Build a Business Lead Database: Fields, Cleanup and Examples

A useful business lead database connects a business to a reason for contacting it, a suitable contact route and an accountable next step. Start by defining the account you want, preserve the evidence behind each record, and test a small import before collecting more data.

This guide walks through that process for a sales operator or agency building a repeatable prospecting workflow. The examples and calculations are illustrative, not customer results. Product details and external sources were reviewed on September 23, 2026.

Define the account before choosing the database

Write a short brief that a second researcher could apply without asking what you meant. “Businesses in Texas” leaves too much open. “Independent plumbing companies serving our delivery territory, advertising installation work and not already customers” gives someone a usable decision.

Specify the offer, geography, business type, exclusions and research capacity. Decide whether a business must operate from the territory or merely serve customers there. Record how you will handle franchises, branches and companies with several service lines. These choices change which rows belong in the database.

Then define what a row represents. A business location, company, person and sales opportunity are different objects. A company can have four offices and one central purchasing team. One office can list a general inbox and several people. None of those facts creates an opportunity until a sales conversation establishes one.

For location-based prospecting, start with one business location per row. Add a separate account identifier when research connects locations to the same organization. Keep people and their email associations in a related contact table, or clearly labeled columns if you are using a spreadsheet. An account-level exclusion should affect every linked location; a correction to one branch address should not.

Your initial database can be a spreadsheet. Move to a CRM or shared database when several people need ownership, activity history and controlled updates. The deciding question is whether teammates can understand and change a record safely, rather than whether the software can hold another thousand rows.

Begin with the business category that matches your intended buyers. Category names help discovery, but your written fit criteria determine whether a record deserves further research.

Design a small set of fields that supports real decisions

Separate imported facts from decisions your team makes. Keep the original business name, address, website and contact details alongside normalized versions used for matching. This lets you correct a cleanup rule without losing what the source actually supplied.

For the working database, use a simple field dictionary:

These are recommended working fields, not a claim that every source exports them. Give each field a clear definition. For example, “reviewed” should mean a researcher completed named checks, not simply opened the row. Use a small controlled status list such as New, Research needed, Ready for review and Excluded, with the reason stored separately.

Leadz CSV exports can include business contact and location details. They can also include names, reported roles and LinkedIn profiles associated with individual email addresses, where available. Keep those associations intact: a person's profile attached to one address should not be reassigned to the business's other inboxes.

The CSV field guide explains the available columns. Use the illustrative business-leads sample to plan your field mapping before working with production data. Blank contact fields should remain explicit unknowns; filling them with guessed roles makes later reporting less reliable.

Evaluate a data source against your actual market

Choose a source by the work it removes from your process. Manual research suits a narrow, high-value account set whose relevance depends on detailed portfolio or service information. A structured business directory is useful when category and geography establish a sensible starting audience. First-party enquiries answer a different question because the business has already interacted with you.

Ask for a sample or inspect available previews before expanding. Select records across your intended territory, including smaller cities and less obvious businesses. A sample of prominent companies with complete websites can flatter a source that performs differently elsewhere.

Judge the sample on distinct questions: Are the businesses in scope? Are locations duplicated? Do websites open? Which contact fields exist? Does the source explain what verification means and when it happened? Can you export and retain the fields needed for your intended use? Document unclear answers instead of converting them into a quality score.

An outside benchmark can help you challenge coverage claims. As of September 23, 2026, the Census Bureau's County Business Patterns data page identifies 2023 as its most current reference year. Those data were released on June 26, 2025 and describe businesses with paid employees. A 2025 release date does not make the observations a 2025 business census, and employer establishments are not the same population as all directory listings.

Use that distinction when someone claims to cover an entire market. Compare compatible geography, industry definitions, reference periods and business populations. A difference may reflect sole proprietors, branches or classification choices rather than missing records.

Evaluate cost after review, too. Suppose two illustrative sources each supply 200 rows. Source A costs $120 and takes three hours of review at an assumed $30 per hour, leaving 144 usable unique accounts. Its reviewed cost is $210 divided by 144, or about $1.46 per account. Source B costs $60 but needs six hours of review and leaves 80 usable accounts: $240 divided by 80, or $3.00. These are planning assumptions, not market prices.

Keep “usable” consistent across both samples. The contact-sample evaluation workflow helps turn that comparison into a repeatable buying decision. For a platform shortlist, use the lead-generation tool comparison workflow to record sample quality, setup effort and ongoing cost together.

Clean duplicates without erasing branches

Preserve a raw import, then work on a copy. Normalize spacing and case for matching, reduce websites to a comparable domain, and standardize phone formatting while retaining the original numbers. Treat postal codes and identifiers as text when importing so formatting changes do not silently alter them.

Use several fields to investigate matches. An identical source identifier is strong evidence of the same source record. A shared domain plus a matching street address may indicate a repeated location. A shared domain with different addresses may indicate branches. A central phone number or generic inbox can legitimately serve several locations.

Consider Cedar Studio, an invented architecture practice. Its downtown and north offices share a website and general email address. A third imported row repeats the downtown address under a slightly different business name. Keep two locations linked to one account, and review the third row as a likely duplicate. Merging by domain alone would erase the north office; treating every row as a company would inflate the account count.

Use three outcomes: merge a confirmed duplicate, link confirmed branches, or hold an ambiguous match for research. Save the surviving ID, removed ID and reason for any merge. Do not let the cleanup script decide that two similarly named businesses must be related.

Here is an illustrative reconciliation:

The unit changes from rows to locations to accounts. Report that change rather than calling every reduction a lost lead. Retain excluded and unresolved records with their decisions so the next import does not create the same work again.

For a fuller import checklist and example file, follow the branch-aware deduplication workflow.

Qualify business fit and contact readiness separately

A company can be a good account while its contact route remains uncertain. Record two decisions: business fit and contact readiness. Combining both into a single “good lead” checkbox hides whether the next step is research, contact validation or outreach.

Business fit should refer to evidence that matters to the offer. For a supplier serving commercial installers, a relevant service page or project description is stronger evidence than a high review score. Record a short observation and its source URL: “Website lists commercial water-heater installation” is more useful than “Looks good.”

Contact readiness asks different questions. Is an address present? Does the associated name or role make sense for this business? Is it a general office inbox, a branch contact or a named person? Has deliverability been checked recently enough for the planned send? Who owns the next verification step?

A populated email field answers only the first question. Reported role information can guide research, but it does not establish purchasing authority or a relationship with a particular branch. The email availability and verification guide separates these checks so they can be recorded consistently.

Keep suppression independent of both scores. Before outreach, reconcile selected accounts and contacts against existing customers, open opportunities, opt-outs and other exclusions. Preserve the scope of an objection: a person's request and an account-wide exclusion may require different handling.

For U.S. commercial email, the FTC's CAN-SPAM business guide explains that the law also covers business-to-business messages and requires honoring opt-out requests. Build an exclusion check into the handoff, and assess other applicable rules for your recipients and channels rather than treating a public address as permission.

The lead qualification scorecard is useful when several researchers need to apply the same business-fit rules. Review a handful of decisions together before splitting a large queue.

Test the export and CRM import before scaling

Use a small reviewed segment for the first end-to-end run. Confirm that the source selection matches your intended category and territory, choose the columns you need, and record the selection date. A public category total is background context; the selected records are your import denominator.

In Leadz, CSV export requires an eligible paid plan. Export availability and newly unlocked records can depend on your plan and available credits, so review the confirmation shown in the product and the current plan details. Choose fields for the job instead of assuming every available column belongs in every file.

Keep one exported business row as one business record during preparation. Multiple email addresses or associated people do not automatically become separate accounts. If your CRM stores people separately, create a contact mapping that retains the business ID and the specific email-to-person association.

Import into a staging list or clearly tagged batch first. Check a known branch group, a business without an email, a record with multiple contact values, and an address with a leading-zero postal code. Confirm that phone numbers remain intact and that the CRM did not overwrite a researched value with a blank import.

Decide your update rules before approving the batch. New source values can enter review when they conflict with a recent verified correction. Imported categories should not erase a salesperson's qualification notes. Most importantly, a new import must not reactivate an opted-out contact.

Reconcile accepted records, rejected records and intentionally skipped records to the original batch. Save an import receipt with the source file, mapping, run date and owner. When the totals disagree, investigate the mismatch before repeating the import; a second upload can turn one confusing batch into two overlapping problems.

Maintain the database around the next decision

Give one person responsibility for data rules and shared exceptions, even if several teammates perform research. Keep account ownership separate: the sales representative manages the relationship, while the database owner resolves conflicting records and maintains the import rules.

For a two-person sales team, a workable routine might give the operations lead thirty minutes each Friday to review unresolved duplicates, failed imports and corrections reported by sales. The representative preparing Monday's campaign reviews that specific segment before handoff. Opt-outs enter the exclusion list as they arrive rather than waiting for the weekly review.

Each review should end with an owner and a next action. An uncertain branch match goes to the account representative to confirm the address; a repeated import problem goes to the operations lead. Keep collection, source-update and review dates in separate fields so the next person can tell which evidence changed. Adjust the routine when the queue grows or the campaign requires more frequent checks.

Track a few operational measures with explicit denominators: usable unique accounts divided by reviewed accounts, confirmed duplicates divided by imported rows, and contact-ready accounts divided by in-scope accounts. Add review time per usable account if research capacity is your constraint. Email opens or replies belong to later campaign reporting, not the initial database-quality score.

Your next step is to write one audience brief, map the sample fields and review a small batch through the entire workflow. Save the exclusions and corrections as carefully as the accepted records. Once that batch can be imported, assigned and revisited without guesswork, expand the category and territory search using the same rules.