A fresh sample for every experiment

35 million companies you can search by what they do, what their website runs and what just changed. Describe the segment in your hypothesis, and email only companies that have never heard from you.

  • ✓ Previously contacted companies excluded
  • ✓ No credit card required
A segment defined by industry, country, website footprint and a hiring signal, returning matching companies SEGMENT IndustryDental Practice Confidence≥ 0.8 CountryGermany Website hasMeeting booking Website lacksCookie consent Signalhiring Sales, last 30 days Never contacted Dental Practice hiring Dental Practice hiring Dental Practice hiring

35 million

companies you can search today

365,268

companies added in the last 7 days

226,630

signals observed in the last 7 days — hiring, filings, launches, website changes

118

industry tags, assigned from each company's own website

Describe the segment exactly

A segment where everyone shares the same problem replies more often, and a higher reply rate reaches a verdict with fewer emails.

Company profiles

  • ✓ Search by the company's own description of what it does
  • ✓ Industry tags with a confidence score, and official industry codes
  • ✓ Size, founding year, and independent companies apart from branches
  • ✓ Geographic data — country, state, city
  • ✓ Website footprint and dated business signals
  • ✓ General company contacts & socials
US software companies with 2–20 employees whose description says “our platform helps”.

The people you'd email

  • ✓ Key professional details & job title
  • ✓ Current & past employment history
  • ✓ Work email & phone with verification status
  • ✓ Professional social profiles
At each of those companies, the founder or CEO, with an email that passes verification.

Industries read from the company's own words

A registry code says what a company filed as, sometimes decades ago. Sellvance reads the company's homepage and assigns industry tags from a taxonomy of 118, each with a confidence score.

  • Choose how sure you need to be. Search uses tags at 0.8 confidence or higher, and you can raise the bar for a stricter sample.
  • Official codes still work. Every tag maps to ISIC, NAICS and NACE. Search by a code directly, and it matches every class beneath it.
  • Say what to do with the unknown. Include or leave out companies not classified yet, instead of guessing for them.
A company's own homepage text is classified into industry tags with a confidence score, and each tag maps to ISIC, NAICS and NACE codes HOMEPAGE TEXT Family dental care in Leipzig since 1998. Implants, check-ups and orthodontics for the whole family. INDUSTRY TAGS CONFIDENCE Dental Practice 0.94 Medical Practice & Clinic 0.41 Allied Health & Therapy 0.12 search threshold 0.8 STATISTICAL CODES ISIC Rev. 5 8620 NAICS 2022 621210 NACE 2.1 86.23

What their website says about how they work

Company websites are re-read on a rolling cycle. Each visit records the tools and settings the page and its DNS give away, and every one of them is a filter: who has it, and who doesn't.

  • ✓ Hiring — careers page, open roles, the ATS in use, the role families they hire for
  • ✓ Marketing & sales stack — analytics, tag manager, ad tags, chat, meeting booking, marketing automation
  • ✓ Website — CMS, ecommerce platform, languages offered, API docs, reviews
  • ✓ Email — mail provider and DMARC policy, read from DNS
  • ✓ Compliance & trust — cookie consent, privacy policy, certifications claimed

“Lacks” only matches sites we have actually read, so a site we never reached doesn't pass as a gap.

Clinics that take bookings online but run no cookie consent tool.
What a company's homepage and DNS reveal: CMS, tag manager, meeting booking, chat widget, cookie consent, mail provider and DMARC policy example.com Book a call Accept CMS WordPress Tag manager Google Tag Manager Meeting booking Calendly Chat widget Intercom Cookie consent Cookiebot DNS Mail providerGoogle Workspace DMARCReject

Reach them when something changes

A company that just started hiring salespeople or changed how its site works has a reason to talk now. Sellvance keeps these as dated signals, and you filter by how recent they are.

Business signals

Hiring, register filings, product launches and website changes, each dated by its source. Ask for the last week, the last month or any window you need.

Website changes

Search for what appeared on a site, what disappeared from it, or a value that changed, such as a DMARC policy moving to reject. A tool counts as gone only after two visits in a row miss it.

Current state, not just events

Combine both: companies that added a chat widget this month and still have no meeting booking tool.

Website parameters read at every weekly crawl: a cookie consent tool appears, a chat widget disappears after two missed crawls, the DMARC policy changes PARAMETER WK 1 WK 2 WK 3 WK 4 WK 5 Cookie consent Usercentrics appeared Chat widget tawk.to missed disappeared DMARC policy from DNS none none reject reject reject changed The first reading is a baseline, not a change. Gone means missing at two crawls in a row, so a slow page is not news.

A sample that refills itself, without repeats

Save the search that describes your segment once. Sellvance pulls new matching companies as the experiment needs them, and never the same company twice.

1

Save the segment

The search from your hypothesis, kept as it is for the whole experiment.

2

Exclude everyone you've contacted

Point the rejected list at All leads and past experiments stay out.

3

Let the campaign set the pace

Autopilot pulls only as many companies as the campaign can email.

A saved search feeds new matching companies into a segment, skipping every company already contacted SAVED SEARCH new matches SKIPPED Already in your leads SEGMENT no company twice Autopilot pulls only as many companies as the campaign can email.

From search to experiment in one place

The database is where an experiment starts. Every company flows straight into the rest of the platform.

  • Add to a segment in one click. Or let a saved search do it continuously.
  • Qualify before anyone is emailed. AI enrichment checks each company against your hypothesis.
  • Send and get a verdict. Qualified companies move into the experiment's campaign.
Explore outreach
Search, qualify and send: companies go from a search to a qualified segment to a campaign that returns a verdict SEARCH Add to segment QUALIFY AI checks each one CAMPAIGN reply VERDICT above threshold 1 · FIND 2 · QUALIFY 3 · SEND AND DECIDE

Why a precise sample gives a better answer

Targeting isn't about sending more. It's what makes the result readable.

One problem per segment

A broad list averages groups that want different things. A narrow one tells you about the group you actually described.

Fewer emails to a verdict

Against a 3% threshold, proving a true 10% takes about 66 sends. Proving a true 6% takes about 282. The sharper the segment and the offer, the cheaper the answer.

Nobody hears from you twice

A company answering your second email tells you nothing about your new offer. Excluding past contacts keeps every result clean.

Test your first segment

New accounts start with 2,000 free connects, enough to search the database and try a first sample.

Create your free account

Start searching the database in under a minute.

Application is loading. Please wait…