B2B list building services, with the sources named
GrowthPal is a B2B outbound and GTM engineering agency operated by Inferno Media, LLC, working from Kathmandu, Nepal.
Last reviewed 29 October 2026
B2B list building services research and assemble a contact list for a defined buyer, rather than exporting one from a database. The useful difference between providers is where the records come from. This page names the public registers, license systems and filing archives we use, what each one returns, where each one is empty, and what has to be researched by hand instead.
When B2B list building services are worth buying
Who this is for
- Your segment is defined by something a database cannot filter on. A current license, a specific filing, a company age, an overdue account, a registered legal form. No contact database is built around any of those, and a register is.
- You sell into a market where person-level data is thin. Where the professional network behind most commercial databases is sparse, the company universe is still verifiable from the register and the contact layer has to be researched per account.
- You need to show where the data came from. If your legal or compliance team asks who compiled a list and from what source, a registry answer is short and checkable. A reseller answer usually is not.
- A database export already failed. You bought credits, pulled the list, and the bounce rate or the job titles were wrong. The next step is research, not more credits.
Who this is not for
- You want the cheapest possible rows. A database subscription is cheaper per record than research and always will be. This work is for the segments where the export does not exist.
- You want a consumer list. Everything here is business contact data at business domains, researched against business registers. Consumer marketing has a different rulebook and we do not do it.
- You want the list and nothing else. We will deliver a standalone list, but a list is a hypothesis about a buyer. If nobody is going to test it with a campaign and feed the results back, most of the value is left on the table.
- The definition of the buyer does not exist yet. If the brief is a sector and a headcount band, the list will be broad and the replies will tell you nothing. The first hour of this engagement is spent fixing that, and if it cannot be fixed the list is not the problem to solve.
What the engagement includes
| Component | What is done | What the client receives | Who owns it at the end |
|---|---|---|---|
| Segment brief | Turn the target into a definition that can be executed: the qualifying signal, the disqualifying signal, the roles in scope, the markets, and the exclusions. | A written brief you sign off before research starts. | Client |
| Source selection | Decide, per segment, which register, license system, filing archive or commercial database is the right starting point, and say which fields each one will and will not return. | A source map per segment, with the known gaps stated. | Client |
| Company universe | Build the account list first, from the source that actually defines the segment, before any person is researched. | An account list with the qualifying evidence against each row. | Client |
| Per account research | Find the people in scope at each account, confirm the role against a current public source, and record where each field came from. | Contact rows with a source field that has a real answer in it. | Client |
| Exclusions and suppression | Screen against your customers, open opportunities, past campaigns, competitors and your own do not contact list before delivery, not after. | A suppression log showing what was removed and why. | Client |
| Verification chain | Validate addresses in a fixed order, drop the invalid, and flag rather than silently keep the risky categories such as catch all domains and role addresses. | A delivered file plus a dropped file, so you can see what did not survive. | Client |
| Field dictionary | Publish the column set with a definition and a fill rate against each column, because a column that is 40 percent empty is a different thing from a column that is full. | A dictionary delivered with the file. | Client |
| Refresh | Re-run the source checks on an agreed cadence and report what proportion of the list changed, rather than quietly replacing it. | A refreshed file and a change report. | Client |
How a list gets built
Define the segment so it can be executed
A sector and a headcount band is not a segment. A qualifying signal is: holds a current license of this type, filed this form, incorporated in this window, runs this technology, has this role on staff. The brief has to be specific enough that two researchers would build the same list.
What you provide. Your best fitting current customers, and the reason each one bought.
Pick the source that actually defines the segment
If the defining signal is a license, the regulator holds it. If it is a filing, the filing archive holds it. If it is a status or an age, the company register holds it. A contact database holds none of those, which is why source selection comes before contact research rather than after.
What you provide. Confirmation of the markets in scope, since the source differs by country.
Build the account universe
Accounts first, people second. Building the account list from the defining source means every row carries the evidence of why it qualified, and the ones that do not qualify never enter the file.
What you provide. Your exclusion list: customers, open opportunities, competitors, anyone a rep already owns.
Research the people
Per account, find who holds the role in scope and confirm it against a current public source rather than a cached database record. Where a commercial database covers the market well this is fast. Where it does not, it is hand work, and that is the part being bought.
What you provide. The roles that matter, and the ones that look right but never buy.
Record provenance per field
Every row carries where it came from. That is what makes the list defensible if anyone asks, and it is also how a bad source gets found later when a segment underperforms.
What you provide. Nothing. This is part of the research step.
Suppress, verify, and split the file
Suppression runs before verification so you are not paying to validate addresses you were never going to use. Catch all domains and role addresses are flagged rather than dropped silently, because they are a decision rather than an error.
What you provide. A decision on whether role addresses are in or out for this campaign.
Deliver with a dictionary and a dropped file
The delivered file, the column dictionary with fill rates, and the records that did not make it with the reason against each. A provider who only sends the good file is hiding the yield.
What you provide. The destination: a CSV, your CRM, or the sequencer directly.
Feed the campaign result back into the source map
Reply and bounce data by source is the only honest accuracy measurement available. If one register-sourced segment bounces at four times the rate of another, that is a finding about the source, and it changes the next build.
What you provide. Campaign results by segment, whoever runs the sending.
Where the data actually comes from
Across the pages ranking for this service, none names a single public register, government database or statutory source. The most specific anyone gets is a professional network or an open data reference. Here is the real list, with what each source returns and where each one is empty. B2B list building services are worth buying only when the provider can answer this question, so we answer it before being asked.
| Source | Market | What it returns | Where it is empty |
|---|---|---|---|
| Companies House Company Data Product | United Kingdom, live companies on the register | Company name and number, registered office address, company type, SIC code, status, last accounts and confirmation statement filed, next due dates, previous names. Free monthly bulk download. | No email, no phone, no website, no named contact. The address is the registered office, which is often an accountant's. SIC codes are self declared and frequently generic. It is a monthly snapshot, not live, and it holds live companies only. |
| Companies House PSC data | United Kingdom | People with significant control, as a free snapshot. | Ownership, not buying authority. A person with significant control is rarely the buyer for an operational purchase, and the record is personal data, so the notice obligations attach. |
| SEC EDGAR full text search | US companies that file with the SEC, filings since 2001 | The full text of every filing including exhibits, searchable by keyword, ticker, company name, central index key, form type and date. Free. | Reporting companies only, so the entire US private mid market is absent. Nothing is structured as a contact record: officer names sit inside documents rather than in fields. No email, no phone. Nothing before 2001. |
| Business Registers Interconnection System | European Union and European Economic Area | Company name, legal form, registered seat and registration number through the e-Justice Portal, free for the minimum set. | Depth varies by national register, so the same query returns different richness per country. No contacts. It is not designed for bulk extraction, and national registers charge for documents and detail. |
| California Department of Consumer Affairs license search | California, licensed professions | License status, including current, expired and disciplined, searchable by business name, person name, registry number and license type. | Several California bodies are outside this system and must be checked on their own sites, including the contractors board and the real estate appraisers bureau. The returned field set beyond status is not documented. |
| Texas Department of Licensing and Regulation license search | Texas, dozens of licensed trades | Searchable by license type, number, expiration date, name, city, county and postcode, with bulk license files published. | Active licenses only, so lapsed and pending records are invisible. One state. The returned field set is not documented on the page. |
| SAM.gov entity registrations | US federal contractors and grantees | Entity registrations, exclusions, and responsibility and qualification data, with public data files. | Entities can opt out of public display. Sensitive records are behind federal sign in. Only organizations that do business with the federal government appear at all. |
| National Tax Agency Corporate Number site | Japan, organizations with a designated corporate number | The 13 digit corporate number, plus English name and head office address where the organization registered an English indication, with change history. | English data exists only where an English indication was registered, so an English-only workflow sees a fraction of the register. No contacts, no sector code, no headcount. |
| Commercial contact databases | Strongest in the United States and Western Europe, large and mid market | Person name, title, company, email, direct dial, firmographics, sometimes technographics and intent signals. | Coverage and accuracy claims are published by the vendors themselves and every comparison we found was written by a vendor or a reseller. Thin where the professional network behind the data is thin, and weak on small and micro business almost everywhere. |
When a register is the only real option
- Any list defined by a license or a permit. No commercial database is built around license status, so a list of currently licensed electricians in one state starts at the regulator. The register gives the universe and the status, and contact discovery is a second step.
- Any list defined by a filing or a disclosure. Companies that filed a particular form or disclosed a particular risk. Full text search of the filings answers this for US reporting companies and nothing commercial does, at the cost of excluding private companies entirely.
- Any list defined by company status, age, legal form or filing behavior. Newly incorporated companies, companies with overdue accounts, a particular legal form. The register holds all of it, and no contact vendor filters on any of it.
- Any market where person-level vendor coverage is thin. The register gives a verifiable company universe and the contact layer is researched per account. This is the case that justifies research over a subscription, and it is the one no competing page makes.
- Any list where provenance has to hold up. The first due diligence question a regulator suggests asking is who compiled the data and where it came from. A registry answer to that is one line long and checkable.
B2B prospect research services: the layer above the list
Prospect research and list building are the same job described in two vocabularies, and the SERPs for them barely overlap, which is why both are covered here rather than on separate pages. The research layer is what happens per account once the universe exists: confirming that the role still exists and who holds it now, reading what the company has published about the problem you solve, and writing the one fact that makes the first line of the email specific to that account. It is recorded in named columns with a definition each, not as a free text note, because a column can be used in a sequence and a note cannot.
The compliance layer, stated plainly and not as legal advice
- The buyer of the data is responsible, not the supplier. The United Kingdom regulator is explicit that an organization using a data broker is responsible for its own processing being compliant, that accepting the broker's assurances is not enough, and that if the broker cannot demonstrate the data is reliable then the data should not be used.
- There is a notice obligation on researched and bought lists. Where contact lists are bought or rented for direct marketing, privacy information has to be given to the people on them within one month of obtaining their data. This is the obligation that almost no page in this market mentions, and it applies whether the list was bought or researched.
- The notice has to include the source. Where personal data was not obtained from the person, what is owed includes the source of the data and whether it came from a publicly accessible source, alongside the purposes, the legal basis, the retention period and the rights. The deadline is the earlier of one month, the first communication, or first disclosure to anyone else.
- That is the practical case for registry sourcing. If the source field has to be disclosed, a list whose source field says a named public register is easier to stand behind than one whose source field says a reseller who will not say.
- The due diligence questions are published. Who compiled the data and where it came from, what privacy information was given at collection, when it was compiled and how, consent records where consent is the basis, whether opt out registers were screened and how recently, and how objections are passed on. Any supplier should be able to answer all six.
Accuracy, and how we talk about it
Six of the eight pages we examined for this service claim an accuracy figure. None states how it was measured, on what sample, by whom, or what counts as a failure. One claims 99.95 percent at field level with no method attached. We do not publish an accuracy percentage, because the only honest measurement is the one your own campaign produces: bounce rate by source, and reply classification by segment. Those are reported back per source so the next build is better than the last, and they are a measurement of our work rather than a claim about it.
What to ask before buying a list from anyone
Six questions separate a research service from a database reseller, and all six have short answers if the provider is doing the work.
| What to check | The question to ask | What a good answer looks like |
|---|---|---|
| Source | Name the sources for this segment, and tell me what each one does not return. | Named registers, named databases, and an honest answer about the empty fields. An answer that stops at verified contacts from trusted sources is a reseller answer. |
| Provenance per row | Will the file carry a source field per record? | Yes, with a real value in it. You need this if anyone ever asks where the data came from, and you need it to find the bad source when a segment underperforms. |
| Accuracy method | How was that accuracy figure measured, on what sample, and what counted as a failure? | Either a described method, or an admission that the real measurement is the bounce rate on your campaign. A percentage with no method behind it tells you nothing. |
| The dropped file | Will you send me the records that did not make it, with the reason? | Yes. A provider who sends only the surviving rows is hiding the yield, and the yield is the most useful thing you learn about a source. |
| Exclusions | How do you suppress our customers, open opportunities and competitors? | Before delivery, from files you supply, with a log. Most delivered lists fail on this rather than on data quality. |
| Fill rates | What percentage of rows will actually have each column populated? | A number per column, delivered with the file. A column that is 40 percent empty is a different product from a column that is full, and only one of them can drive a sequence. |
What we can show you for list work
| What GrowthPal has published | What it counts | |
|---|---|---|
| 4M+ cold emails sent for 40+ B2B companies, across Prabesh Khanal's five years of outbound work, including campaigns led before GrowthPal was founded in 2025 | Emails sent, cumulative across every engagement in the founder's career. Not one campaign, not an average, and not confined to GrowthPal. | Source |
| The interested-reply definition and counting rules | Not a figure. A link to the published methodology. | Source |
What stands behind the research
- No list-only result is published yet. GrowthPal has not published a standalone list building outcome, so there is no case study for this service on its own. One registered campaign result credits the list research directly, and it is held on the manufacturing page rather than restated here, because the engagement it describes was more than list building.
- Every source on this page is named and dated. The registers, license systems and filing archives listed above are cited at the foot of the page with the date each was checked, including the ones where the published field set could not be confirmed. Those are marked rather than smoothed over.
- The dropped file is the evidence. Every delivery includes the records that did not survive research, suppression or verification, with a reason against each. That is a harder number to show than an accuracy claim and it is the one that tells you what the source is worth.
- Accuracy is measured on your campaign, not asserted. Bounce rate by source and reply classification by segment come back to you per build. If a source underperforms, it is visible in your own data within weeks.
Questions about outsourcing B2B list building
How is this different from buying a database subscription?
A database answers who matches these firmographic filters. Research answers who matches this defining signal, which is often a license, a filing, a company status or something written on the company's own site. If your segment can be expressed as a database filter, buy the subscription, it is cheaper per row. This work exists for the segments where that export does not exist or comes back wrong.
Are any of these records resold from a data broker?
No. Records are researched from named sources and the source is recorded per row. That matters commercially as well as legally, because when a segment bounces or replies badly you can see which source produced it and stop using that source.
Is it legal to email people on a researched list?
Legality depends on where the people are, and the sending rules and the data rules are two separate questions. On the data side, where personal data was not obtained from the person, the information owed to them includes the source of the data, and the deadline is the earlier of one month, your first communication with them, or first disclosure to anyone else. The practical consequence is that your first email usually has to carry or link to that information. We are not your lawyers and this page is not legal advice.
What does prospect research add on top of the list?
The per account layer: confirming the role still exists and who holds it now, reading what the company has published about the problem, and writing the one specific fact that makes the opening line about them. It lands in named columns with definitions so a sequence can use it, not as a free text note nobody reads.
What fill rate should we expect on each column?
Fill rate varies by source and by market, which is why the number is delivered with the file rather than promised before the work. The pattern is consistent: company-level columns from a register fill close to completely, person-level columns fill well in markets where the commercial databases are strong and badly where they are not, and direct dial numbers are the thinnest column in almost every build.
How do you handle catch all domains and role addresses?
Catch all domains and role addresses are flagged, not silently dropped and not silently kept. A catch all domain accepts everything and so cannot be verified, which makes it a risk decision rather than a data error. Role addresses such as a shared info mailbox are a different decision again. Both come to you labelled so you can choose, and the usual answer is to hold them back for a separate, later batch.
How often does a list need refreshing?
Refresh cadence depends on the source and the market, and anyone who gives you a single number across all of them is guessing. Register-sourced company data moves slowly and is republished on a published schedule. Person-level data moves with the job market in that segment. What we do is re-run the source checks on an agreed cadence and report the proportion that changed, so the cadence is set from your own data after the first refresh rather than from a convention.
Can you work in markets outside the United States and Western Europe?
Yes, with the caveat stated up front: in markets where person-level vendor coverage is thin, the company universe is still verifiable from the national register but the contact layer is hand research, which is slower. We will tell you which of your target markets fall into that category before the work starts rather than delivering a thin file and explaining afterwards.
Do you deliver into our CRM or as a file?
Either. A CSV with the column dictionary is the default, and pushing to a CRM or straight into a sequencer is common. Whichever it is, the dropped records come with it, because the records that failed research are part of what you paid to find out.
Sources
- Companies House data products, Companies House, GOV.UK, published 18 June 2019, last updated 25 October 2022. https://www.gov.uk/guidance/companies-house-data-products. Accessed 29 October 2026.
- Free Company Data Product download, Companies House, snapshot dated 1 October 2026 at access. http://download.companieshouse.gov.uk/en_output.html. Accessed 29 October 2026.
- EDGAR full-text search frequently asked questions, US Securities and Exchange Commission. https://www.sec.gov/edgar/search/efts-faq.html. Accessed 29 October 2026.
- Organizations using marketing services of data brokers, Information Commissioner's Office, no last updated date shown; the page states it is under review. https://ico.org.uk/for-organizations/direct-marketing-and-privacy-and-electronic-communications/organizations-using-marketing-services-of-data-brokers/. Accessed 29 October 2026.
- Article 14, information to be provided where personal data have not been obtained from the data subject, legislation.gov.uk, The National Archives, latest available revised version. https://www.legislation.gov.uk/eur/2016/679/article/14. Accessed 29 October 2026.
- California license search, California Department of Consumer Affairs. https://search.dca.ca.gov/. Accessed 29 October 2026.
- License data search, Texas Department of Licensing and Regulation. https://www.tdlr.texas.gov/LicenseSearch/. Accessed 29 October 2026.
- Entity information, US General Services Administration, SAM.gov. https://sam.gov/content/entity-information. Accessed 29 October 2026.
- Business Registers Interconnection System dashboard, European Commission, Digital Building Blocks, last updated 2 February 2026. https://ec.europa.eu/digital-building-blocks/sites/spaces/DIGITAL/pages/210798097/Business+Registers+Interconnection+System+dashboard. Accessed 29 October 2026.
- Corporate Number Publication Site, National Tax Agency, Japan. https://www.houjin-bangou.nta.go.jp/en/. Accessed 29 October 2026.
- GrowthPal reporting methodology and counting rules, GrowthPal. https://growthpal.io/glossary/#methodology. Accessed 29 October 2026.
Related
- Cold email agency what happens once the list exists.
- B2B lead generation agency the full program, list research included.
- B2B database reactivation for contacts you already hold rather than new ones.
- Cold email deliverability why a list is verified before it is sent to.
- Glossary and reporting methodology what every figure on this site counts.
- Enterprise outbound who is controller and who is processor for the records above.
Send us the segment you cannot export and we will tell you which source defines it, what that source returns, and what has to be researched by hand.
A plain text version of this page is at /b2b-lead-list-building.md, and the site index for language models is at /llms.txt.