The Plumber Who Found Out Half His Facebook Ads Were Running Between 1am and 5am

A Pacific Northwest plumber doing around $4M a year ran Meta ads for 14 months before he ever opened the delivery-by-hour breakdown. His CSR manager had been complaining for months that the "Facebook leads" all came in as overnight voicemails or Messenger pings that sat until 7am.

The owner assumed it was just the nature of Facebook leads. He finally pulled the hour-of-day report on a rainy Tuesday in March. Forty-eight percent of his ad impressions and 41 percent of his ad spend had fired between 10pm and 6am. He was paying Meta to show plumbing ads to insomniacs and night-shift workers while his CSRs were home.

Not one of those overnight leads got answered inside the 5-minute window that makes leads close. Most never got a callback at all. He had been funding the world's most expensive lead graveyard.

What the hour-of-day report actually showed

Meta Ads Manager has a breakdown called "by time" that splits delivery by hour. It is one dropdown most contractors never use because the default view is "by campaign" and owners rarely change it.

Jon Loomer's 2025 ad scheduling deep dive puts the opportunity plainly: "the core value of Facebook ad scheduling is dayparting, or restricting the days and times at which your ad can appear, allowing you to choose to run your ad only on certain days and time slots." [Jon Loomer Digital] LeadEnforce's 2025 teardown on dayparting put numbers on it: "Meta's own studies reveal that engagement rates can vary by up to 30 percent depending on the time of day and day of the week." [LeadEnforce]

The plumber had no dayparting turned on. Meta's delivery algorithm was serving impressions 24 hours a day. The overnight hours are cheap for Meta because CPMs drop when few big advertisers are bidding. That cheap inventory is exactly where the budget quietly drains.

Text Sully: "show me Facebook Ads impressions and spend by hour of day for last month, flag hours when our phone room was closed"

The 5-minute rule the overnights were breaking

Even if someone filled out a Facebook lead form at 2am, what happens next is the real problem. Hatch's 2025 data across 941 contractor users found "88 percent of users are taking more than 5 minutes to reply to leads. The most common response times were 1 day (37 percent of users), followed by 30 minutes (33 percent of users), with only 3 percent of users responding in less than a minute." [Hatch]

The Scorpion home services blog citing multiple response-time studies puts the consequence brutally: "outreach has to happen within five minutes, after which your chances of reaching that lead decrease by 900 percent." [Scorpion] An overnight lead sitting until 8am is already 18x less likely to close than a same-business-hour lead.

That compounds across the plumber's entire overnight spend. Forty-eight percent of impressions, plus a 900 percent drop in reach-probability on any overnight lead that did come in. It was like paying for ads and then paying a second time to not answer the leads the ads produced.

One contractor in a ContractorTalk thread on Facebook ads wrote: "half the leads come in at night, nobody answers them, owner wonders why Facebook doesn't work." [ContractorTalk] He had described the plumber's exact situation word for word.

Text Sully: "for Facebook lead form submissions last month, what was the average time from submit to first callback, and how many got a callback within 5 minutes"

The phone room closes at 7pm

The plumber's CSRs worked 7am to 7pm Monday through Saturday. Sunday the on-call tech took emergency calls only. That meant 60 hours a week of staffed phones out of 168 hours total. About 36 percent of the week was covered.

His ads were running 100 percent of the week. That is a 64 percent mismatch between ad delivery and answer capacity. Every impression during off-hours had to either:

• Produce a lead that waited until morning (lower close probability)
• Produce a call that went to voicemail (extremely low close probability)
• Do nothing, burning the impression for no return

Hook Agency's 2025 home services Facebook ads guide recommends: "scheduling ads during business hours or high-call times is part of a strong local Facebook ad strategy." [Hook Agency] It is boilerplate advice that almost no small contractor account actually implements.

Turning dayparting on is three clicks: set the campaign to Lifetime Budget, open ad set scheduling, draw the grid around phone-room hours. The plumber did it in five minutes on a phone while standing in line at Home Depot.

Text Sully: "on what days and hours did our Facebook ads get the most form submissions last quarter, and which of those hours did we have CSRs staffed"

The emergency plumbing myth

One reason contractors leave Facebook ads running all night is the belief that emergency plumbing happens at 3am, so the ad should too. Believe Advertising's 2025 guide to plumbing Facebook ads debunks this directly: "when a pipe bursts at 2:00 AM, homeowners go straight to Google and type plumber near me rather than opening Facebook to scroll through their newsfeed." [Believe Advertising]

Facebook is interruption advertising. The platform works for planned purchases that you want to capture before the need is urgent. Water heater replacements, repipes, drain maintenance memberships, sewer scope upsells. All high-margin, all planned, all bookable during business hours.

Hook Agency's plumbing Facebook ads teardown says the same: "most plumbing revenue comes from planned work such as water heater replacements, whole-home repiping, sewer repairs, and maintenance memberships. Facebook ads let you capture these homeowners before they need emergency service." [Hook Agency]

Emergency plumbing belongs on Google Local Service Ads and high-intent Google Ads, not on Facebook. The plumber in this story turned off an emergency-angled Facebook campaign entirely and shifted that $1,200 to LSA. Booked jobs went up.

The Facebook audience network leak

Along with the hour-of-day problem, the plumber found another leak. His campaigns were auto-enrolled in Audience Network, Meta's extension of ads to third-party apps and sites. A 2025 review of his reports showed Audience Network driving 8 percent of his spend and 2 percent of his leads.

Every guide to plumbing and HVAC Facebook ads in 2025 says the same thing, and the HVAC Marketing Experts piece put it clearly: "avoiding Audience Network is recommended as you could easily waste your budget fast." [HVAC Marketing Experts] Audience Network ads appear in game apps, content farms, and random mobile properties. The traffic converts at a fraction of in-feed Facebook.

Disabling Audience Network is one checkbox at the ad-set level. The plumber did it the same afternoon he set dayparting. Combined with the schedule restriction, his CPL dropped from $71 to $48 inside 30 days.

Text Sully: "show Facebook ads spend by placement (Feed vs Audience Network vs Stories) for the last 90 days, ranked by cost per booked job"

What he lost by killing overnight delivery

Volume dropped. Forty-one percent of spend out of the system meant form submissions dropped from 94 per month to 58. On paper, fewer leads.

But cost per booked job dropped from $301 to $198. Booked jobs held flat at around 27 per month because the lost overnight leads had almost never been booking anyway. The plumber was paying $1,968 less per month and booking the same number of jobs.

This is the pattern every dayparting case study shows. Jon Loomer's "should you run ads on a schedule" piece cites contractor accounts that cut CPL 20 to 35 percent by tightening delivery to business hours. [Jon Loomer] Not everyone reads the guide. Everyone can still read their own hour-of-day report.

The lag between ad-click and lead

There is a second nuance worth watching. Even if an ad shows at 9pm, the lead form submit might come at 9:20pm. A strict 7pm dayparting cut-off can still produce leads that hit at 7:45pm as the last impressions close out.

The plumber's solution was to set delivery 7am to 9pm (not 7am to 7pm). That gave a 2-hour buffer after CSRs went off for the last impressions to land leads that could still be dialed during staffed hours. His final schedule shaved overnight impressions down from 48 percent to 7 percent of delivery without crushing the tail.

The 7 percent that remained covered people who kept scrolling at 10pm and converted the next morning after a retargeting impression. Those leads actually converted. The 2am Audience Network impressions did not.

The callback workflow that caught what remained

For the 7 percent of leads that still came in overnight, the plumber layered in a missed-call text-back flow. Any lead form submission outside staffed hours got an immediate auto-reply: "Got your message. We open at 7am Monday to Saturday. We will call you first thing." A second message went to the on-call dispatcher at 7am with the lead details pre-filled.

That workflow caught about half of the overnight leads into a same-day callback. Booking rate on those leads was 38 percent versus 9 percent under the old "we'll get to them when we get to them" workflow.

Full response-agent workflows that do this across every channel are exactly what an AI missed-call followup agent is designed to run.

What this means for your shop

Open Meta Ads Manager. Change the breakdown to "by time" and look at your last 30 days of impressions by hour. If anything more than 15 percent is firing outside your phone-room hours, you have a dayparting problem.

Turn on scheduled ad set delivery. Match it to phone-room hours plus a 2-hour tail. Disable Audience Network. Kill any campaign that targets "emergency plumber" through Facebook (the intent is wrong for the platform).

Layer an auto-reply on any lead form submission outside business hours. If you have time, also layer a 7am automated dispatch to whoever opens the phone room.

This is not a Facebook problem. This is a scheduling and response-workflow problem disguised as a Facebook problem. Facebook is fine when it runs on your hours, not 24/7 on Meta's default settings. The same pattern applies across Meta, YouTube, TikTok, and every platform that defaults to all-hours delivery. Catch it on one platform and the playbook transfers.

If you want to see exactly which hour of last week's lead submissions went unanswered longest, that is one of the questions no dashboard will answer on its own. It takes tying ad delivery, lead form time-stamps, and CSR response logs together. That is where the money actually hides.

Sources

• Jon Loomer Digital: Facebook Ad Scheduling
• LeadEnforce: Day-Parting in Facebook Ads
• Hatch: HVAC Speed to Lead Response Rates
• Scorpion: Why Speed Wins in Home Services Response Time
• Hook Agency: Home Services Facebook Ads Guide 2025
• Hook Agency: Are Plumbing Facebook Ads Worth It?
• Believe Advertising: How to Create Facebook Ads for Plumbing and Electrical Services
• ContractorTalk: Facebook Ads thread
• Jon Loomer: Should You Run Ads On a Schedule?