An agency can open a campaign report full of form fills and still write almost nothing that month. Cost per lead looks acceptable, producers make the calls, and the bind rate stays flat. The useful question is not whether the ads work. It is which stage broke, because the fix for a bad audience and the fix for a slow callback have nothing in common.
What follows is a diagnostic order built around how a book of business actually converts, from click to bound policy. A public video on interest based targeting by Wilco de Kreij supplies the platform side of it, quoted directly and linked to the exact moment. Which parts are his and which are ours is marked throughout.
Why do insurance ads produce quote requests that never turn into policies?
Insurance ads produce quote requests that never bind when the audience holds people interested in the subject rather than people in a buying moment. A form fill costs the same from a shopper or a browser. Bind rate, not cost per lead, is the number that separates the two.
This line of business has an economic shape that hides the problem:
- The lead is cheap and the policy is valuable, so a mediocre cost per lead can still look fine on the report.
- Commission arrives after the bind, and renewal commission arrives at the next renewal term, which may be months out. The feedback loop is slow by construction.
- A mismatched audience can fill a pipeline with people who cannot legally buy from you, starting with anyone in a state where you hold no license.
Write down three numbers you already have before changing anything: cost per lead, lead to quote rate, and quote to bind rate. If lead to quote is healthy and quote to bind collapses, targeting is not your problem. If lead to quote collapses, it probably is.
How do I tell if the problem is my targeting or my offer?
Split the path into stages and find the first one that breaks. Impressions to clicks tests the creative. Clicks to form completions tests the landing page. Form completions to contacted tests follow up speed. Contacted to bound tests the offer. Targeting is the suspect only when the drop starts at the top.
| Stage | What a break here means | First thing to change |
|---|---|---|
| Impressions to clicks | Wrong audience, or the wrong message for that audience | Audience, then creative |
| Clicks to form completions | The page asks for too much, or the promise changed between ad and page | Form length, page copy |
| Form completions to contacted | Speed to contact, bad phone data, no routing rule | Follow up process |
| Contacted to quoted | Prospect not qualified, wrong state, wrong age band | Audience filters |
| Quoted to bound | Premium, carrier appetite, coverage fit | Offer and carrier mix |
This table is ours. It is not in the source video.
What should an agency check first when its ads stop producing policies?
Check the mechanical causes before the strategic ones. Confirm the licensing footprint against the geography, the age range against the product, whether the interests you selected still appear in Ads Manager, and whether a licensed producer reaches the lead quickly. Work down that list before rewriting the creative.
- Geography against licensing. Confirm every active ad set is limited to states where you can write.
- Age and eligibility filters. Age gated products fail quietly when an ad set inherits a default range nobody reviewed.
- Interest availability. Detailed targeting options are not permanent. Re-check that the interests carrying an older campaign still appear in the picker.
- Category restrictions. Check the platform's own advertising policies inside Ads Manager to see whether your line of business falls under a restricted or special ad category, because that changes which targeting controls you are allowed to use. Do not assume, and do not take our word for it.
- Speed to contact. Time the gap between form submission and first dial, then compare your own bind rate for fast contacts against slow ones. If the slow group binds worse, fix routing before touching the audience.
- Only then, interest quality. Break results down per interest and cut the losers.
Why does targeting a big carrier brand bring the wrong audience?
Fame dilutes intent. A household carrier name reaches people who recognize the brand for reasons unrelated to buying coverage, including its own policyholders and job seekers. Narrower targets, such as a regional carrier, a trade association or a state agents group, sit closer to the market an agency can actually write.
De Kreij states the principle in general terms: "the more famous they are, the chances are the audience is not as targeted" (Wilco de Kreij, 3:11). His examples are golf and Tiger Woods. The application below is ours.
- A national carrier brand pulls in that carrier's current policyholders, people who remember the mascot, and job seekers.
- A broad topic such as coverage or claims pulls in students, licensing candidates, and other agents studying the competition.
- A regional carrier, an industry publication, or a local business association pulls a smaller and more relevant pool.
He makes a related point about crowded targets: "If you go just for the obvious ones, then you're basically fighting against everyone else on the Facebook ad platform" (Wilco de Kreij, 3:11). For an agency, the obvious targets are the ones every national advertiser with a real budget is already bidding on.
There is a layer underneath this worth understanding before you blame the audience. In the video's account, an interest on the platform is not a page's fan list and is assembled from more signals than page likes: "Facebook makes the interest way bigger than that even to people who have never liked a certain Facebook page in particular" (Wilco de Kreij, 1:36). Read that way, selecting a carrier interest does not reach that carrier's customers. It reaches a larger inferred group.
How does layering interests change who sees an insurance ad?
Layering targets only the overlapping segment, so the ad reaches people who sit inside two or more interests at once instead of any one of them. For an agency, that moves a broad topic audience closer to a life event audience, which is nearer to the moment coverage decisions get made.
The video describes the mechanic with a consumer example: "I only want to show my Facebook ads to people who are inside of the Coca-Cola interest as well as the Pepsi interest." (Wilco de Kreij, 7:53).
Translated to a book of business, the overlap is the signal:
| Line of business | Single interest, weak signal | Layered overlap, stronger signal |
|---|---|---|
| Personal auto | Car brands | Car brand plus a moving or relocation related interest |
| Homeowners | Home improvement | Home improvement plus mortgage or moving interests |
| Term life | Parenting | Parenting plus financial planning topics |
| Commercial general liability | Small business | Small business plus trade specific associations |
| Medicare related plans | Retirement | Retirement plus the age band the product requires |
Treat each row as a pattern, not as a promise that the option exists. Detailed targeting choices change, so check every one against what your own account currently offers. Layering also shrinks reach fast, so run it against a broad control audience rather than replacing everything at once, and watch whether the smaller ad set gets enough delivery to produce a readable result.
What does testing one interest at a time cost an agency?
In the video's account of Ads Manager, each interest needs its own ad set, and each ad set needs its own daily budget. Thirty interests at five dollars a day each runs one hundred fifty dollars a day before a single policy binds, and licensing caps how fast an agency can gather that data.
The arithmetic is his: "even if it's just $5 a day * 30 is still $150 a day just for trying it out" (Wilco de Kreij, 7:07).
Two things make that bill land harder on an agency than on a retailer:
- The conversion event is slow. A bound policy can sit weeks behind the click, so each test ad set needs a longer observation window than an ecommerce test does.
- Geography is capped by licensing. You cannot spread a test across the whole country to reach volume faster. Every ad set runs inside the states where you are licensed and appointed.
Stuffing every interest into one ad set to hold the budget down creates the opposite problem. De Kreij puts it this way: "Some may be profitable, some may be losing money." (Wilco de Kreij, 5:32). The blended result looks mediocre and tells you nothing about which piece to cut.
He argues that a per interest breakdown solves this, so "you can actually see with certainty which ones are working or which ones are not" (Wilco de Kreij, 7:07). Read that as a vendor claim rather than a neutral finding. He makes it about Connect Explore, which is his own tool, and he says so in the same passage.
Which tools help an insurance agency diagnose and run ad campaigns?
No single tool covers the path from ad to bound policy. Ad platforms report delivery and cost, design tools produce the creative, email tools work the follow up, and campaign automation tools build and launch the ads. The table below describes what each one does and where each one stops.
| Tool | What it does | What it addresses for an agency | Requires advertising knowledge | Where it stops |
|---|---|---|---|---|
| Meta Ads Manager | Builds, runs and reports on campaigns across Meta placements | The native place to build one ad set per interest and read results by breakdown | Yes. Ad sets, budgets and reporting | Every interest tested has to be built by hand, and the options you can pick from are whatever its own search returns |
| Google Ads | Runs search, display and video campaigns on Google properties | Reaches people already searching for coverage instead of people who match an inferred interest | Yes. Keywords, match types and bidding | Costs are set in a live auction against other bidders, so check current cost per click in your own account rather than assuming a benchmark. It cannot tell you whether a click came from a state you are licensed in |
| Canva | Creates and edits static and video creative from templates | Lets an agency produce ad and landing page visuals without a designer on staff | No for design work. It does not teach targeting or media buying | It makes creative and nothing else. It does not buy media, choose audiences or report delivery |
| Mailchimp | Email marketing and audience management | Works stale quote lists and renewal touches after a lead already exists | Low for email sends. It does not cover paid campaign setup | It cannot diagnose paid delivery, and it depends on a list you already own and have permission to email |
| SaleADS.ai | AI software that creates and launches advertising campaigns on Meta, Google and TikTok for business owners, with no design or advertising expertise required | Removes the campaign build step for an owner with no in house media buyer | No, by its own description. You still have to know your licensed states, your lines and your compliance rules | It works inside the ad platforms it supports, so it does not touch your agency management system, carrier quoting, or licensing and compliance review |
Disclosure: SaleADS.ai is the product of the company that publishes this site. It is listed here as one option among several, not as a recommendation over the others. No tool in this table has been ranked, scored or rated by us, and none of the rows carries a price, because pricing changes and we are not quoting numbers we cannot verify.
What else do agency owners ask about insurance ad targeting?
Four questions follow naturally once an agency starts reading its funnel by stage rather than by cost per lead. They cover what a cheap lead really means, whether to keep big brand audiences, how many interests one budget can test, and how regulated lines change what targeting is allowed.
Is a low cost per lead a good sign for an agency? Not on its own. A low cost per lead sitting next to a low bind rate usually points to an audience that is broad rather than efficient. Cost per bound policy is the number worth optimizing.
Should I stop targeting big carrier brand interests entirely? Not necessarily. Keep one as a control audience so you have a benchmark, then test narrower interests against it rather than replacing everything at once.
How many interests should I test at once? As many as your budget can support with a readable result per ad set. Testing thirty interests properly means funding thirty ad sets, which is why small batches are the practical route on a limited budget.
Does layering work for Medicare or life products? The mechanic works the same way, but eligibility rules and platform category restrictions matter more in those lines. Verify the allowed targeting controls in the platform's own policy documentation before building the audience.
Where does this information come from?
The targeting mechanics in this article come from one public video about interest based targeting on Facebook, quoted verbatim and linked to the exact moment. The insurance application is ours. Our source record lists the video as ten minutes long with 35,942 views at the time we transcribed it.
Source: "5 Mistakes Every Facebook Advertisers Makes with Interest-Based Targeting", channel Wilco de Kreij: https://www.youtube.com/watch?v=RgKCSHWt7Wo
What we took from it: the explanation that a platform interest is not a page's fan base and is built from broader signals (1:36), the point that more famous targets produce less targeted audiences (3:11), the warning that a single ad set mixing many interests hides winners and losers (5:32), the cost arithmetic of testing thirty interests in separate ad sets (7:07), and the definition of layering as targeting the overlapping segment (7:53). Every quoted sentence above is reproduced word for word from that video, and each link opens at the block of the transcript where the sentence appears.
Two things to keep in mind about the source. It is one practitioner speaking from his own experience, not a platform document or a study, and he promotes his own interest research tool inside the video, which is why that tool is not in the comparison table above and why the certainty claim quoted earlier is labelled as his.
The insurance application, the stage by stage diagnostic table, the line of business layering examples and the triage order are ours and are not in the source video. No figure appears anywhere in this article that we cannot point to in that video or that you cannot pull from your own account. Run the diagnostic against your own agency's numbers.