A contact record can look complete while individual fields have very different levels of evidence. The business name may be confirmed on the official website, a phone number may appear in several public sources, and an email may be older or ambiguous. A responsible confidence system scores fields—not people—and preserves the evidence needed for review without pretending uncertainty has disappeared.

VISUAL LESSON

What you will learn

  1. 01Design a field-level confidence rubric.
  2. 02Store source, review date, agreement, and verification evidence.
  3. 03Create action rules for high-, medium-, and low-confidence fields.
Several public source nodes feeding contact fields that receive confidence layers and review-time markers before entering a CRM tray
Confidence belongs at the field level because a single business record can contain both strong and weak evidence.

ILLUSTRATIVE 100-POINT RUBRIC

A score summarizes evidence; it does not replace it

Source relevanceOfficial or direct context
35 pts
RecencyObserved review date
25 pts
Cross-source agreementIndependent consistency
25 pts
Verification evidenceField-specific check
15 pts
Hypothetical rubric—not an industry standard or accuracy guarantee. Adapt weights to the field, use case, and review capacity.

DATA-QUALITY INTERFACE MAP

Make uncertainty actionable

Field01Store value and source

Attach the source URL or system, observed date, research method, and whether the value was copied, normalized, or inferred.

Review02Apply the rubric

Score source relevance, recency, agreement, and any field-specific verification; record conflicts instead of silently choosing a winner.

Action03Route by evidence

Use high-confidence fields after final review, recheck medium-confidence fields, and hold or remove low-confidence values rather than inventing certainty.

Conceptual CRM map. Field names, confidence features, and source systems vary; preserve the underlying evidence even when the CRM stores only a simplified grade.

THE FIELD-EVIDENCE LAYER

Source + recency + agreement + verification

RecordAttach evidence to each field
ScoreApply one documented rubric
RouteUse, recheck, hold, or remove

THE LEAD ATLAS METHOD

Lead Atlas Data can research a campaign-specific contact list for the customer's chosen locations, markets, and business categories, while a confidence-and-source layer helps the customer decide which fields need its own final review before use.See how custom list research works ↗
01

Choose fields that need confidence

Start with campaign-critical fields such as business identity, domain, location, category, email, phone, contact name, and role. Define what error would harm the customer experience or reporting for each one.

Score fields individually. A verified domain does not automatically validate a named contact, and a current phone number does not prove that a general email reaches the intended role.

02

Define source and recency rules

Rank sources by their relationship to the field: an official contact page may be strong for a published phone, while a third-party directory can be a useful clue that requires confirmation. Record the exact page or system and the date it was reviewed.

Recency should reflect when the evidence was observed, not when the spreadsheet was downloaded. Use different refresh windows for fields that change at different rates.

03

Handle agreement and conflict

Independent agreement can raise confidence, but copied directories are not truly independent. When two credible sources disagree, keep both values in the exception log, inspect the official business context, and assign a reviewer.

Never average phone numbers, merge names, or infer a personal email pattern just to resolve conflict. Unknown is a valid state.

04

Connect scores to actions

Write action bands such as use after review, recheck before use, alternate channel, hold, and remove. Tailor the threshold to risk: a website used for fit research may need a different rule than a direct contact field used for outreach.

A confidence score is not permission, consent, deliverability, or a guarantee that someone will respond. Keep compliance, suppression, and customer-choice controls separate.

05

Run the 30-record calibration

Have two reviewers independently score thirty records, compare disagreements, and inspect the fields that produce inconsistent decisions. Adjust definitions before scoring the full list.

Deliverable: field rubric, source hierarchy, freshness windows, action bands, conflict log, calibration results, reviewer guide, and the CRM columns needed to preserve evidence and decisions.

LEAD ATLAS WORKBOOK

Turn this lesson into a research brief.

Apply “Create a Contact-Data Confidence and Source Review System” to one campaign before requesting or using a list.

  1. 01Market boundary

    Name the locations and business categories this decision applies to.

  2. 02Fit evidence

    Write the public signals that would make a business relevant enough to review.

  3. 03Exclusions

    List the business types, markets, and records that should not enter the campaign.

  4. 04Outreach use

    State who will review the list, personalize the message, and record outcomes.

THE TAKEAWAY

Score each critical field from explicit evidence, preserve source and review dates, route low-confidence values to recheck or alternate channels, and never present a score as certainty or consent.

OFFICIAL REFERENCES

Check the platform’s current instructions.

Platform labels, eligibility, and workflows can change. These official help pages were used to validate this lesson.