Your ad has been live for nine days. It's spent $340. You have 11 likes, one comment from your cousin, and zero phone calls. So you do what every owner does: assume Facebook ads don't work for your business.
They might not — cause #7 is exactly that, and it's real. But usually the ad isn't failing for a mysterious reason. It's failing for one of nine, and they're not equally likely. Here they are in the order they actually show up: the symptom, what's happening underneath, and a fix with a number on it.
1. Your budget is too small to ever leave the learning phase
The symptom: results are erratic. Two leads Tuesday, nothing for four days, one on Sunday that cost $130. Nothing stable enough to judge.
What's happening: every ad set enters a learning phase where Meta spends your money figuring out who converts. Meta's own guidance is that it stabilizes at roughly 50 conversions per week — at a $50 cost per lead, that's $2,500 a week, $10,000 a month. Essentially no local contractor spends that, which means most small accounts never fully exit the learning phase. That's normal, not a bug. The practical floor is lower: 15 to 20 conversions in the first two weeks. Below that, you and the algorithm are both guessing.
The fix: work backwards from your cost per lead. At $50 CPL, 20 leads in two weeks is $1,000 — about $70/day. Can't do that? Halve the radius, cut to one offer, and run $30 to $40/day in one ad set rather than $10/day across three: you're feeding one learner instead of starving four. Our breakdown of what Facebook ads actually cost has the per-trade numbers.
Launch your first Facebook & Instagram ad in 2 minutes
Describe your business: the AI writes the copy, designs the visual, sets the targeting, and publishes your ad. Leads arrive with their phone number.
Try Leadria free7-day free trial — no credit card — cancel anytime
2. You killed it on day three
The symptom: "I tried Facebook ads. Spent $150, got nothing, turned it off."
What's happening: you ended an experiment before it produced a result. The learning phase runs roughly 7 days, and days 1 to 3 are the most expensive, least representative days of an ad's life — cost per lead on day 2 is routinely 2 to 3 times what it settles at by day 10. Judging on day 3 is like judging a new hire on their second morning.
Worse is the owner who keeps poking it. Every meaningful edit — new image, new audience, a budget change beyond about 20% in one step — pushes the ad set back into learning. Tweak it Monday, Wednesday and Friday and it lives permanently in the phase where it performs worst. That's not optimizing, it's resetting.
The fix: a written rule. Don't touch it for 7 days. Not the budget, not the photo, not the copy. Judge at day 14 on cost per lead and answered calls. If you can't afford 14 days at your daily number, you can't afford the test.
3. You don't have an offer
The symptom: decent impressions, a handful of clicks, no leads. The ad looks fine to you.
What's happening: the most common creative failure by a wide margin, and it's invisible to the person who wrote it. "Quality workmanship since 1998. Licensed and insured. Call for a free estimate." That is not an offer — it's a description of your existence. Every competitor within 30 miles could run it with a different phone number. An offer answers one question: what do I get, and what does it cost me to find out? "Free estimate" fails, because asking for an estimate means admitting you're about to spend $9,000. The homeowner isn't there yet — she's at "I wonder if that stain means anything."
The fix: make the yes smaller and more specific. Not "free quote" but "free 20-minute roof check — we'll tell you how many years you have left, and we don't sell you a roof you don't need." Not "AC service" but "$89 tune-up, 45 minutes, this month only." Swapping a description for a real offer routinely moves cost per lead 30% to 50% on the same budget and audience. Highest-leverage change on this list, and it costs nothing.
4. You're optimizing for the wrong thing
The symptom: likes and reach, a few comments, zero phone numbers. The ad feels popular and produces nothing.
What's happening: you hit the blue Boost button. Boosting optimizes for engagement — Meta finds people likely to react to a post, a different group from people willing to hand over a phone number. It does that job well, which is the problem: you got what you asked for. A boosted post can deliver engagement at $0.05 to $0.20 a pop and not produce one lead in $400.
The fix: run the same money through Ads Manager with the Leads objective and an instant form. Same photo, same copy, same audience, same $400 — except Meta is now hunting form-fillers instead of thumbs-up-clickers. Owners who change only this commonly go from zero to 8 to 15 leads on that same $400. Mechanics in our guide to getting leads instead of likes.
5. Your audience is too narrow
The symptom: high CPM, the same people seeing the ad over and over, performance that decays after a week.
What's happening: you built what feels like a smart audience — 3-mile radius, homeowners, 35 to 60, interested in home improvement, HGTV, Home Depot, gardening and DIY. That stack leaves 8,000 people, and Meta can't find your buyers because there aren't enough people to find them among. CPM climbs, frequency passes 5 impressions per person per week inside ten days, and the ad burns out. Each interest layer is also a guess, and six guesses multiplied together exclude the buyer who never liked a home-improvement page.
The fix: go broad and let the machine work. A 15-to-25-mile radius (or whatever your crews actually drive), homeowners if that option exists in your market, ages 30 to 65+, and zero interests. Aim for 100,000 to 1,000,000 people. It feels reckless. It isn't — the creative does the filtering now, and Meta's delivery finds your buyer better than your guesses do.
6. Your creative doesn't stop anybody
The symptom: click-through rate under 1%, and a CPM that keeps climbing.
What's happening: the ad looks like an ad. A stock photo of a model in a clean hard hat, or your logo on a gradient. A homeowner scrolling past their nephew's baby photos doesn't stop for that. And low CTR isn't just fewer clicks — Meta reads it as a low-quality ad and charges more per thousand impressions. Weak creative is expensive twice.
The fix: three things, in order of impact. Before-and-after beats everything in the trades — same spot in the yard, every job; two minutes a job builds a library worth thousands. A real face — you, your crew, in the truck, on the roof — beats a polished object shot. And real local houses, not a mountain lodge from a stock library: a house like the ones on your prospect's street reads as "from here." Refresh every 4 to 6 weeks, because local audiences are small enough to exhaust. Trade examples: roofers, HVAC, contractors.
7. Your business is on the wrong channel
The symptom: everything above is done right and it still doesn't produce. This one is real, and nobody selling you ads will say it.
What's happening: Facebook creates demand. It doesn't capture it. If your customer only thinks about you when something is broken right now, they aren't scrolling — they're typing "emergency plumber near me" into Google and calling whoever picks up within 5 minutes. No amount of creative work puts you in front of someone whose water heater just flooded the garage.
Wrong channel: burst pipes, AC out in August, lockouts, active infestations, a roof actively leaking — that money belongs in Google Search and a well-fed Google Business Profile. Right channel: anything planned — roof replacement, repipes, remodels, panel upgrades, maintenance plans, EV chargers, listings. The decision takes weeks, and Facebook is where the idea gets planted.
The fix: split by demand type. If 80%+ of your revenue is emergency work, don't fix your Facebook ads — move the budget. Our Facebook vs Google comparison has the rule per trade.
8. The leak is after the click
The symptom: Ads Manager shows leads. Your calendar shows nothing. You conclude the leads are junk.
What's happening: some are. Most went cold in a Meta dashboard nobody opened. A Facebook lead is lukewarm by nature — they filled a form in 8 seconds and went back to scrolling. Calling in 5 minutes versus 30 moves contact rate by a multiple, not a few points, and a lead called the next morning is often unreachable: the contractor who called first is already in the driveway. The other version — speed is fine and the form is the problem: seven fields, a required email, six seconds to load on a phone.
The fix: get leads onto a phone screen the second they arrive, and assign one named person to call in under 5 minutes during business hours. Keep the form to name, phone, ZIP, and one qualifying question — "Do you own the home?" cuts junk without killing volume. If you can't call fast, don't advertise: you're buying leads for whoever calls second.
9. You're expecting sales but measuring likes
The symptom: you genuinely can't say whether it worked. You know you spent $600.
What's happening: nobody ever calculated what a lead is worth, so every number gets judged on vibes. "$70 a lead" sounds expensive to everybody, always — until you attach it to a job. The fix is one line of arithmetic: close rate x job margin = what a lead is worth to you.
| Business | Job value | Close rate | A lead is worth | So $70/lead is |
|---|---|---|---|---|
| Roofer | $10,000 (30% margin) | 20% | $600 | Cheap — buy every one |
| HVAC replacement | $7,000 (30% margin) | 15% | $315 | Very profitable |
| Cleaning (recurring) | $2,400/yr (40% margin) | 25% | $240 | Works on retention |
| Salon | $60 (50% margin) | 40% | $12 | A disaster |
Same $70 lead, four different verdicts. Until you've done this, you have no basis to say the ads failed — and the roofer who quit at $70 a lead walked away from an 8x return because the number sounded big.
Diagnose yours in 10 minutes
Open Ads Manager and walk the funnel in order. Find the stage where the number collapses — that's your cause, and you can stop looking.
- Impressions. Under 5,000 in a week? Budget or audience size (causes 1 and 5). Meta isn't spending because it can't find anyone.
- CTR. Below 1%? Creative or offer (causes 3 and 6). They saw it and kept scrolling. Healthy local lead ads run 1% to 3%.
- Clicks to leads. Under 10% of clicks becoming a lead? Your form (cause 8). Instant forms should beat that comfortably.
- Leads to answered calls. Reaching under 50%? Follow-up speed (cause 8 again). Check your call timestamps against the lead timestamps — the gap is the answer.
- Answered calls to jobs. Under 10% closing after real conversations? That's not the ads. Keep reading.
One caveat: if you see no impressions at all, the ad may still be in review or was rejected — check the delivery column before diagnosing anything else.
Sometimes the ads are fine and the business isn't
The uncomfortable one. Ads amplify what's already there — they don't create it. If people see your ad, click, fill the form, talk to you and don't buy, the ad did its job and handed off to something that isn't working.
- Capacity. Booked eight weeks out? New leads get "we'll call you back" and go elsewhere. You don't need ads — you need to raise prices 10% to 15% and see who's left.
- Reviews. Every lead Googles you before calling back. A 3.2-star average with 11 reviews against a competitor's 4.8 with 200 means you're paying to send homeowners to their reputation. Fix that first: it's free and it moves everything downstream.
- Pricing. Sitting 40% above market with no visible reason why just buys more conversations that end the same way. Justify the premium in the ad, or fix the number.
None of this is a reason to skip advertising. It's a reason to sequence it: fix the leak, then turn on the tap.
What to do this week
Pick the first cause that's true for you and fix only that — almost nobody needs all nine. In most accounts it's budget, patience, and the offer: causes 1, 2 and 3.
Then: one ad set, one audience, $40/day minimum, a real offer with a small yes, Leads objective, broad targeting, a before-and-after photo, and a rule that you don't touch it for 7 days. Judge at 14 on cost per lead and answered calls. If the build itself trips you up, start with our complete Facebook ads guide for small business. And if Ads Manager is why the ad has sat untouched for six weeks, Leadria exists for that: you describe your business in a sentence, and the AI writes the copy, generates the visual, sets the Meta targeting, and publishes to Facebook and Instagram. Leads come back with a phone number attached — 7-day free trial, no credit card.
The discipline is short: one offer, enough budget, seven days of not touching it, a call in five minutes. Most "Facebook ads don't work" stories end before any of those four ever happened.