- Run OEP as three phases with three different jobs. Prospecting builds the audience Meta needs. Retargeting closes the leads prospecting found. Scale spends into whatever is converting once volume proves itself.
- Special Ad Category rules, enforced on insurance since January 2025, strip out age, gender, and zip targeting. Your creative and your first-party data have to do the work your audience settings can’t.
- Lead-form ads win on volume and cost. Click-to-call wins on lead quality. Most agencies need both running at once, split by where in OEP you are.
- QA lead quality daily during OEP, not weekly. A bad creative or a broken pixel can burn a week of budget before a weekly review catches it.
Quick answers:
- What’s the right Meta ad budget split across an OEP campaign?
- Should I use lead forms or click-to-call for ACA and U65 leads?
- What targeting is actually left after Special Ad Category restrictions?
- How do I keep health ad creative compliant without going generic?
- How often should I QA lead quality during OEP?
Every U65 and ACA lead-gen operation running paid social already knows the Medicare AEP playbook. Hit the audience hard before the deadline. Retarget the ones who almost converted. Ride whatever’s working until the window closes. ACA Open Enrollment runs on the same clock. For 2026 coverage, the window opens November 1, 2025 and runs through January 15, 2026 in most states. California, Connecticut, D.C., Illinois, New Jersey, New York, Pennsylvania, and Rhode Island run through January 31 (HealthCare.gov). Miss the December 15 cutoff and a lead’s coverage doesn’t start until February 1 instead of January 1. That date matters more to your creative than most agencies treat it.
Meta’s own rules changed the game underneath this playbook. Health insurance ads fall under Meta’s Financial Products and Services Special Ad Category, enforced since January 2025. That single classification removes most of what a media buyer used to lean on. No age or gender targeting beyond a blanket 18-65+ bucket. No zip-code radius. No detailed interest or lookalike audiences built the old way. Lead forms can’t even collect age or gender as fields (Meta Special Ad Categories; Meta Transparency Center). You’re targeting states, not zip codes. You’re targeting everyone 18+, not the 45-64 bracket you know actually needs a plan. This is the constraint the whole campaign has to be built around, not an inconvenience to work past.
What’s the right Meta ad budget split across an OEP campaign?
Structure the campaign in three phases. Each one solves a different problem. A single “OEP campaign” running flat from November 1 to January 15 wastes budget on the wrong lever at the wrong time.
Phase 1: Prospecting (roughly the first three to four weeks). With detailed targeting gone, Meta’s algorithm needs volume and signal to find the right audience on its own. This phase feeds the pixel enough conversion data, fast, so the delivery system has something to learn from before the window gets busy. Run broad state-level targeting. Push volume. Accept a higher early cost per lead than you’d want long-term. You’re buying signal, not efficiency, in these first weeks.
Phase 2: Retargeting (overlapping, starting once you have real pixel data, typically week two onward). This is where lead-form abandoners, site visitors who didn’t convert, and video-view engagers get a second and third look. Retargeting audiences are exempt from a lot of the detailed-targeting restriction, because you built them from your own first-party events, not from Meta’s demographic categories. This is the highest-ROI phase of the whole window, because you’re spending against people who already showed intent.
Phase 3: Scale (the final two to three weeks before the deadline). By now you have weeks of conversion data telling you which ad sets, creative, and placements are actually producing enrolled leads, not just cheap leads. Shift budget hard into whatever’s proven. Let the algorithm’s own delivery data, not your gut, decide where the last dollars go. The mistake agencies make here is spreading budget evenly across every ad set that’s “still running” instead of concentrating it on the two or three that earned it.
What to do: Map your total OEP budget to these three phases before the window opens, not as you go. Decide the split ahead of time. Something like 30/40/30 prospecting, retargeting, scale is a reasonable starting shape. Then you’re not improvising allocation decisions during the busiest two weeks of the year.
Should I use lead forms or click-to-call for ACA and U65 leads?
Both. Just not for the same reason, and not in the same phase.
Lead-form ads (Meta’s native Instant Forms) convert at a lower cost per lead because the entire flow stays inside the app. No page load. No redirect. Fewer steps to drop off. They’re the right tool for prospecting volume, where the goal is filling the top of the funnel cheaply enough to feed the algorithm’s learning phase. The tradeoff is quality. A form filled out in eight seconds on a phone screen produces more junk: wrong numbers, dead interest, someone who tapped through without reading. A call someone chose to place doesn’t have that problem the same way.
Click-to-call ads cost more per lead because a phone call is a higher-commitment action than a tap. But the person who calls is self-selecting for real intent. That’s exactly what a lead-gen agency’s buyer is paying for. Click-to-call also sidesteps a real compliance exposure. A lead form captures written consent data you have to store and defend. A call routes straight into your existing TCPA-compliant call-handling process, if you already have one.
The operator’s answer: run lead forms in prospecting to build volume and pixel signal cheaply. Shift weight toward click-to-call as you move into retargeting and scale, where the buyer on the other end is paying for quality over quantity. Agencies that run 100% lead-form through the whole window usually see their cost per lead look great and their cost per sold lead look terrible. The metric that was cheap wasn’t the metric that mattered.
What targeting is actually left after Special Ad Category restrictions?
Less than agencies think. What’s left is mostly stuff you build yourself.
- Geography at the state level. You can still target by state, and by DMA in some setups, just not by zip radius. If your buyer only wants leads in states with certain carrier coverage, this is still workable. It’s just coarser than it used to be.
- Broad age band only. 18-65+ as a single bucket. You can’t isolate the 50-64 segment that’s statistically more likely to need marketplace coverage. Your creative and your first-party retargeting have to do that filtering instead of your audience settings.
- Custom audiences built from your own data. Uploaded customer lists, website visitors, and lead-form engagers are yours, not Meta’s demographic taxonomy, so they aren’t stripped by the Special Ad Category restriction. This is the actual lever left for precision. Not ad-platform targeting, but pixel and CRM data you own.
- Lookalikes are restricted, not fully off. They’re built from a smaller, coarser seed than pre-2025 campaigns allowed. Expect a wider, less precise lookalike than you’re used to from other verticals.
The common mistake: Treating this like a targeting problem you can out-clever with a workaround. It’s not. Meta enforces this at the account level, and testing the edges risks the ad account, not just the campaign. The workable response is investing in first-party data collection, retargeting pools, and CRM uploads early enough in the window that you have a real custom audience by the time you hit scale phase.
How do I keep health ad creative compliant without going generic?
Meta’s personal-attributes policy prohibits ad copy that implies you know something about the viewer’s health status or condition. Not just direct “you have diabetes” language, but indirect framing too. Enforcement expanded in Q1 2026 to catch phrasing like “for people managing diabetes” or “those dealing with blood sugar spikes.” Meta’s automated review now factors in the audience and interest signal alongside the copy itself, not the copy in isolation (Meta personal attributes policy summary).
That kills the classic direct-response health hook. What still works:
- Talk about the deadline, not the diagnosis. “Open enrollment closes January 15” is a fact about a calendar, not an implied statement about the reader’s health. It converts because the deadline is real and urgent, and it doesn’t touch the personal-attributes line.
- Talk about the cost problem, not the health problem. “Subsidies you might qualify for” and “average premium after tax credits” are financial claims, not health claims. That only works if you can back the number with a real, current source. That’s non-negotiable for this kind of copy.
- Lead with the outcome of acting, not a description of the reader’s situation. “See your options before the window closes” describes an action, not a condition.
Operator note: Run every piece of health-adjacent creative through the direct-and-indirect test before it ships. Read it out loud and ask: does this describe my audience’s body or their calendar? If it’s their body, in any phrasing, it’s not going up.
How often should I QA lead quality during OEP?
Daily, not weekly, for the length of the window. This is the discipline that separates agencies that come out of OEP profitable from the ones that come out with a pile of leads nobody wants to buy.
Weekly QA fails specifically during OEP because volume moves fast. A broken pixel event, a creative that started attracting the wrong audience, or a lead form with a typo in the phone-number field can burn several thousand dollars before a Friday review catches it. A daily pull of lead quality, disconnect rate, invalid numbers, duplicate submissions, and buyer feedback on sold leads, catches the failure in hours instead of days.
Build the QA loop around three checks run every day the campaign is live:
- Volume vs. baseline. A sudden spike in lead volume with no budget change usually means the audience drifted, not that you found a goldmine. Check it before you celebrate it.
- Disconnect and invalid-number rate on click-to-call. A rising rate here usually traces back to a specific ad set or placement. It’s fixable in an afternoon if you catch it same-day.
- Buyer acceptance rate. The number that actually matters. Cost per lead is a vanity number if the buyer is rejecting a third of what you send them. Feed buyer rejection reasons back into which ad sets get more budget in the scale phase.
What most agencies get wrong: They QA cost per lead daily and lead quality weekly, when it should be the reverse. Cost per lead is stable and slow to move. Lead quality during a high-volume window like OEP can turn in a single day.
Frequently Asked Questions
What’s the right Meta ad budget split across an OEP campaign?
A rough starting shape is 30% prospecting, 40% retargeting, 30% scale. Front-load enough spend to build pixel signal in the first three to four weeks. Put the heaviest weight on retargeting once you have real conversion data. Concentrate the final two to three weeks on whatever the data has already proven works.
Should I use lead forms or click-to-call for ACA and U65 leads?
Use lead forms to build cheap volume during prospecting. Shift toward click-to-call during retargeting and scale, where lead quality, not lead cost, is what your buyer is actually paying for.
What targeting is actually left after Special Ad Category restrictions?
State-level geography, a single 18-65+ age band, and custom and lookalike audiences built from your own first-party data. Zip-code radius, narrow age brackets, gender, and detailed interest targeting are all off the table for health insurance campaigns under Meta’s Financial Products and Services Special Ad Category.
How do I keep health ad creative compliant without going generic?
Anchor the creative to the deadline and the cost, not the reader’s health status. Direct and indirect health-condition language are both enforced under Meta’s personal-attributes policy as of the Q1 2026 update.
How often should I QA lead quality during OEP?
Daily. Weekly QA catches problems days after they’ve already burned budget. A daily pull of disconnect rate, invalid numbers, and buyer acceptance catches a bad ad set or broken pixel the same day it happens.
Book a free call with Elevarus if you want a second set of eyes on your OEP campaign structure before the window gets busy.
