Should your funeral home run Facebook ads? What 2 campaigns and $437 showed

In March 2026 we spent $82.89 on Facebook ads over 3 days and got 24 leads. In August we spent $354.79 over 15 days and got 96. Both campaigns worked, at almost the same price per lead — $3.45 then, $3.70 now.

The reason to write this twice is that everything else changed. The placement that produced 83% of our leads in March produced 34% in August. The platform that produced nothing in March turned out to be our cheapest. If we had spent the second campaign defending the conclusions of the first, we would have spent it in the wrong places.

A disclosure before anything else: we make software for funeral homes, so we are a vendor writing about vendors. The numbers below are our own ad account, not a case study about someone else. If you saw a FuneralGuestbook ad in your feed this spring or this month, this is what it cost us to put it there.

What we were actually buying

We were advertising to funeral home owners. You were the audience. That matters for how much of this transfers to a funeral home advertising to families in its own county, so it’s worth being exact about it.

What transfers cleanly is the machinery — how Meta spends a budget across placements when you let it, what it does to your images after you approve them, and which settings are on before you touch anything. That machinery is identical whoever you’re targeting.

What doesn’t transfer is our cost per lead. A funeral home reaching bereaved families in one metro is bidding in a different auction against different competitors, and anyone who tells you their number predicts yours is selling something. Take the method, not the $3.70.

Run 1: March, $82.89, 3 days

One ad, run with automatic placements, budget spread by Meta across every surface it could find.

Placement	Spend	Impressions	Leads	Cost per lead
Facebook feed	$48.37	1,777	20	$2.42
Facebook Reels	$20.83	717	3	$6.94
Facebook Stories	$3.75	82	1	$3.75
Instagram Reels	$5.41	115	0	—
Instagram feed	$2.40	85	0	—
Facebook profile feed	$0.80	53	0	—
Instagram Stories	$0.48	16	0	—
Facebook in-stream video	$0.72	28	0	—
Facebook Marketplace	$0.13	8	0	—
Facebook search	$0.00	3	0	—
Total	$82.89	2,884	24	$3.45

The feed took 58% of the spend and returned 83% of the leads. Instagram, all 3 of its placements together, took $8.29 and produced nothing. At the time we drew the obvious conclusion: the people who hire funeral directors were in the Facebook feed, Instagram was a different room, and a small budget should pick its placements by hand rather than pay to rediscover that.

That conclusion lasted 5 months.

Run 2: August, $354.79, 15 days

Different campaign, and honestly a different animal: better creative, per-placement copy instead of 1 image, and a longer run — long enough that Meta’s optimizer finished exploring, which the 3-day run never did. We left placements automatic on purpose this time.

Placement	Spend	Impressions	Leads	Cost per lead
Facebook Reels	$136.50	6,889	42	$3.25
Facebook feed	$135.97	8,399	33	$4.12
Facebook search	$44.68	9,558	8	$5.58
Instagram Reels	$13.45	480	5	$2.69
Instagram feed	$6.26	309	2	$3.13
Facebook Stories	$5.94	272	2	$2.97
Facebook in-stream video	$4.57	311	2	$2.29
Instagram Stories	$2.22	110	1	$2.22
Facebook profile feed	$1.34	117	1	$1.34
Facebook notifications	$1.14	13	0	—
Facebook Marketplace	$0.62	60	0	—
Audience Network	$0.55	16	0	—
Facebook Reels overlay	$0.28	35	0	—
Unattributed	$1.27	98	0	—
Total	$354.79	26,667	96	$3.70

Read those 2 tables side by side and 3 of our March conclusions come apart.

Reels went from worst to first. In March it delivered leads at 2.9 times the feed’s rate and looked like waste. In August it was the single largest source of leads and cheaper than the feed — $3.25 against $4.12.

Instagram went from nothing to our cheapest platform. Its 3 placements produced 0 leads in March. In August they produced 8, for $21.93, at $2.74 each — a better rate than any Facebook surface.

Facebook search went from a rounding error to 9,558 impressions. It took $0.00 and served 3 impressions in the first run. In the second it took $44.68 and produced 8 leads, at the highest cost per lead in the table but not zero.

The feed is still excellent. It is simply no longer the whole story, and the map we drew in March would have told us to buy only the thing that used to work.

So which is right — pick your placements, or let Meta pick?

The honest answer is that our 2 campaigns do not settle it, and we would rather say so than pick the tidy version.

They differ in more than placement strategy. The August run had better creative, per-placement copy, a longer window, and enough conversions to finish the learning phase. Any one of those could explain the shift. What the pair does establish is narrower and more useful:

A 3-day placement map is not a finding. It is a snapshot of one creative in one auction in one week, and ours did not survive contact with the second run. That applies to our March table, and it applies to any placement breakdown a vendor shows you — including this one. The August numbers will age too.

So the rule we would now give a funeral home is about reading, not about picking:

• Run long enough to get past exploration before you conclude anything. If your budget is small, that means running a small budget for longer rather than a bigger one briefly.
• Open the placement report every week, not once. Ours inverted inside 5 months.
• Cut a surface when it has spent real money and produced nothing over a window you would defend, not after 3 days.

The one placement judgment we would still make by hand, and it is not about cost: Audience Network. Automatic placement includes it — third-party apps and websites, which in practice means your ad rendering between levels of somebody’s puzzle game. Ours took $0.55 there. For most advertisers that is an aesthetic complaint. For a funeral home it is a judgment about where the name of your business appears next to an invitation to grieve, and no cost-per-lead number answers it.

The settings that change your ad after you sign off on it

This is the part a funeral home should read twice, and nothing about it changed between the 2 runs.

Meta applies a set of automatic improvements to ad creative called Advantage+ creative enhancements. Its own help page is direct about what that means: “Although creative enhancements will try to maintain the core message of your ad campaign, the media and text you upload may be adjusted.” Same page, one sentence later: “some enhancements are turned on by default but you can turn them off at any time” (turning them off).

Read the enhancement list and the abstraction gets concrete fast. Two entries from Meta’s catalog of them:

• Add animation (AI) — “designed to automatically add common movements like panning, zooming, rotating and more to your eligible images and transform them into short, engaging videos.”
• Add music (AI) — “If you choose automatic, we will automatically apply either AI-generated music or music from the existing music library based on what we think will be more performant.”

Put those against a funeral home’s actual creative. You upload a photograph of your chapel. The system may pan and zoom across it, turn it into a short video, and score it with music it selected on your behalf because that combination tests well. Nobody approved that version. It is not what you would have made, and the first time you see it may be when a family mentions it.

There’s a trap inside the trap, and it’s in Meta’s own instructions: “Some enhancements can only be turned off in Advanced preview, and won’t be visible in the Ad creative. To review the status of all enhancements, check both sections.” A director who opens the obvious panel, switches everything off, and publishes has not switched everything off.

None of this is Meta behaving badly. These features lift performance for a furniture store, which is who they were built for. Deathcare is one of the few businesses where a more engaging version of your ad can be a worse ad, and the platform has no way to know that. You have to tell it, twice, in two different panels.

Meta may label your funeral home’s ad as AI

If you let the enhancements run, there’s a consequence beyond taste. Meta labels ads that its own generative tools made or altered. From its transparency announcement: “Our labels are designed to help people understand when images and videos have been created or significantly edited with our in-house generative AI tools.” The same post says Meta will “begin automatically detecting ads created or edited using third-party AI tools through industry-standard signals” and label those too.

So the choice isn’t only whether an animated, auto-scored version of your chapel is dignified. It’s whether a family looking at your ad sees a tag telling them the picture was made by a machine. In an industry that sells trust and continuity, that label costs more than it does almost anywhere else. If you do use generated imagery, label it yourself and be plain about it — a photoreal scene of mourners who never existed is the one place we’d argue the disclosure is owed regardless of what any platform requires.

What a small budget actually controls

Three levers, in the order they matter.

The daily budget is the guard, not the bid. Whatever you set as a daily amount is the worst case for that day. Bid settings, cost targets and optimization goals all shape delivery, but the daily number is the only one that answers “how much can this cost me by Friday.” Ours ran at $25/day in August. Set it to a figure you’d shrug at, and treat the first weeks as the price of the placement report — because that report is the actual deliverable of a first test.

A kill rule, written down before launch. Ours is simple: if a placement has spent real money and produced nothing by the time the campaign has enough data to be read, it goes off. Deciding that in advance is what stops a campaign from running for a month on the strength of a status chip that says Active. Note the second half of that sentence, which the March run taught us to take seriously: enough data to be read.

Speed of answer. A lead form gives you a name and a number and nothing else. We hold ourselves to answering the same hour, including weekends, because a family or a director who filled in a form on Saturday has moved on by Monday. We can’t show you a conversion study proving the size of that effect, so treat it as our operating practice rather than a finding.

One craft note from building the ads. Headlines get cut short, and where they get cut varies by placement — we trimmed ours to 27 characters because that’s where they stopped rendering in full. Write for the shortest surface you’re buying, or write different copy per placement. Both work; only one of them is free.

The setting that lies to you

A campaign’s status field and its delivery are different facts. We have watched a campaign read Active, funded, past its start date, with every serving-state field clean, and serve nothing at all. The status column describes configuration. Impressions describe reality.

If you take one operational habit from this: never judge a campaign by whether it says it’s running. Open the report and look for impressions. Zero impressions on a funded, active, started campaign is not a slow start; it’s a problem, and it’s usually something a settings screen won’t tell you.

What we still have not proven

We have now generated 120 leads across 2 campaigns for $437.68. We have not shown that those leads became customers.

That distinction is most of what’s wrong with marketing writing in this trade, so we’d rather stay on the correct side of it — including now, when an update would be a convenient place to quietly drop it. The August leads are moving: 12 have become real conversations and 3 funeral homes have booked a demonstration. None of that is revenue. We are also not going to divide those numbers by 96 and call the result a conversion rate, because the funnel is 2 weeks old and most of the leads in it are younger than that. When there is a number worth reporting, we will report it here.

Lead generation on Facebook is proven for us, at that price, on that audience, on those days. Revenue is a separate experiment, and it runs on the phone rather than in Ads Manager. Anyone publishing a cost-per-lead as though it were a cost-per-customer is quoting the easy half.

The same caution applies to the wider picture. Facebook is where a lot of funeral homes already do their community posting, and organic reach on the obituary page a town actually reads is a different and often better asset than anything you can buy. Paid social is a supplement to that, not a replacement for it — and if you’re weighing where a limited marketing budget goes against a consolidator down the road, the local tactics piece is the better first read.

The short version

If a funeral home came to us tomorrow asking whether to run Facebook ads, the answer would be yes, with 4 conditions: run long enough to read the report before you trust it, turn off the creative enhancements in both panels, set a daily cap you’d shrug at, and answer every lead the same hour.

The $437.68 was worth it. Not for the 120 leads — for 2 tables that disagree with each other. The first campaign bought us a map. The second one bought us the more valuable thing, which is the knowledge that the map expires.