The Prospect Research Source Reference: What Each Source Proves

Before you reach out, one question decides whether the research was worth doing: which public source will actually confirm the fact your opener depends on, and how far can you trust it? This reference pairs every account-level and buyer-level fact a rep needs before first contact with the source that proves it, how fresh that source is, and the specific way it misleads. It is built for founders selling their own product, account executives, SDR leads, and partnerships teams who want to jump to one row and leave.

Most pre-call checklists list sources and stop there. They tell you to check LinkedIn, company news, and job posts, but never say, source by source, what each one can confirm versus how it quietly lies through stale titles, ghost posts, and planted press. That gap is expensive. In Refolk's index of professional profiles, the buyer you need to reach is far scarcer than the seller reaching out, so a mis-targeted title wastes disproportionate effort. This document closes the gap.

Why source selection matters before the first message

Picking the right source is the difference between an opener built on a verified, dated fact and one built on a title the buyer abandoned months ago. Research separates the top of the field from everyone else, and the separation is large.

According to a LinkedIn State of Sales figure, 82% of top performing salespeople say they always research before contacting prospects, against 49% for other sellers. Separately, 76% of top reps research before a cold call. The volume of buyers to reach is small relative to the sellers reaching them, which raises the cost of every wrong guess.

The pyramid is the point. A rep is common; the person a rep must reach is scarce. When you mis-target the buyer title, you do not waste effort at a 1:1 rate, you waste it against a pool 16x thinner. That is the mechanism behind spending a few disciplined minutes on the right source rather than skimming three wrong ones.

Time spent varies by how you measure it. Manual research runs 1 to 3 hours per account, while a single contact takes 4 to 7 minutes. Field data shows reps spend about 9% of their time on prospecting research before making a single contact. The goal here is not more time; it is spending the time you have on the source that actually proves the fact.

The source reliability table: freshness against distortion

Here is the spine of the reference. Each public source has a freshness cadence and a documented distortion. Read both columns before you trust a cell.

The counterintuitive lesson is that freshness and trust move in opposite directions. The fastest hard source, the 8-K, is also the most verifiable, because a regulatory penalty enforces its accuracy. The freshest buyer signal, a self-edited profile, is the least reliable, because nothing enforces its accuracy at all.

Treat the review-site row with the most care. Verification reshapes the distribution more than the average. Verified Blind shows 22.0% five-star reviews against Glassdoor's 37.7%, yet the mean ratings differ by only 0.15 points. A give-to-get incentive reduces the likelihood of extreme 1-star and 5-star reviews by 3.6 and 2.1 percentage points respectively. Selection bias concentrates at the extremes, so read the shape of the ratings, not the headline star average.

What each source proves, source by source

Each source confirms a narrow band of facts well and everything else poorly. Match the fact you need to the source built to hold it.

Professional profiles: current role and public voice

The professional profile is the primary buyer-level source. It confirms current title, tenure, and, through linked activity, the buyer's public voice. Beyond the profile itself, review personal websites, podcast appearances, speaking engagements, and industry awards to confirm what a person actually cares about. What the profile does not reliably confirm is a past claim, because past roles and skills are the fields most often edited retroactively.

Regulatory filings: dated corporate events

An SEC 8-K discloses material corporate events like mergers, leadership changes, and bankruptcy within four business days, filed to EDGAR. It is the strongest confirmation of a dated, account-level trigger because it is fast, timestamped, and legally enforced. An auditor's restatement letter must be reported within two business days. What it does not prove is completeness: the absence of a filing is not proof that nothing happened.

Job postings: hiring and growth intent

A careers-page posting signals hiring and growth intent, which is useful for timing an outreach around a team's expansion. It confirms intent only when the post is genuinely active. Post age is your trust dial: with an average fill of 41 days, a posting past 30 days with no visible activity has crossed from signal into likely noise.

Company news and press releases: claims, not validation

Company news confirms what the company wants said and when it wanted to say it. A press release carries the sender's message verbatim. A self-issued wire release is a claim the sender controls end to end, not third-party validation. Major paid wires reach thousands of sites, but much of that placement lands on low-value aggregators, so raw link count tells you nothing about credibility.

The CRM: what already happened

Your CRM confirms prior touches, inbound path, and what the buyer said last time. Deal history is the highest-value research you can do, because what the buyer told you in a previous conversation outranks anything on their profile. Its failure mode is decay, covered in the next tables.

The account-versus-buyer split

Divide facts into two buckets and each maps cleanly to its proving source. Account-level facts resolve to filings, news, and postings; buyer-level facts resolve to the profile and the person's public voice.

The scarcity of the buyer bucket is why title precision matters. Refolk's index shows the buyer pool is thin, and it is not uniform across geographies.

The US and UK leadership pools also differ by more than market size explains.

The ratio is 2.70x, driven partly by title inflation and differing "Head of" conventions by geography. A US-calibrated title filter under-returns qualified UK buyers, so widen your title synonyms per country before you conclude a market is thin.

That query pairs a scarce buyer title with the most trustworthy account signal in one pass. Instead of confirming the title on one tab and hunting the filing on another, Refolk returns people already joined to a dated 8-K event, which is exactly the load-bearing combination step 3 and step 4 of the procedure ask you to assemble by hand.

The pre-call research procedure

Run these six steps in order. The whole sequence fits inside the time a disciplined rep already spends, and it ends with one verified opening line rather than a folder of unverified tabs.

The 3x3 method is the guardrail against the opposite failure, over-research. Top reps reportedly spend 6x more research time than average reps, but the return is diminishing once you have three verified facts, so cap the effort deliberately.

How pre-call research goes wrong

Every source has a documented failure mode, and each produces a specific false positive. This is the most important section in the reference, because a source that misleads you confidently is worse than no source at all.

Stale profile title. The buyer left the role months ago and the opener names a dead job. The false positive is a current-looking title that was simply never updated. Check by cross-referencing the profile against the company's team or careers page, or a recent 8-K.

Time-traveled profile. Skills and titles are retro-added for the market, not the truth. Nearly 1 in 5 established U.S. LinkedIn users, 19.7%, retroactively edited a past role between 2020 and 2026. Check by matching claims to dated posts or press mentions, not the profile's self-description.

Ghost or evergreen job post. A posting read as an active buying trigger when it is pipeline-building. Between 18 and 22% of quarterly posts qualify as ghost jobs. Check by confirming the same role in the company's ATS and reading its post age against the 41-day fill benchmark.

Planted PR. A self-issued wire release read as third-party validation. The message is fully controlled and often lands only on aggregators. Check for independent, bylined coverage before trusting the claim.

Review-site polarization. A bimodal 1-star and 5-star split misread as company-wide sentiment. Check by reading volume and modal ratings, discounting firms with few reviews, and comparing against a verified-review source.

Filing materiality gap. Absence of an 8-K does not mean nothing happened, because materiality is a judgment call. Check by pairing filings with news and never treating "no filing" as "no event."

CRM decay. Building sequences on contacts who moved 6 to 12 months ago, a hidden failure rate that only surfaces when outreach fails at scale. Check by re-verifying title and employer at the time of outreach, not at list-build.

When to trust freshness versus verifiability

Use this judgment call whenever a source is fresh but weak, or slow but strong. The two variables that matter are how recent the signal is and how enforceable its accuracy is.

The top-right quadrant is where you want your opener's load-bearing fact to come from: fresh and verifiable, like a recent leadership-change 8-K. The top-left is the trap: fresh, attention-grabbing, and unenforced. A profile edited yesterday feels current, but 19.7% of edits rewrite the past. Move any top-left fact into the top-right by confirming it against a second, enforceable source before you write the line.

Keep the reference current

Sources drift, so treat this as a living lookup and re-check the mechanisms, not the values. The values in the distortion column change with each new study; the mechanisms do not.

Re-verify contact currency at outreach, never at list-build, because the 6-to-12-month role change is the decay window that only appears at scale. Re-read job-post age against the current fill benchmark rather than a fixed number, since the 41-day average is a moving figure. And when a source you rely on publishes a new distortion figure, update the row and keep the check that goes with it.

Fact needed | Primary source | Second source | Freshness check | Verdict
Current title | Profile | Company team page | Edited in last 90 days? | Confirm / discard
Account trigger | 8-K or news | Independent coverage | Filed within 4 business days? | Confirm / discard
Hiring signal | Job post | Company ATS | Under 30 days old? | Confirm / discard
Prior relationship | CRM | Last thread date | Role changed 6-12 mo ago? | Confirm / discard

The reference earns its place when you stop reading it start to finish and start jumping to a single row mid-task. Find the fact you need, read its proving source, read how fresh it is, and read the one way it lies. Then confirm it against a second source and write the line.