A local service business needs one afternoon of phone footage and about four to six ad variations, not a content calendar. The reason is audience size: a plumber advertising inside a 15 mile radius is buying against tens of thousands of people rather than millions, so frequency climbs slowly at small budgets and creative lasts far longer than it does for a national brand. Film the work, show the before and after, and start on a free plan rather than a production budget.
What should you film in one afternoon?
Pick one real job and film it end to end. A single completed job produces more usable ad footage than a day of staged shots, because the before state is genuinely ugly and the after state is genuinely yours.
The shot list, in order:
- The before state, filmed properly. Wide, then close on the specific problem. Say the date out loud on camera.
- Arrival. The van, the door, the shoe covers if you use them. This is the moment a customer is nervous about.
- The work itself. Three or four clips of hands doing the job. No commentary needed.
- The mess, then the cleanup. Almost nobody films this and it answers the objection everyone has.
- The after state. Same angle, same framing, same light as the before. Say the date again.
- The customer, at the door, in their own words. Thirty seconds, one take, permission asked first.
- You, saying your first name and how long you have worked in the area. Two takes.
- The phone being answered. Local buyers want to know a human picks up.
That is 15 to 20 clips and roughly two hours including the job itself. Lock exposure before each take, put the light in front of the subject, film in 4K if the phone allows so you can crop to vertical later, and keep the product or problem in the middle third so the same clip works at 9:16, 4:5, 1:1 and 16:9.
One thing to avoid: do not narrate over the whole thing. The work is more convincing without a voice explaining that you are professional.
How many variations does a local audience actually need?
Far fewer than a prospecting set for a national brand, and this is where most agencies oversell local businesses. The number falls out of arithmetic rather than opinion.
| Daily spend | Weekly impressions at a €8 CPM | Audience in radius being served | Weekly frequency | Variations you need live |
|---|---|---|---|---|
| €10 | 8,750 | 25,000 | 0.35 | 3 |
| €25 | 21,875 | 30,000 | 0.7 | 4 to 5 |
| €60 | 52,500 | 35,000 | 1.5 | 6 to 8 |
| €150 | 131,250 | 40,000 | 3.3 | 10 or more, plus a wider radius |
Compare that with a DTC brand spending €4,000 a day, which crosses a weekly frequency of 2.5 inside week one and needs 12 to 20 variations refreshed weekly. At €25 a day in a single town, four variations can run for a month before anyone has seen one more than twice. Our 2026 fatigue benchmark, a synthesis of published platform and agency figures rather than our own experiment, puts the CTR drop at 45 percent after a fourth exposure to the same creative and identifies weekly frequency of 2.5 on Meta prospecting as where decline begins. At a local budget you are nowhere near either number.
The catch is the other end. A small radius means you exhaust the audience over months rather than weeks, so plan a refresh quarterly or when the job type changes with the season, not because a calendar says fortnightly. To run these numbers with your own radius and spend, the ad variation calculator does it faster than a spreadsheet.
The four variations worth having, if you only ever make four:
- Before and after of the ugliest job you have done, with dates on screen.
- You, first name, years in the area, one sentence on what you will not do (upsell, subcontract, leave a mess).
- The customer at the door, unedited.
- Price or process transparency: what a call-out costs, or what the first visit involves.
Why does completed work footage outperform everything else?
Because local buying is risk reduction, not aspiration. Nobody chooses a roofer because the ad was beautiful. They choose the one where they can see a finished roof, a real street, a named person and evidence that the house was left tidy.
Three things make the proof work harder:
- Hyper-local specificity. Name the neighbourhood, the type of property, the era of the building. "A 1930s semi in Kingsholm with the original cast iron" is more persuasive to the next owner of a 1930s semi than any claim about quality.
- Show the difficult part. The cracked joist, the corroded valve, the roots in the drain. Difficulty is what justifies the price.
- Timestamp both ends. A date on the before and the after shot turns a claim about turnaround into evidence.
Trust credentials belong on screen as text rather than in the voiceover: registration numbers, insurance, guarantees, years trading. They are checkable, which is the point.
Worth conceding: for genuine emergencies, search intent usually beats video. Somebody with a burst pipe is typing, not scrolling, and your money is better spent being findable. Video earns its place on considered jobs: kitchens, bathrooms, landscaping, solar, dentistry, extensions, anything where the customer researches for weeks before calling.
Is the free plan actually enough to start?
Yes, and this is not a soft-sell. The free plan gives 5 video ad variations plus one AI-generated video of up to 10 seconds, with no card required and no watermark on the export. For a local service business, five variations covering four different arguments is a complete first test set, not a trial.
Run those five for two to four weeks at whatever you normally spend. If they work, one standard variation costs one credit on a paid plan, so Starter at €29 for 15 credits covers roughly a year of quarterly refreshes at a small local budget, and Growth at €99 for 65 credits is aimed at teams testing weekly rather than at a single-van business. Credits never expire, which matters for a trade with a seasonal pattern: buy in spring, spend through summer, sit idle in January. The pricing page lists the credit rates, including AI-generated video at 2 credits per second, and our roundup of free video ad tools covers what else in this category has a genuinely free tier.
What we do not do, so nobody is surprised: no direct publishing to Meta or TikTok, so you will upload the files in Ads Manager or boost from the page yourself. No static banners, so display and profile graphics need something else. No AI avatars, which for a local business is fine, because your face is the asset. And no predicted performance score before you spend.
Frequently asked questions
How many video ads does a local business need?
Usually three to six. At €25 a day against an audience of about 30,000 in a small radius, weekly frequency stays below 1, so four variations can run for a month before anyone sees one more than twice. Variation counts of 12 to 20 apply to national prospecting budgets, not local ones.
What is the single most effective local service ad?
A before-and-after of a difficult job, filmed on a phone from the same angle with the dates on screen, followed by the customer saying one sentence at the door. It answers the two questions a local buyer has: can they do it, and can I trust them in my house.
How often should local ads be refreshed?
Quarterly, or when the seasonal job mix changes. Fatigue is driven by frequency, and at small budgets in a small radius it takes months to reach the exposure levels where creative decline sets in. Refresh sooner only if you are spending above roughly €100 a day.
Can you run local video ads for free?
You can produce a first set for free. Genyad's free plan includes 5 variations and one AI-generated video of up to 10 seconds with no card and no watermark, which is enough for a real test. The media spend itself still costs whatever you choose to put behind it.