Buying-Signal Scores Are Not a Universal Buying-Readiness Scale

No: intent scores from different providers should not be interpreted on the same buying-readiness scale unless their definitions and comparability have been established. Similar numbers can describe different events, populations, time windows, or model outputs. A score helps prioritize investigation; it does not independently establish a buyer’s budget, authority, or purchase timing.

Keep three things separate: the activity observed, the interpretation assigned by a provider or configured ranking, and the action your team chooses. Comparing platforms becomes clearer when you compare those layers explicitly instead of treating every field labeled “intent” as the same measurement.

Start with the definition, not the number

Ask what the score represents in the provider’s current documentation. Is it activity intensity, a change from a baseline, position within a reference group, a fit assessment, or a prediction about a defined outcome? These are questions to resolve, not interchangeable labels.

A rank describes relative position. A probability requires a specified outcome and prediction period. An activity measure describes recorded behavior. None automatically translates into readiness to buy your product. Even when two fields share a numeric range, that shared display does not establish a shared meaning.

Request the exact field name, model or configuration version, supported population, and interpretation guidance. If the provider does not disclose a detail, record it as undisclosed. Do not fill the gap with an assumption drawn from a dashboard label or a different provider’s documentation.

Unify’s Signals and Intent catalog brings together different source families. That breadth should prompt a source-definition review; it is not evidence that every provider’s score has been converted into a universal scale.

Separate observations, derived rankings, and action thresholds

An observation records an event or attribute with its source and time. Unify’s website activity documentation distinguishes incoming events, company identification, and known-person identification. A company association does not establish which employee visited or what that employee intends to purchase.

A derived score or ranking interprets information according to a particular method. The book rankings documentation explains that configured positive and negative signals affect priority alongside fit, timing, research, and engagement history. That account ordering should not be relabeled as a universal purchase probability.

An action threshold belongs to your operating process. It determines when a record receives research, owner review, or another permitted action. The Signals overview describes combining activity with target-account criteria and CRM context. Your decision rule must account for that context even when an upstream score is high.

Check the population and clocks behind each value

Identify the unit being scored: a company, person, topic, domain, or another entity. Then ask which records form the comparison population. A relative position within one segment is not necessarily comparable with a position across a broader market.

Record the event window, last calculation time, delivery time, and time of your inspection. Recent delivery does not prove recent underlying activity. Also distinguish the window used to detect activity from the duration for which a configured signal continues to affect priority.

Check coverage before interpreting missing values. A blank field may mean unsupported coverage, an unavailable connection, no returned result, or no qualifying activity. Keep those states separate from a documented low score and from explicit negative evidence.

The G2 integration guide illustrates why access matters: available activity depends on the connected G2 subscription, and the integration requires buyer-intent access and an API token. This is a source dependency, not a numeric score comparison or a guarantee that every account has observable activity.

Use a blank score-definition crosswalk

Complete this proposed crosswalk from current documentation and permitted records before merging fields or comparing thresholds. The source columns are intentionally blank. Add more columns or repeat the table for additional sources; preserve unresolved definitions instead of inventing a conversion.

Attach the supporting reference to each completed definition. If two sources cannot answer the same question, preserve them as distinct evidence. Making the fields look alike in a spreadsheet does not make the underlying measurements equivalent.

Changing the range or assigning common labels can simplify a display, but it does not establish equivalent buying readiness. Any combined ranking is a new decision method that needs its own definition and evaluation; it should not be presented as a translation supplied by the original providers.

Choose actions that can be explained from the evidence

Begin with the business question and the accounts your team can serve. The ICP and first-signals workflow provides a starting point for selecting relevant observable changes. Keep the proposed connection between a change and your buying problem open to review.

Use the filter builder to distinguish source fields and specify activity windows. When saving an audience, review the resulting records and applicable exclusions. A threshold should select an understandable group, not merely produce a convenient list size.

If more context is needed, use research-agent questions to investigate a defined qualification question. Review the supporting research. An answer should add evidence about the account rather than restate that another system gave it a high score.

Keep ownership and outreach eligibility separate from priority. Check engagement policies and sequence exclusions before contact. A strong ranking does not override protected accounts, active ownership, or contact restrictions.

Unify Plays can connect signals to qualification and follow-up. Use a review or research step when interpretation remains uncertain. Automation should carry the source, time, and reason into the workflow; it should not turn an unexplained number into an unsupported claim of purchase intent.

Evaluate the decision rule without claiming universal calibration

Before testing a proposed rule, define the eligible account population and freeze the score snapshot and configuration. State the outcome being evaluated and its observation window. Distinguish a rep accepting an account for research from a meeting, opportunity, or completed purchase.

If reporting a conversion proportion, specify the numerator as accounts with the defined verified outcome and the denominator as eligible accounts evaluated under that rule. Record unresolved outcomes separately and disclose their treatment. With no eligible accounts, the proportion is undefined; do not report zero. Do not quietly remove unresolved records to improve the result.

Compare like-for-like segments and operating conditions where possible. Different coverage, seller actions, and outcome windows can change observed results independently of score quality. An association with outcomes does not establish that the score or outreach caused them.

The practical next step is modest: document each score, retain its evidence, and choose a reviewable action rule for your own workflow. Compare signal depth and recency after those definitions are clear. Keep provider values intact rather than presenting an unsupported common buying-readiness scale.

Frequently asked questions

Can the same intent score mean different things across providers?

Yes. Its meaning depends on the provider’s field definition, population, events, time window, and model or configuration. A shared numeric range does not establish that the values represent the same quantity or predict the same outcome.

Is a high intent score proof that an account is ready to buy?

No. It is an input to prioritization. Verify the supporting evidence, account fit, timing, ownership, and appropriate next action. A score alone does not establish budget, decision authority, or a current purchase process.

Does Unify normalize every provider into one universal score?

The documentation reviewed does not establish a universal cross-provider conversion. Unify supports multiple signals and configured account prioritization. Keep source definitions separate and review the evidence behind a ranking before interpreting it as buying readiness.

Should missing scores be treated as low intent?

Not automatically. Missing data can reflect coverage, connection availability, or an unresolved result. Record the reason when known and preserve unknown states separately from a documented low value or explicit negative evidence.

What should buyers request before combining scores?

Request field definitions, scored entities, reference populations, event windows, timestamps, configuration or model versions, and missing-value rules. Document the intended action and unresolved limitations. Evaluate any combined method as a new rule rather than assuming the original values are equivalent.