Every enrichment tool promises some version of the same thing: give us a name and a company, and we'll give you their email and phone number. Some claim 95% accuracy. Others say 90%. A few say "industry-leading" without putting a number on it.
What none of them explain clearly is where that data actually comes from, how it gets into their database, and what "95% accuracy" means when you look at the methodology behind the claim.
After digging into how a few of these tools work under the hood, the process is more straightforward than most vendors let on.
The five main data sources
Enrichment databases are built from a mix of sources. The proportions vary by provider, but almost every database draws from the same five categories.
The first is web scraping. Automated bots crawl company websites, staff pages, press releases, conference speaker lists, and public directories. If a company has a "/team" or "/about" page listing employees with their titles, that data gets scraped and indexed. The coverage is decent for companies that maintain public-facing team pages, and useless for companies that don't. Scraping also picks up data from job boards, where candidates list their work history publicly.
The second is data partnerships. Enrichment providers buy data from other companies that collect contact information as part of their business. Business credit agencies, event registration platforms, trade associations, and B2B media companies all have contact databases they license to third parties. These partnerships are why some tools have surprisingly good coverage of specific industries or regions. If a provider has a partnership with a European trade association, their European data will be better than a competitor who doesn't have that deal.
The third is user-contributed data. Some tools operate on a give-to-get model. You install the extension, and in exchange for free credits, the tool collects contact data from your email headers, calendar invites, or email signatures. Every email you send or receive with a business signature becomes a data point. This is how tools like Lusha and Apollo built their early databases. The coverage is concentrated wherever the tool's user base is largest, which usually means US tech and sales teams.
The fourth is social media and professional networks. LinkedIn profiles are the most obvious source, but enrichment providers can't just scrape LinkedIn directly without getting blocked. Instead, they use a combination of public profile data, third-party data aggregators who have licensing deals, and cached versions of profiles. Some providers also pull from GitHub, Twitter/X bios, personal websites linked from social profiles, and public posts that mention work affiliations.
The fifth is public records and registries. Government business registrations, SEC filings, patent databases, court records, and professional licensing boards all contain contact information. This source is strongest for executives at registered businesses and licensed professionals (lawyers, doctors, CPAs). It's weak for individual contributors and anyone at a company that doesn't file public paperwork.
How data gets matched to a person
Having raw data is one thing. Turning it into a reliable "Name + Company = Email + Phone" lookup is a different problem.
When you search for "Jane Smith at Acme Corp," the enrichment tool isn't looking up a single record. It's running a matching algorithm across multiple data points to find the most likely correct result.
The tool starts by looking for exact matches: someone named Jane Smith who is currently listed as working at Acme Corp in the database. If it finds one, that's the easy case. But people change jobs, companies get acquired, and names aren't unique. There might be three Jane Smiths who have worked at Acme Corp at different times.
To pick the right one, the tool uses recency signals. Which data source was updated most recently? Does the LinkedIn profile still show Acme Corp as the current employer? Is the email domain still active? Has anyone sent or received email from this address in the last 90 days?
Some tools also use email pattern matching. If Acme Corp uses the format firstname.lastname@acme.com, and the tool knows this pattern from other Acme employees in its database, it can generate jane.smith@acme.com as a probable email even without a confirmed data point. It then verifies whether that address accepts mail (more on verification below).
The matching step is where a lot of errors enter the system. If the recency data is stale, the tool might return an email from Jane's previous employer. If two Jane Smiths worked at Acme, it might return the wrong one's phone number. If the email pattern has changed (Acme switched from firstname.lastname to first initial + lastname), the generated email will bounce.
What "verified email" actually means
Most enrichment tools label their results as "verified." This sounds like a strong guarantee, but the verification process is more limited than you'd expect.
Email verification typically works by connecting to the recipient's mail server (via SMTP) and asking "does this mailbox exist?" without actually sending an email. The server responds with either a "yes, this address exists" or "no, unknown user." If the server says yes, the email gets labeled as verified.
The problem is that this only confirms the address exists. It doesn't confirm that it belongs to the person you're looking for. It doesn't confirm that the person still works at that company. It doesn't confirm that the inbox is monitored. An employee who left six months ago might still have an active mailbox that the company hasn't deactivated.
There's another wrinkle: catch-all domains. Some companies configure their mail servers to accept email sent to any address at their domain, whether or not the specific mailbox exists. If Acme Corp is a catch-all domain, then any.random.string@acme.com will pass verification. The enrichment tool will mark it as "verified" even though it's just hitting a catch-all. The email might get delivered to a general inbox that nobody checks, or it might get silently dropped.
When a tool claims "95% verified emails," they mean 95% of the addresses they returned passed the SMTP check. They don't mean 95% of those emails will reach the right person at their current job. Those are very different numbers.
Where the "95% accuracy" claim comes from
Accuracy claims in enrichment are almost always self-reported, and the methodology behind them is rarely disclosed.
Here's how it typically works: the provider runs their tool on a sample of contacts, checks how many emails pass SMTP verification, and reports that number as their accuracy rate. The sample is usually curated to favor the tool's strengths. If the tool is strong on US tech contacts, the sample will be heavy on US tech contacts.
What the accuracy number doesn't account for: contacts where the tool returned no result at all (those aren't counted as "inaccurate," they're just excluded from the denominator), contacts where the email is technically valid but belongs to a previous employer, contacts where the phone number is in the wrong country, and contacts in geographies or industries where the tool has weak coverage.
A tool could have a 95% accuracy rate on the contacts it finds, while only finding 50% of the contacts you need. The accuracy rate tells you about data quality. It tells you nothing about coverage.
When I ran my 25-profile test, the tool with the widest coverage (20+ providers) found emails for 76% of profiles. A competing tool found 44%. Both tools' accuracy rates on the emails they did return were probably similar. But the 44% tool simply didn't have data for more than half the list. If you only measure accuracy on found results, both tools look good. If you measure on your full prospect list, they look very different.
Why databases go stale
Enrichment data has a shelf life. People change jobs, companies get acquired, domains expire, and phone numbers get reassigned. The average tenure at a US company is about 4 years, which means roughly 25% of your database is changing employers every year.
Providers try to keep their data fresh through re-scraping, monitoring LinkedIn profile changes, tracking email bounces, and processing new user-contributed data. But there's always a lag. A person might change jobs today and the enrichment database might not reflect it for weeks or months.
The staleness problem is worse for some data types than others. Work email addresses go stale immediately when someone leaves a company (though the mailbox might stay active for a while). Phone numbers are stickier since people keep their cell numbers across jobs. LinkedIn profiles are usually updated within a few weeks of a job change, but that update has to propagate through all the downstream data partnerships before enrichment tools see it.
This is why tools that check multiple sources tend to have fresher data. If Provider A hasn't updated yet but Provider B scraped the person's new company website last week, the multi-source tool catches the change faster.
What this means for how you use enrichment tools
Understanding where the data comes from changes how you should evaluate and use these tools.
Don't take accuracy claims at face value. Ask the vendor what their methodology is. If they can't explain how they measure accuracy, or if the answer is just "we verify emails via SMTP," that tells you the number is less meaningful than it sounds. The metric that matters is coverage on your specific prospect list, which is something you can only measure by testing.
Know your own prospect profile. If you sell to US tech companies with 200+ employees, almost any tool will have decent data. If you sell to European SMBs or niche industries, you need a tool that either specializes in those segments or checks enough sources to cover the gaps. Asking "how many data providers do you query?" is a more useful question than "what's your accuracy rate?"
Plan for data decay. Any contact data you collect today has a half-life. Build re-verification into your workflow. If you're running outreach on a list that's more than 3 months old, re-enrich before sending. The bounce rate will tell you how much has changed.
Treat enrichment output as a starting point, not a guarantee. A "verified" email still needs to be validated against bounces after sending. A phone number still needs to be checked for country code. If your workflow assumes enrichment data is 100% correct, every error becomes a wasted touchpoint.
ShareCo SalesSync queries 20+ enrichment providers per lookup via waterfall, which is one approach to the coverage problem described above. You can try it free on the Chrome Web Store, or pick whatever tool fits your prospect profile and test it yourself.