How do I know a lead is real

Two different claims hide inside the question. One is that a real web session produced this lead. The other is that a real person wants a loan. Almost everything attached to a lead answers the first, cheaply and well, and nothing attached to it answers the second.

What the question is really asking

Separating the two claims answers most of the question. Evidence that a session happened is plentiful: a time, a page, a device, a consent record. Evidence about the person behind it barely exists at the point of purchase, and no vendor can hand it over with the lead, because it is not the kind of thing a web form collects.

That is not an accusation about vendors. It is the shape of the data. A buyer who knows which claim a given artifact supports stops over-reading the strong-looking ones and stops dismissing the weak-looking ones for the wrong reason.

The artifacts behind one lead, and what each one proves

Everything that arrives with a lead falls into three tiers. Some items evidence a session: a moment, a page, a device. Some evidence a live channel: a number that rings, a mailbox that accepts mail. A few evidence an act by whoever held that channel. None of the three reaches identity, and reading a lower tier as a higher one is where the money goes.

The certificate is the item most often read as more than it is. It is a record of the web session, held at a URL on the issuer's server, listing a time stamp, the page, and the consent language on screen. What it is, how it is created and who holds it afterwards is set out in TrustedForm certificate.

Source: ActiveProspect developer documentation and TrustedForm product page, read 20 September 2026, for what a certificate records and for the certificate URL. Availability is bounded and time-sensitive and no duration is stated here, because the window depends on the account and on when the retain request is made. Read the figure for a given certificate from the issuer's own response, never from a page like this one.

Checks a buyer can run without taking anyone's word

Most of these checks cost nothing and run in under a minute, and each one is worth doing on arrival rather than during an argument three weeks later. A buyer who runs them holds facts. A buyer who asks afterwards holds an assurance, and an assurance is the one thing every seller has in stock.

• Open the consent link before you need it. It resolves or it does not. Delay changes that answer, and no amount of arguing changes it back.
• Compare the timestamp to the moment it reached your system. A long gap describes the route the lead travelled, not the person at the end of it.
• Read the lead against itself. A ZIP code in one state and a property in another is a contradiction visible without asking anybody anything.
• Ask which channel was confirmed, and how. A number confirmed by a code says nothing about the email address sitting next to it. One check covers one field.
• Ask for the page as it was, not as it is. Today's version of a form is not evidence of what was on screen then.
• Dial before you judge. A lookup and a dial agree on exactly one thing, which is whether the line is live.

What none of this evidence proves

None of it reaches the two things a buyer wants: that the person is who the lead says, and that they want a loan. Identity is not in the file. Intent is a judgement made on a call. A real session by a real person with no appetite for a mortgage produces a flawless set of artifacts and a dead pipeline, and every check above passes.

The phone is the most over-read item in the pile. Reachability proves the line is live. The safe harbour in the federal calling rules covers a caller who queried the reassigned numbers database, received the answer "no", and reached a number that had in fact changed hands, which is to say it operates only where the database was wrong. It settles a question about liability. It says nothing about who is holding the handset.

Source: 47 CFR 64.1200(m), full text read 20 September 2026 at law.cornell.edu. It requires both that the caller queried the Administrator's database and received a response of "no" verifying the number has not been permanently disconnected since consent was obtained, and that the call "was the result of the database erroneously returning a response of 'no'". It is a liability provision and supports nothing about identity, intent, or who answers a call. The database itself is the FCC's Reassigned Numbers Database.

There is also no honest figure for how many purchased leads are fabricated. Every rate in circulation was published by a company selling detection against it, which makes each one a sales document rather than a finding, and treating a vendor's estimate of its own market as a benchmark is how the last several lead-buying myths got started.

Searched 20 September 2026 across regulator publications, academic work and trade coverage for a rate of fabricated or automated lead form submissions. Every figure found, including the widely repeated one-in-four, originates with a fraud-detection vendor's own report on its own traffic. No independent study puts a number on it, so none is printed here.

When the record and the person disagree

A complete record and a person denying all of it is the ordinary case rather than the strange one, and the artifacts do not settle it. What the person says is a statement about memory. What the file holds is a statement about a session. Both can be accurate in the same conversation, which is why the argument goes nowhere.

The causes behind that denial are their own subject, worked through in why leads say they never filled out a form. Whether the disagreement makes the lead disputable at all is a separate line again, drawn in what counts as a bad mortgage lead. This page stops where the evidence does, which is earlier than most people assume.