Group Insurance Leads: How to Build a List That Ranks Itself

Published 2026-09-13 · Updated 2026-09-13 · By Benefit Signals · Markdown version

Group insurance leads come from four places: purchased lists, referrals and centers of influence, inbound marketing, and the public filed record. Only the filed record tells you which employer group holds which carrier, who the incumbent broker is, roughly what that broker is paid, and when the policy year ends. Ranking a territory by filed renewal date and by signs the incumbent relationship has come loose gives you a call order. A purchased list gives you names in whatever order the vendor exported them.

Every benefits broker has bought a list at least once. You pay for a spreadsheet of employers in your counties, you sort it by employee count, and you start at the top. Two weeks later you have learned three things: most of those groups renewed four months ago, half the contact names are gone, and every other agent in your market bought the same file from the same vendor. The list was not wrong. It was just unranked -- and an unranked list is a list you have to work alphabetically, which is the same as working it randomly.

This page is about the difference between a name and a lead. A name is an employer that exists. A lead is an employer that exists, is reachable, and has a reason to take the call this quarter. The gap between those two things is where most prospecting hours go to die, and it is a gap you can close with records that are already public.

What actually counts as a group insurance lead

Strip the word "lead" of its vendor packaging and a usable group insurance lead has to answer four questions before you pick up the phone. If a source cannot answer them, it is selling you a directory, not a lead.

• Who holds it now. The incumbent broker or agency of record, by name. If you do not know who you are displacing, you cannot judge whether the seat is defensible.
• What is in force. Carrier, contract, and which coverages are actually on file -- health, dental, vision, life, disability, stop loss.
• When it turns. The policy year end. A group is not persuadable on your calendar; it is persuadable on its own renewal calendar.
• Whether anything has come loose. A broker who has left the filings, a business that changed hands, economics that sit outside the norm for a group that size. Those are the conditions under which an employer entertains a stranger.
• Who signs. A decision-maker you can reach, not a generic info@ address.

The four sources of group insurance leads, compared honestly

None of those four is useless. Referrals close best and always will. The point is that only one of them gives you an ordering principle across a whole territory -- and it is the one most brokers never open, because reading it by hand across thousands of employers is not a job a person can do.

Why a bought list underperforms even when the data is clean

The usual explanation is data quality. Bad emails, dead phone numbers, wrong headcounts. That is real, and it is not the main problem. The main problem is that a bought list is sorted by attributes of the employer -- size, SIC code, ZIP -- and none of those attributes change. An employer with 140 employees in Dayton had 140 employees in Dayton last year too. Sorting by a constant tells you nothing about which employer is movable now.

An employer does not switch brokers because you called. It switches because something changed first, and then someone called. The trigger comes before the conversation. So the useful sort is not "who is biggest" -- it is "where did something just change, and whose contract year is about to close." Those are both properties that move, and both are visible in filings.

The truth is, the reason this ordering is rare has nothing to do with secrecy. The Form 5500 Schedule A is stamped "This Form is Open to Public Inspection" right at the top. It is public because ERISA section 103(a)(2) requires insurers to hand the plan the information and the plan to file it. Anyone can pull one filing. Almost nobody reads every filing in a state and sorts the result.

Source: 2024 Schedule A (Form 5500), U.S. Department of Labor (as published for plan year 2024).

What the filed record gives you that a list does not

Schedule A is filed per insurance contract, not per employer, and it names things a lead list never will. Line 1(a) is the name of the insurance carrier. Line 1(d) is the contract or identification number. Lines 1(f) and 1(g) are the policy or contract year, from and to -- that is the renewal date, in the record, in the employer's own filing. Line 3(a) is the name and address of the agent, broker, or other person to whom commissions or fees were paid, listed in descending order of amount. Part III, line 8 is the benefit and contract type: health, dental, vision, life insurance, temporary disability, long-term disability, prescription drug, stop loss, HMO, PPO, indemnity.

Put plainly: the incumbent broker, what the incumbent is paid, the carrier, the coverages, and the date the clock runs out -- all of it, on a public form, for every group large enough to file one. We walk through the form field by field in our Schedule A guide, and the practical lookup procedure is in how to find a company's insurance carrier.

Group benefits leads and group life insurance leads are the same record

Brokers search for "group benefits leads," "group health insurance leads," and "group life insurance leads" as if they were three different data problems. On the filed record they are one problem viewed through three checkboxes. Part III line 8 of Schedule A asks the carrier to check every applicable box on a single contract, so a group that files health, dental, vision, and life shows up in all four universes with the same EIN, the same broker of record line, and the same policy year dates.

That has a practical consequence worth sitting with. If you write life and disability, the groups you want are not a separate market -- they are a filter on the same market, and the ancillary lines frequently renew on the same contract year as the medical. A group that is in play at all is usually in play across its whole benefits stack. Filtering by coverage type is a targeting decision, not a data-sourcing decision.

Two questions a ranked list has to answer: who, and when

Benefit Signals runs on two mechanisms, and they do different jobs. Change signals tell you who. The Renewal Clock tells you when. Neither one is useful alone -- a group that is loose but eleven months from renewal is a relationship to start, not a proposal to write, and a group renewing in six weeks with a happy incumbent is a courtesy call.

The change signals fall into three categories. We name the categories, not the detectors, for the obvious commercial reason: a published detector list is a detector list a competitor ships next quarter.

• The incumbent broker's position changed. The agent of record is no longer where they were in the filings.
• The business itself is in motion. Sold, renamed, or newly under different ownership.
• The economics sit out of line for a group that size. Premium or compensation well outside the norm for comparable groups -- the kind of gap a CFO notices.

The Renewal Clock is the simpler half and the one brokers underrate. Filed policy periods are exact dates, not quarters. Sorting a territory by whose contract year closes next puts your week-by-week call order on the record rather than on a guess, and it kills the habit of treating January 1 as the whole market. Plenty of groups do not renew on the calendar year, and those are the ones your competitors are not calling.

Does the ranking actually rank? Here is the test it had to pass

A ranking claim is worth nothing without a grading method, so here is ours in plain terms. The model was trained on filings from 2016 through 2022 and then pointed at a 2023 filing cohort held out from model fitting. A group counts in that test only if it has a top-compensated broker seat on file and a next-year filing to check the outcome against -- so the denominator is the observable cohort, not every employer in the country.

Likelihoods, not certainties — how the model was tested is published on the home page. "Winnable" is itself a residual -- a transition counts only after excluding co-broker reorders and bulk transfers consistent with book sales or rebrands, because employer intent is not in the filings and we are not going to pretend to read it. What a ranking buys you is a better call order, not a promise. If a prospecting vendor tells you otherwise, ask them for their test results; how the model was tested is published on our home page.

How to work a ranked territory in the first two weeks

• Set your territory once -- ZIP and radius, or a set of counties -- and stop re-drawing it. Depth beats breadth in a market where the incumbent has a ten-year head start.
• Sort by renewal date first, then read the signals inside the next two quarters. Anything renewing inside 30 days is usually already committed; the workable window sits roughly 90 to 180 days out.
• Before you dial, read the group's own filed record: carrier, coverages, the broker of record line, and what that broker was paid. Your opener should be about their plan, not your agency.
• Treat a live change signal as a reason to move a group up the list, not as evidence that the employer is unhappy. It is a condition, not a complaint.
• Log outcomes against the signal that fired. Within two renewal cycles you will know which category converts in your market, which is knowledge no vendor can sell you.
• If a group says yes, know the mechanics before you ask -- see broker of record letter.

Where bought lists still make sense

Two places. First, below the filing threshold: employers too small to file a group-health Form 5500 do not appear in the Schedule A universe at all, and a list vendor may be the only way to reach them. That is a real gap, and it is why we also carry 831,896 small employers that file a retirement plan and have no filed group-health plan -- fewer than 100 retirement-plan participants apiece. Some of them carry coverage that is not required to be filed; others may not offer it yet. Either way they are invisible in the filed group-health universe most prospecting tools start from.

Second, for pure contact enrichment. If you already know which groups you want and you need a mobile number, a contact vendor is a contact vendor. Just do not confuse buying a phone number with buying a reason to use it. The full comparison of the tools in this category, ours included, is in insurance prospecting tools.

What this approach costs

Benefit Signals is $60 a month, or $600 a year. The monthly plan comes with a seven-day free trial: card at checkout, first charge on day 8, cancel before then and pay nothing. The annual plan has no trial and is billed in full the day you start. Either way the subscription renews automatically until you cancel. For context on where that sits in the category, two of the better-known platforms in this space publish no dollar pricing at all -- see our pages on BenefitFlow and miEdge.