context as historical information rather than presenting it as a current mandate. A digital marketing agency evaluating signal quality should be able to explain what identifies the account, what identifies the person, and which connection remains an inference.

Also inspect duplicate and merged records. HubSpot documents record matching and deduplication using email addresses, domains, and record identifiers. Deduplicate records in HubSpot Those mechanisms deserve testing when history crosses systems. A merge can bring old events into a newer-looking record, while a duplicate can hide a previous conversation or preference. Review the history relevant to the intended action, not just the latest visible card. A recently created contact record is not proof that every event attached to it is recent or belongs to the same decision.

How do you decide how long a signal should remain actionable?

Choose a review window based on the underlying need, the proposed action, the evidence strength, and the cost of being wrong. Treat the window as an operating hypothesis that your own outcomes can refine. It should trigger reconsideration, not certify that all younger signals are valid or all older signals are worthless.

Begin by naming the action. Preparing account research requires less certainty than telling a person that you know they are purchasing. Responding to a direct unanswered request differs from adding someone to a new prospecting sequence. Set the review condition for that specific use, then document why it is reasonable. If the business cannot explain the window without appealing to a software default, it probably has not connected freshness to the real buying process.

Use meaningful events as well as elapsed time. A stated purchasing deadline, a project completion, a role change, a clear decline, or an active conversation can change the appropriate action before a timer expires. Conversely, a customer may explicitly ask for a later discussion, making that agreed date more relevant than a generic aging rule. Keep the commitment and its context visible. Do not let the system restart a prospecting sequence simply because the record again satisfies a numerical score threshold.

Set a separate review window for each signal type when their meanings differ materially. Do not overcomplicate the model with dozens of distinctions the team cannot maintain. A small set of understandable categories, such as direct requests, observed account activity, and historical research context, can be more usable. The companion article on validating audience segments helps connect selection to suitable outcomes. Freshness review adds the question of whether the evidence still supports the selection at the moment the business acts.

When should a new event refresh an old assessment?

A new event should refresh only the part of the assessment it actually supports. New relevant activity can strengthen evidence of current interest, but it does not automatically confirm identity, service fit, authority, or an unresolved need. Preserve the old event and record the new one separately so the updated conclusion remains explainable.

Suppose a hypothetical prospect researched a service some time ago and later attends a relevant discussion. The new attendance can justify revisiting the account. It does not prove the original project is still open. A practical next step might be to review the relationship and ask whether the topic is currently useful. Treating the attendance as a total reset would silently carry old assumptions forward, including the original budget, timeline, and decision role, none of which the new event confirmed.

Review how software combines events. HubSpot's score-building guidance describes criteria based on properties and actions, including aggregation choices for associated records. Build lead scores to qualify contacts, companies, and deals The underlying configuration can therefore affect whether a new activity dominates an older assessment. Test the behavior with a known example instead of assuming the displayed score has the meaning your team expects. A total can rise while the original signal becomes less relevant, especially if the new activity concerns a different topic or associated person.

Check for duplicate observations before calling something corroboration. A provider alert, an exported file, and a CRM update may all represent the same underlying event. Record an event identifier or another practical matching method where available. Independent evidence can strengthen a conclusion; repeated copies do not. If you cannot determine overlap, preserve that limitation. Your review should be able to say that several records were received without claiming that several separate buying actions occurred.

How should score decay influence a real decision?

Score decay can reduce the influence of older activity, but it cannot determine whether the need still exists or whether contact is appropriate. Use it to manage priority after checking eligibility and relationship status. Keep its rule visible and test it against known events, because a declining number can otherwise create a false sense of careful judgment.

HubSpot documents decay applied to individual scored events and notes limits for segment-membership criteria. Overview of the lead scoring tool That product-specific behavior illustrates why teams should inspect the actual implementation rather than assume every score ages automatically. A score based partly on persistent properties can behave differently from one based on dated events. If the business reads both as “current intent,” the interface may imply more than the formula measures. Give the score a label that reflects its components and intended use.

An illustrative policy might move a record from active consideration to research review once the relevant evidence exceeds the team's chosen window. The exact window would be a business assumption, not an industry benchmark. Another policy might retain the record for content planning while removing it from sales prioritization. Both approaches distinguish the continued usefulness of the information from its ability to support a specific contact action. That distinction is often more important than finding a sophisticated decay formula.

Never let a decayed score erase a decisive newer fact. A clear stop request, incompatible need, or completed purchase requires an appropriate state change, not merely a points adjustment. Likewise, a current direct request deserves attention even if an older engagement score is low. Use the score as one input to an explainable decision. The record should still tell the operator what happened, when it happened, and what the person or business expects next.

A hand writes on cream paper beside a brass desk clock.

What must you check before reusing an old audience?

Before reusing an audience, recheck identity, service fit, the underlying need, relationship ownership, preferences, and the original source dates. Reuse should be a new eligibility decision, not an assumption that the earlier campaign's judgment still applies. Keep the historical selection criteria so reviewers can understand what changed and why the new campaign is appropriate.

Start with the original offer. A list built for one service may not be suitable for a different one, even if the contact details remain accurate. Review what the people saw, what they requested, and how the prior campaign ended. An old interest signal should not be detached from that context merely because the business has a new message ready. If the new offer solves a different problem, do fresh audience work rather than describing the change as a routine refresh.

Next, reconcile the existing relationship. Someone might already be a customer, discussing a proposal, waiting for a promised response, or no longer interested. HubSpot's workflow documentation describes suppression and unenrollment controls for managing records that should leave workflows. Set unenrollment triggers in workflows The practical application is to check the current state before reuse and ensure the downstream action honors it. A new audience name should not conceal an old unfinished conversation or create competing messages from the same business.

Finally, inspect how automatic selection works. HubSpot's configuration guidance describes saved views and ways to add companies into records, segments, or workflows. Configure buyer intent Review the dates and criteria behind any such automation in your own setup. A saved view can preserve a rule while the business assumptions behind that rule become outdated. Assign an owner to review the definition, not only the records flowing through it. The intent activation guide explains the wider connection between selection, messaging, and the next step.

What should a freshness review tool tell you?

A freshness review tool should tell you whether to stop, route to an existing owner, revalidate the evidence, or consider a limited next step. It should show which missing or outdated information controls the result. It should not present age alone as a probability of purchase or imply that a chosen review window is a verified market standard.

Intent freshness decision worksheet

Choose the evidence you have checked, then review the next action.

The tool uses your chosen review window solely as a planning input. Enter the age of the underlying event, not the time since a file was uploaded. If the event date is unknown, use the unknown option rather than guessing a favorable value. Then indicate whether identity and current relevance have been checked and whether a stop condition or existing conversation changes the route. The output follows an explicit order: stop first, preserve ownership next, resolve unknown evidence, then compare event age with the chosen window.

Consider a hypothetical event that is older than the team's review window but belongs to a known suitable account. The tool recommends revalidation; it does not declare that the account will never buy. If a new direct conversation confirms a current need, the operator can record that evidence and make a fresh decision. In another hypothetical case, an event inside the window still receives a hold when the person or current relevance is unverified. Passing a timer is not equivalent to passing the whole review.

The tool stores nothing and sends no inputs elsewhere. Its rules are recommendations for organizing a review, not a validated prediction or a substitute for source inspection. The same method can be used in a worksheet: record the event, preserve its age, check decisive changes, and choose the least presumptive useful next action. When the evidence supports only research, that is a complete and meaningful result. It is better than forcing every alert to produce another message.

How can you evaluate a freshness policy without fooling yourself?

Evaluate freshness policies by comparing clearly defined age groups and downstream outcomes while accounting for differences in signal type, audience, and treatment. Use the original event age at the time of action. Report uncertainty and alternative explanations rather than assuming that newer records caused better results or that older records have no commercial value.

Start with a decision you can observe. For example, you might examine whether revalidated older accounts produce suitable conversations and how much research they require. Keep the outcome definition consistent across groups. A reply, a booked appointment, an attended meeting, and an accepted opportunity are different events. If one group receives personal research and another receives a generic sequence, any performance difference reflects more than signal age. Record that treatment difference before drawing conclusions.

Avoid measuring only the records that were contacted. The policy may keep valuable records on hold indefinitely, making the active campaign look efficient while hiding missed opportunities. Review samples from the held group and document why they remained there. Equally, a campaign that reaches many people with weak old signals may create work that never appears in acquisition reporting. Track the effort to investigate, correct, and handle unsuitable conversations alongside the outcomes the business values.

Use enough observation time for the decision being studied and preserve the original cohort definitions. A long purchase process can make an early comparison misleading. Do not quietly move successful older records into a newer group after a later activity event, then claim that fresh signals produced all the opportunities. A digital marketing company should explain the measurement rule before presenting the result. This is an analytical recommendation, not a promise that a particular experimental design or sample will be available for every business.

A man reviews an open folder while a woman sits across his desk.

How should your team handle conflicting dates and evidence?

When evidence conflicts, preserve the conflict, identify the most direct reliable source for the question, and hold any action that depends on resolving it. Do not average incompatible facts into a confident-looking score. A current direct statement can outweigh an older inference, while a new system update may have no relevance to the buyer's circumstances.

Imagine a hypothetical record with an old proposal request, a later note saying the project was postponed, and a new company-level topic alert. The alert does not cancel the postponement. Review the note and any commitment about future contact before deciding what to do. A relevant next action might be a reminder to the existing owner to inspect the account, not automatic enrollment in a new sequence. The output should preserve the difference between a possible change and a confirmed reopening.

Distinguish a data correction from a new buying event. A supplier may correct a company name, a domain may change, or an internal user may update an industry label. Those changes can improve record quality without providing new evidence of demand. HubSpot's buyer-intent FAQ distinguishes enrichment outputs from signal outputs. Buyer Intent, Intent Signals, and Custom Signals: Frequently asked questions That specific distinction reinforces the need to understand what each integration writes. Review the source field, transformation, and trigger before allowing an administrative change to activate outreach.

Name the person who resolves conflicting cases. Without ownership, records can remain indefinitely suspended or be repeatedly released by whichever system runs last. Give the reviewer a concise evidence timeline and an explicit decision to make. Record both the conclusion and what remains unknown. If the evidence is insufficient, keep the next action proportionate: research, wait for the agreed event, or close the current pursuit. Uncertainty is a valid state and should remain visible until something actually resolves it.

How do you make freshness review part of everyday work?

Build freshness into record fields, campaign entry checks, delayed-action checks, and review ownership. Start with the few facts needed for the next decision and test them across the full workflow. The process should help an operator understand why a record is actionable today, rather than require someone to reconstruct its history every time.

Use a compact record structure: original event, event date, source, entity level, current relevance check, relationship owner, stop condition, review due date, and next action. Keep factual evidence separate from the recommendation. An operator should be able to see that the account showed topic activity on a particular date and that a reviewer later recommended research. Combining those into a single “hot lead” label hides both the observation and the judgment that followed it.

Test records that exercise the difficult transitions. Include a newly imported old event, a current event with unknown identity, an existing customer, a record with a stop request, and a signal that ages out while waiting in a sequence. Verify the downstream result after each transition. If the message tool cannot check current eligibility immediately before acting, account for that limitation in your operating process. Do not claim that a scheduled nightly update delivers an instantaneous stop across every connected system.

Assign a periodic review of the policy itself. Business offers, delivery capacity, data sources, and buying processes can change even when the automation remains technically functional. Ask whether each field still changes a decision and whether the review queue is manageable. Retire unnecessary complexity, investigate repeated exceptions, and preserve a simple explanation of the current rules. Freshness becomes useful when it improves the next interaction, not when a dashboard displays more dates than the team can interpret.

An effective operating note also identifies what was not checked. A reviewer might confirm a person's current role without confirming the project's status, or establish that an account still fits without identifying the relevant buyer. Preserve those partial results instead of treating the word verified as a universal endorsement. A later colleague should know which question was answered and which remains open. That reduces the chance that a modest research update becomes a stronger assertion as it travels through reports and handoffs.

When introducing a new source, run a short reconciliation exercise with familiar records. Compare its event labels and timestamps with what your team can independently establish. Ask the supplier to explain differences without requiring it to disclose information it does not possess. If the explanation remains unclear, narrow the permitted use of the feed. An unfamiliar dataset should earn a role through understandable evidence and useful decisions, rather than inherit the same authority as a direct conversation simply because both appear in the customer database.

What other freshness questions do buyers ask?

Freshness decisions become clearer when you separate data accuracy, current interest, and the action being considered. The questions below address common edge cases without inventing a universal expiration period. Use each answer to identify the additional evidence needed, then preserve that evidence in the record so the decision can be reviewed later.

Is real-time data always better than older data?

Real-time delivery can reduce one kind of delay, but it does not establish that the event means what your team assumes. Check identity, source meaning, relevance, and the intended action first. A carefully documented older customer conversation may support a better next step than an immediate alert with unclear context.

Does a verified email address make intent current?

A verified email address concerns the contact channel, not whether the person has an active purchasing need. Treat address quality and intent freshness as separate checks. A reachable person can have completed the project, changed roles, or asked to stop. Review the relationship before converting technical reachability into a sales action.

Should old intent records be deleted immediately?

Old intent may remain useful as appropriately managed historical context even when it should no longer drive a campaign. Decide retention separately under your organization's applicable requirements. Preserve the distinction between history and current eligibility, and avoid allowing retained records to re-enter active outreach merely because another system imports them again.

Can a new salesperson restart an old opportunity?

A new owner should review the original request, later conversations, commitments, and outcome before choosing a next step. Reassignment does not make the opportunity new. If a current need is unconfirmed, begin with research or an appropriate relationship-based check rather than representing an old signal as evidence of present readiness.

What if a provider will not disclose the event date?

Record the event date as unknown and ask what the available timestamp actually represents. Limit use to decisions that can tolerate that uncertainty, such as preliminary account research. Do not substitute delivery time for occurrence time. If recency is essential to the campaign, unavailable event timing is a material limitation.

How can The Mangione Group help you use intent data more carefully?

The Mangione Group can help connect buyer research to the decision your campaign needs to make. That includes examining what a signal represents, whether it still matters, and how much confidence the next action requires. Useful research should improve relevance without making claims about a person's circumstances that the evidence cannot support.

Our buyer intent services focus on audience questions, comparisons, signals, and practical next actions. A freshness review can identify where a list needs new research, where an existing conversation should remain with its owner, and where an old assumption should stop driving the campaign. Your business supplies the service boundaries and relationship context that data alone cannot establish.

Our outreach funnels work helps connect selection, messages, destinations, follow-up, and scheduling. A marketing company should be able to explain that connection in ordinary language, including what happens when evidence changes between the first contact and the next step.

For a 30-minute Marketing Review, bring the campaign objective, the kinds of signals you use, and appropriately handled examples of records that are difficult to judge. We can discuss where clearer dates, better source definitions, or a more suitable follow-up process would make the next decision easier.

Michael Mangione

Michael Mangione, founder of The Mangione Group.

Founder, The Mangione Group

Michael Mangione brings 18 years of marketing expertise. He's assembled a leadership team whose track record includes more than $400 million in advertising spend managed, over 600,000 leads generated, and more than $1 billion in client outcomes.

He built The Mangione Group to help businesses compete with greater precision, win more customers, and grow revenue. His approach brings innovative technology and AI together with sharp strategy, persuasive copywriting, and compelling design to make businesses stand out, make their value unmistakable, and give customers a reason to choose them.

About Mike Mangione | Mike Mangione on LinkedIn

References and Sources

  1. Overview of the lead scoring tool. HubSpot Knowledge Base. Updated September 2, 2026. Checked October 25, 2026.
  2. Set unenrollment triggers in workflows. HubSpot Knowledge Base. Updated August 18, 2026. Checked October 25, 2026.
  3. How to Use Buyer Intent Data to Identify Companies Ready to Buy. HubSpot Knowledge Base. Updated October 13, 2026. Checked October 25, 2026.
  4. Configure buyer intent. HubSpot Knowledge Base. Updated October 13, 2026. Checked October 25, 2026.
  5. Use intent signals. HubSpot Knowledge Base. Updated September 29, 2026. Checked October 25, 2026.
  6. Use Company Surge signals, powered by Bombora. HubSpot Knowledge Base. Updated September 16, 2026. Checked October 25, 2026.
  7. Buyer Intent, Intent Signals, and Custom Signals: Frequently asked questions. HubSpot Knowledge Base. Updated September 24, 2026. Checked October 25, 2026.
  8. Build lead scores to qualify contacts, companies, and deals. HubSpot Knowledge Base. Updated August 14, 2026. Checked October 25, 2026.
  9. Deduplicate records in HubSpot. HubSpot Knowledge Base. Updated October 2, 2026. Checked October 25, 2026.
  10. PROV-Overview. W3C. Published April 30, 2013. Checked October 25, 2026.