Facebook Ads CCS

1How to use this guide

Who: Everyone · Time: 10 minutes

This guide takes CCS from no Facebook advertising at all to a running, measured campaign. It is written for someone who has never opened Meta Ads Manager. Nothing is assumed.

Work through it in order. Every section ends with a checklist. Do not start a section until the checklist before it is fully ticked, because each section depends on the one before.

Who does what

Each section names its role at the top. If one person does every job, that is fine. The labels still tell you what kind of work it is.

How the checklists work

• On the website version, tap a box to tick it. Your ticks are saved in your own browser on that device only. Other people do not see them.
• In the Google Doc, type an x in the box, or use Docs' own checklist: select the lines first, then Format, then Bullets and numbering, then Checklist.
• A box marked Owner decides cannot be ticked by anyone else.

2What we are building, in one picture

Who: Everyone · Time: 10 minutes

The whole campaign is one path that a stranger walks. Every later section exists to make one step of this path work, or to measure it.

The offer every ad must match

• Two free months. Two visits and two sets of photographs. No card and no charge.
• Opening December 1, 2026. Anyone who signs up before then joins a waiting list. Her first visit comes after opening.
• 10% off the enrollment service and every month after, for seniors and military. It applies if either the person paying or the person buried is a senior or served in the military.
• 60 cemeteries across King County north of Covington and southern Snohomish County. That includes cemeteries in and around Seattle, Bellevue, Kirkland, Bothell, Redmond, Kent and Everett. The full list of 60, with a map, is in the Service Areas section of the CCS home page (ccs2.akroadside.com). Keep it open when you call leads.
• After the free months: a $150 enrollment service, then Essential at $40, Remembrance at $50 or Complete at $60 a month.

3The words you will see

Who: Everyone · Time: 15 minutes

Meta uses a lot of jargon. These are the terms this guide uses, in the order you meet them. Come back here whenever a word is unclear. The full glossary at the end has more.

4Decisions only the owner can make

Who: Owner · Time: One sitting

Seven questions block the work. Nobody else can answer them, and guessing wrong costs money or creates legal risk. Answer them first and write the answers in the "Decision" column of the decision log at the end of this section.

1. Where does the customer live?

The grave is always inside the CCS service area. The woman paying may not be. There are two very different customers:

Recommendation: do not guess. Run both as two ad sets (two audiences, each with its own $13 a day, inside one campaign) with equal money for three weeks, and let the cost per lead decide (sections 16 and 18). The rest of this guide assumes you do that.

2. Who appears in the videos?

The scripts are written as a customer telling her own story: "So I hired somebody. They sent me pictures." CCS opens December 1, so there are no customers of this service yet who could say those words truthfully.

Recommendation: option C for launch, then switch to option A as real December customers agree to film. Every before-and-after photo in every video must be real CCS work, whichever option you choose.

3. The privacy policy

The CCS site has no privacy policy. The footer (the strip of small links at the very bottom of every page) has a Privacy link, but it goes nowhere. Meta's terms require a privacy policy for any site with a Pixel, and without one Meta can refuse the ads, so this blocks launch. Section 8 has the steps. The owner's decision is simply to approve the text.

4. The budget

The recommended first test costs about $26 a day for 21 days, about $550 in total: two ad sets at $13 a day each. Section 18 explains why less than that teaches you almost nothing, and what doubling it buys.

5. Capacity

How many first visits can the crew do in December? If the ads produce 80 signups and the crew can do 30, fifty people wait and some leave angry. Pick a number. When the waiting list reaches it, pause the ads. Pausing is safe: in Ads Manager, click the blue switch next to the campaign so it turns grey. Nothing is lost, and flipping it back restarts the ads (section 24).

6. Who calls the leads, and how fast

Name one person. Leads go cold within hours. A lead called in the first hour is far more likely to become a customer than one called the next day (section 20).

7. Launch date

Ad prices in late November are some of the highest of the year, because every store is buying ads for Black Friday (November 27, 2026). Two sensible choices:

• Launch about November 2. The three-week test finishes around November 22. Its last few days overlap the start of holiday prices, but both ad sets pay the same higher prices, so the comparison stays fair. The waiting list fills ahead of the December 1 opening.
• Launch December 1, the day CCS opens. The ads can then say "now open", but the first two weeks run during expensive December traffic.

Decision log

5Accounts and access

Who: Owner, then ads operator · Time: 1 to 2 hours, once

Do this once, in this order. The most common disaster with Meta ads is not a bad ad. It is being locked out of your own account, and Meta support can take weeks to restore access. Most of these steps exist to prevent that.

1. Your own Facebook profileTurn on two-factor authentication. This means that logging in needs your password plus a code from your phone, so a stolen password alone is not enough. On facebook.com, click your profile picture at the top right, then Settings and privacy, then Settings, then Accounts Center, then Password and security, then Two-factor authentication. Choose text message if unsure (Meta texts you a code each time). The other choice, an authenticator app, is a free phone app such as Google Authenticator that shows a new 6-digit code every 30 seconds. Meta blocks ad access for admins without two-factor.
2. business.facebook.comClick Create account (the button may say Create a business portfolio). Enter the business name (Cemetery Cleaning Services), your name, and a business email you will keep for years.
3. Settings, then PeopleIn business.facebook.com, click the gear icon (Settings) at the bottom left, then People. Click Add and invite a second person with full control, which means they can do everything you can, including adding and removing people. If one person is locked out, the other can still reach everything and add them back.
4. Settings, then Accounts, then PagesAdd the CCS Facebook Page, or create one here. Fill in the logo, a cover photo, the phone number (206) 312-0001, the website (https://ccs2.akroadside.com until CCS moves to its own address) and the service area.
5. Settings, then Accounts, then Ad accountsClick Add, then Create a new ad account. Time zone: Pacific Time (Los Angeles). Currency: USD. Neither can be changed later. If either is wrong you must make a new account.
6. Same screenClick the new ad account, then Assign people. Tick your name and the second admin's, choose Full control, and click Save.
7. Settings, then Billing and paymentsAdd a payment card. Then set an account spending limit, a hard cap on total spend, a little above the test budget. $700 is about right for the $550 test. It is a seatbelt against a typo like $130 a day instead of $13. When the limit is reached, every ad stops at once, so on day 21 raise it before you raise any budget: add about $30 for each extra day you plan to run.
8. Settings, then Accounts, then Instagram accountsOptional. You need the CCS Instagram username and password. Linking the account lets the ads show under the CCS name on Instagram. Without it, they still run on Instagram under the Facebook Page's name.

6The Pixel, explained from zero

Who: Everyone should read this · Time: 15 minutes

This is the most important idea in the guide. If you understand this section, most of the other rules make sense.

Why it matters so much

Meta does not show an ad to everyone in the audience. There are millions of women 45+ in the US, and a $13 budget reaches a few thousand of them a day. Meta picks which few thousand. It picks based on what you told it you want, and on what it has learned from the people who already did that thing.

Meta can only learn what a signup looks like if someone tells it every time a signup happens. The Pixel is the only thing that tells it. The landing page form is on the CCS website, not inside Facebook, so without the Pixel Meta never finds out who signed up.

The example numbers are illustrations, not promises. The pattern is real. More clicks and fewer signups is exactly what optimizing for clicks tends to produce.

What the CCS site already reports

7Switching the Pixel on and proving it works

Who: Ads operator and developer · Time: 1 hour

Create the Pixel and get its ID

1. business.facebook.com, then Events ManagerTo get there, click the menu icon (three lines, top left), then All tools, then Events Manager. Click Connect data sources (or the green plus), then Web (meaning a website, as opposed to a phone app). Name it CCS website. Click Create.
2. The "install" screen that followsDo not paste any code. The site already has it. Close this screen.
3. Events Manager, CCS website, SettingsClick CCS website in the list on the left, then the Settings tab across the top. Copy the dataset ID, a number about 15 or 16 digits long.
4. DeveloperPaste to Claude: "Set the CCS Pixel ID to [the number] and redeploy the CCS site." Redeploy means publish the updated website so the change is live on the internet. You get back: a message saying the site is live with the Pixel on. If the number has a typo in it, the site refuses to publish and Claude tells you, so a wrong ID cannot go live.
5. Events Manager, Settings, Automatic advanced matchingTurn it on, and tick email, first name and last name. This lets Meta connect a signup to the right Facebook account more often, which makes its learning and its counts more accurate.
6. Settings, Track events automatically without codeTurn it off. Otherwise Meta invents events from button taps. They clutter the reports, and they also show up as choices when you build the ad set, where one could be picked by mistake. Lead is the only event that should count as a result.
7. Settings, Traffic permissionsChoose Allow only and add ccs2.akroadside.com (the address of the CCS website, typed with nothing before or after it). This stops other websites from sending fake events into the CCS Pixel. Add the main CCS domain later if the page moves there.

For reference, this is what Claude does when you ask. Owner and ads operator: you do not run this.

# Developer: in D:\Pycharm\Amicus1\.env add one line
FB_PIXEL_ID=<the dataset ID, digits only>

# then rebuild and deploy the CCS site (deploy runs the build first)
bash ccs2_site/deploy.sh

Prove it works before spending a dollar

1. On a computer, open the Chrome browser. An extension is a small free add-on to the browser. Go to chromewebstore.google.com, search Meta Pixel Helper, and click Add to Chrome. Its icon appears under the jigsaw-piece icon at the top right of Chrome. Click the pin next to it so it stays visible.
2. Open the landing page. Click the Pixel Helper icon. It should show the CCS Pixel ID and one PageView with a green tick. If it shows nothing, the ID is not set or the site was not redeployed. Paste to Claude: "Pixel Helper shows nothing on the CCS welcome page."
3. In Events Manager, click CCS website, then the Test events tab. Type the landing page address and click Open website. This opens the page in a way that shows its events live.
4. Fill in the form with the first name TEST, last name TEST, and an email inbox you can check. Submit it.
5. Within a minute, the Test events screen should list Lead. That is the moment the Lead option unlocks in Ads Manager. It can take up to an hour to show there.
6. Check that inbox for the "you're on the list" email, and check info@ for the signup notice. That proves the server side works too.
7. Paste to Claude: "Delete the TEST signups from the CCS intake database." You get back how many rows were deleted. It should be 1. This stops the test counting as a real lead.

# Developer: delete the test signup
bash ccs2_site/d1q.sh "DELETE FROM intake WHERE first_name='TEST' AND last_name='TEST'"

8The privacy policy and domain verification

Who: Owner, developer, ads operator · Time: 2 hours

What the privacy page must say

• Who CCS is and how to contact it: email and phone.
• What the form collects: her name and email, and optionally her loved one's name, cemetery, plot and date of passing.
• Why: to contact her about the service and schedule visits.
• That the site uses the Meta Pixel, which shares page visits and signups with Meta to measure and improve advertising. Name it plainly.
• That CCS does not sell her information.
• How to ask CCS to delete her information: an email address that someone reads.
• The date it was last updated.

1. DeveloperPaste to Claude: "Draft the CCS privacy page from the privacy section of the Facebook Ads CCS guide." You get back the text to read.
2. Owner decidesRead the draft. Reply to Claude with any changes, or "Approved, publish it at /privacy and link it from every footer and under the landing page form."
3. DeveloperYou get back a message that ccs2.akroadside.com/privacy is live. Open it on your phone, and tap the Privacy link at the bottom of the home page to check it arrives there.

Domain verification

1. Settings (the gear icon), then Brand safety and suitability, then DomainsClick Add and type akroadside.com. The CCS site lives on akroadside.com for now, so that is the address to verify.
2. Choose a methodPick Meta-tag verification. Meta shows one line of code. Copy it.
3. DeveloperPaste to Claude: "Add this Meta domain verification tag to the home page of akroadside.com itself (not only ccs2.akroadside.com) and redeploy: [the line]." Meta checks the top address, akroadside.com. You get back a message that it is live.
4. Back in DomainsClick Verify. It should turn green. If not, wait 10 minutes and try again.

9Two scoreboards: knowing the numbers are true

Who: Ads operator and developer · Time: 5 minutes a week

Every signup is counted in two separate places. You need both: Meta's number is the one Meta learns from, and ours is the true one. Comparing them once a week is how you know whether Meta is seeing what really happens.

So Meta's number is normally a little lower than ours. A small gap is normal. A big gap is a problem, for two reasons.

• The ads look worse than they are. Meta reports 6 leads at $43 each when the truth is 10 leads at $26 each. You might switch off an ad that is actually working.
• Meta learns from fewer examples. Every missed Lead is a signup Meta never studies, so it gets slower at finding the next one.

How to do the check

1. Pick the same date range in both places, for example last Monday to Sunday. Expect a small mismatch at the edges of each week: Meta dates a lead by the day of the tap on the ad, and our database dates it by the day of the signup, which can be a few days later.
2. In Ads Manager, read the total in the Results column for the campaign. Write it down as Meta's count.
3. Paste to Claude: "How many fb-welcome signups were there from [Monday's date] to [Sunday's date], Pacific time?" You get back one number. That is our count. (The form tags each signup with where it came from. Signups from the ad page get the tag fb-welcome.)
4. Work out the gap: our count minus Meta's count, divided by our count, times 100 for a percent. Example: 40 minus 34 is 6, and 6 divided by 40 is 0.15, which is 15%.
5. Write both numbers in the weekly report (section 23).

# Developer: signups from the Facebook landing page in a date range
bash ccs2_site/d1q.sh "SELECT COUNT(*) FROM intake WHERE source='fb-welcome' AND submitted_at >= '2026-11-02' AND submitted_at < '2026-11-09'"

10Conversions API: the backup messenger

Who: Developer · Time: About a day of developer work

When to build it: not before launch. Launch on the Pixel alone. Build it when the weekly count check (section 9) shows a gap above 15% two Mondays in a row, or above 25% in a week with 20 or more leads, or before spending more than the first test budget, whichever comes first. To ask for it, paste to Claude: "Build the Meta Conversions API for the CCS signup form, as described in the Facebook Ads CCS guide." You get back a message that it is live. Then Events Manager should show Browser and Server, as in the last row below.

11Why the video matters more than anything else

Who: Everyone · Time: 5 minutes

In 2026, Meta does most of the targeting by itself. You tell it "women 45+", and it decides which women 45+ see the ad. What it uses to decide is who stops scrolling and who keeps watching. So the first three seconds of the video do much of the targeting.

A woman who has a parent buried in the Seattle area sees a mossy, unreadable headstone, then the same stone clean and bright. She stops. Someone with no grave to care for scrolls past. Meta notices the difference and finds more people like the first woman.

12Which scripts to film first, and the edits they need

Who: Owner and video maker · Time: 1 hour

The scripts are in the Google Doc "CCS scripts v1", in the CCS shared Drive folder (search Google Drive for its name): 9 in a short, choppy style and 16 in a relaxed, conversational style. That is too many to film at once. Start with six, chosen to give each audience three strong stories and to keep away from claims that might not be true.

The first six

Hold these for later

• 04 Mother's Day (both styles): film it about six weeks before, in late March, and run it from about April 20 to May 10, 2027. Running it in November looks odd.
• Memorial Day, Father's Day, Veterans Day and the anniversary scripts: the same idea. Seasonal scripts run in the two or three weeks before the date.
• The choppy versions: use them in the second test (section 18), filmed with the same woman, to learn whether short and broken-up beats relaxed and conversational.

13Filming: the shot list

Who: Video maker · Time: Half a day for nine videos: six for launch, three spares

Film nine scripts at the first shoot: the six above for launch, and three spares from the list so that fresh videos are ready (section 25). Each video is built from four kinds of shot. Film all four for every script. The editor needs choices.

Camera and sound settings

• Film on a recent phone held vertically, at 1080p or 4K, 30 frames per second. On an iPhone: Settings, Camera, Record Video, choose 4K at 30 fps. On Android the same choice is in the camera app's own settings. Wipe the lens first.
• Lock the focus and exposure on her face (tap and hold on most phones), so the picture does not pulse.
• Sound matters more than picture. Use a small clip-on microphone that plugs into the phone (about $20 to $40). Film in a quiet room with soft furnishings, not a kitchen with hard walls.
• Leave about a hand's width of empty space above her head, and more below her chin. Section 14 explains why: the app covers the top and bottom of the screen.
• Film each script three to five times. Natural slips are fine. Reading off a card is not. She should know the lines well enough to say them like a person talking.
• Keep her clothing plain. No logos, no stripes, no busy patterns.

14Editing: one recipe for every video

Who: Video maker · Time: 1 to 2 hours per video

Every video follows the same recipe. Keeping the recipe fixed is what makes the tests in section 18 fair: when only the story changes, a difference in results is about the story.

Technical settings

Safe zones

• Keep all text and her face out of the top 14% of the frame (about the top 270 pixels of a 1920 pixel tall video).
• Keep them out of the bottom 35% (about the bottom 670 pixels).
• Keep them a little away from the right edge, where the icons sit.
• When you build the ad (section 16), the preview on the right shows where the app's buttons sit. Check every video there before publishing. Nothing people need to read should be under them.

Name every file the same way

File names become ad names (without the size and without .mp4), and ad names are how you read results. A consistent name lets anyone see at a glance which script, style and size an ad is.

CCS_<script number>_<short name>_<style>_<size>_v<version>.mp4

CCS_01_CantBeThere_Conv_916_v1.mp4
CCS_01_CantBeThere_Conv_45_v1.mp4
CCS_08_Overseas_Conv_916_v2.mp4

15The words next to the video

Who: Ads operator, approved by owner · Time: 30 minutes

Each ad has three text fields. Meta lets you enter up to five versions of each (in the ad, under Primary text, click Add text option for each extra version), and it tries combinations. Give it three good ones.

Ready-to-use lines, taken from the landing page

The ad and the page should sound like one voice, so tapping feels like turning a page rather than arriving somewhere new. All of these lines come from the welcome page or match it.

16Building the campaign, click by click

Who: Ads operator · Time: 1 to 2 hours

Before you start, make the exclusion audience in section 17 ("Make the exclusion audience before launch"). This section uses it. Then open Ads Manager: go to adsmanager.facebook.com, or in business.facebook.com click the menu icon (three lines, top left), then Ads Manager. Click the green + Create button. Meta's defaults are tuned for large advertisers with big budgets. Several of them are wrong for a small test, so every setting below says what to choose and why.

Campaign level: the goal

1. Choose a campaign objectiveLeads.
2. Choose a setupPick Manual leads campaign, not Advantage+ leads campaign. The automatic one ignores the age and gender limits this guide relies on. If you do not see this choice, Meta has renamed it: pick the option that lets you set age and gender yourself.
3. Campaign nameCCS_FB_Leads_Welcome_2026-11. A name you can sort and search later.
4. Special ad categoriesLeave empty. These are for housing, jobs, credit and politics. CCS is none of them. Ticking one by mistake removes age and gender targeting.
5. Advantage+ campaign budgetOff. When it is on, Meta moves money between the ad sets as it likes, and one usually gets almost everything. For a fair test, each ad set needs its own fixed budget.
6. A/B testOff. Section 18 explains when to use it.

Ad set level: who, where and how much

Build the first ad set (Away) fully, then Duplicate it to make the second (Local) and change only the location. To duplicate, hover over the ad set name in the list on the left, click the three dots, then Duplicate, and choose Original campaign. That guarantees every other setting is the same.

1. Ad set nameCCS_Away_W45+ and later CCS_Local_W45+.
2. Conversion locationWhere the signup happens. Choose Website. The form is on the CCS site.
3. Performance goalMaximize number of conversions. A conversion is the action you chose, here a Lead.
4. Dataset and conversion eventDataset CCS website. Event Lead. If Lead is greyed out, Meta has not received a test Lead yet. Go back to section 7.
5. Cost per result goalLeave it empty. Setting a target this early usually stops the ad from spending at all.
6. Daily budget$13.
7. ScheduleStart: the launch date. End: none. You will stop it by hand after the test.
8. Audience: locationSection 17 says exactly what to enter for each ad set.
9. Audience: switch to original optionsClick Switch to original audience options, a small blue link under the audience box. If Meta asks you to confirm, click Switch. Then set Age 45 to 65+ and Gender: Women. With the newer Advantage+ audience, age and gender are only suggestions and Meta will show the ad outside them. The original options make them hard limits.
10. Audience: excludeUnder Custom audiences, Exclude the audience "CCS leads 180 days". If it is not in the list, stop: make it first (section 17), then come back. Excluding it stops paying to show ads to people who already signed up.
11. Audience: Advantage detailed targetingOff, if it appears. Leave detailed targeting (the box for interests like "gardening") empty. The video finds the right people.
12. PlacementsA placement is a place where the ad can show. Under Placements, choose Manual placements, then untick boxes. Keep Facebook Feed, Facebook Reels, Facebook Stories, Instagram Feed, Instagram Reels and Instagram Stories. Untick everything else, especially Audience Network (ads inside other apps and games) and Messenger. Those produce many accidental taps from people who never read the ad.
13. Attribution settingLeave Meta's default, which counts a lead up to 7 days after a tap, or 1 day after a view (she saw the ad without tapping it, then signed up within a day). Never change it during a test, or the numbers before and after stop being comparable.

Ad level: one ad per video

1. Ad nameThe file name without the size and without .mp4, because one ad carries both sizes. For example CCS_01_CantBeThere_Conv_v1.
2. IdentityThe CCS Facebook Page, and the CCS Instagram account if it is linked.
3. Ad setupCreate ad, then Single image or video.
4. MediaUpload the 9:16 file. Then, under Media, hover over the Feeds group, click the pencil, then Replace, and choose the 4:5 file. One ad, two sizes, each shown where it fits.
5. Primary text and headlinePaste the three versions of each from section 15.
6. Call to actionSign up.
7. Website URLhttps://ccs2.akroadside.com/welcome/
8. URL parametersPaste the line below. It adds a label to the link with the campaign and ad name, so a signup can later be matched to the ad that sent her. The CCS form does not save this label yet. Paste to Claude: "Make the CCS welcome form save the utm labels from the link with each signup." You get back a message that it is live.
9. TrackingUnder Website events, tick CCS website.
10. Advantage+ creative enhancementsThese are Meta's automatic changes to your ad. The section is in the ad, below the media, with a row of switches. Open the list. Turn off: music, visual touch-ups, image animation, "relevant comments", and anything else that adds or changes things. Leave on only text improvements, if you want them. Meta otherwise may add an upbeat soundtrack to a video about a grave.
11. PreviewClick through every placement in the preview on the right. Check that nothing important is hidden by the app's buttons.
12. RepeatMake the other two ads in the ad set the same way: Duplicate the ad and swap the video and name.
13. PublishClick Publish. Ads go to Meta's review, which usually takes minutes to 24 hours. If an ad shows Rejected, see section 26.

utm_source=facebook&utm_medium=paid_social&utm_campaign={{campaign.name}}&utm_content={{ad.name}}

17Audiences

Who: Ads operator · Time: 30 minutes

For location, type the county name in the location box, choose it from the list, and use the Include / Exclude menu next to it. For Away, add United States first, then add the two counties and set each to Exclude.

Make the exclusion audience before launch

1. Menu, All tools, AudiencesIn business.facebook.com, click the menu icon (three lines, top left), then All tools, then Audiences. Click Create audience, then Custom audience, then Website.
2. SourceCCS website.
3. EventsChoose Lead. Retention: 180 days (how long a person stays on the list after signing up).
4. NameCCS leads 180 days. Click Create. It fills up by itself as people sign up.

Make the retargeting audiences on day 14, so they fill up. Start ads to them on day 28

• Website visitors: Custom audience, Website, CCS website, people who visited the landing page address in the last 30 days. Name it "CCS LP visitors 30 days".
• Video viewers: Custom audience, Video, choose "People who viewed at least 50%" of the CCS videos, last 30 days. Name it "CCS video 50% 30 days".
• Each audience shows its size next to its name. Wait until it reads 1,000 or more before using it. A smaller audience mostly will not spend. That is why retargeting waits until the main ads have run for a few weeks.

18A/B testing, done properly

Who: Ads operator, decisions by owner · Time: Planned once, run for 12 weeks

Why only one change at a time

Suppose B has a different video and a different headline, and B wins. Was it the video or the headline? You cannot tell, so you learned nothing you can reuse. Change one thing, learn one thing, keep it, then test the next thing.

Why small numbers lie

Flip a fair coin 10 times and you will often get 7 heads. That does not make the coin unfair. Leads work the same way. If ad A gets 6 leads and ad B gets 4, that can easily be luck. With small numbers, only a big difference means anything.

At $13 a day per side for 21 days, each side spends about $275. At a typical $30 to $60 per lead, that is roughly 5 to 9 leads per side. That is often not enough to decide, because the winner needs 10 leads of its own (see the decision rules). Plan from the start for a fourth week, or for $20 a day per side. Twice the leads on one side is the kind of difference this test can see. A small difference it cannot.

Two ways to run a test in Meta

The test plan, in order

Decision rules

Write the rule down before the test starts, so nobody can pick a winner afterwards because they liked it.

Mistakes that ruin a test

• Changing anything in the middle, including adding a new video. Every edit restarts learning and mixes "before" with "after". The only exception is the stop rule above.
• Stopping after two good days. Early results swing wildly. Run the full time.
• Different budgets on the two sides.
• Testing two things at once.
• Comparing a test that ran over Thanksgiving or Black Friday week with one that did not. Ad prices that week are higher for everyone.
• Not writing the result down. A month later nobody remembers which won or why.

The test log

Copy this table into a Google Sheet called CCS Facebook test log, one row per test.

19Launch day

Who: Ads operator · Time: An hour in the morning, 15 minutes at night

1. Morning: in Ads Manager, click the Ads tab. The Delivery column should say Active for every ad, not "In review" or "Rejected".
2. If an ad was rejected, read the reason (click the ad, then the warning). Section 26 has the usual causes.
3. Open the landing page from the ad on your own phone. To find your own ad, click the ad in Ads Manager, then the preview's Share button, copy the link and open it on your phone. Tap Sign up. Confirm the page loads quickly and the form works, but do not submit it.
4. Tell the lead caller the ads are live.
5. Evening: check that money is being spent on both ad sets. Do not judge results on day one.

20Answering every lead

Who: Lead caller · Time: 5 to 10 minutes per lead

Ads produce leads. Only people produce customers. A lead that waits a day has often forgotten why she signed up. Make contact the same day, within an hour when you can.

Where leads arrive

• An email to info@CemeteryCleaningServices.com for every signup, with everything she typed.
• She receives an automatic "you're on the list" email at the same moment. It comes from intake@amicusanalytics.com.
• Every signup is also saved in the CCS intake database. To get a copy, paste to Claude: "Export the CCS intake database to a spreadsheet." You get back a file you can open in Google Sheets.

The follow-up, step by step

1. Within the hour, email her, using her first name. The ad form does not ask for a phone number. If she gives one in her reply, call her.
2. Thank her, and confirm what she signed up for: two free months, starting after December 1.
3. Ask which cemetery her loved one is in, if she did not say. Check it against the list of 60 on the CCS home page.
4. If it is: note the loved one's name, the cemetery and the plot if she knows it. Tell her when the first visit will be and that photos will follow.
5. If it is not one of the 60: thank her warmly, say CCS does not reach that cemetery yet, and ask if she would like to hear if that changes.
6. Ask whether she or her loved one is a senior or served in the military, for the 10% discount.
7. Write the result in the lead tracker the same day.

The lead tracker

A Google Sheet called CCS lead tracker, one row per lead. Without it, nobody can say which ads produce customers rather than just signups.

21Comments and messages

Who: Ads operator or owner · Time: 10 minutes a day

People comment on ads, and everyone who sees the ad later sees the comments. A grave-care ad will attract kind comments, questions, and some unkind ones. An unanswered question or an ugly comment sitting under the ad lowers results.

1. Meta Business Suite, InboxIn business.facebook.com, click the menu icon, then Inbox. Or use the free Meta Business Suite phone app. This one screen shows comments on the ads and messages to the Page, from both Facebook and Instagram. Check it once a day.
2. A questionReply within the day, briefly and warmly. Point to the page: "Yes, we care for graves in 60 cemeteries in King and southern Snohomish County. You can sign up here for two free months."
3. Unkind but not abusive ("this is creepy")Leave it, or reply once, gently: "We understand it's not for everyone. For families who live far away, the photos mean a lot." Never argue.
4. Spam, scams, abuseHide it. A hidden comment stays visible to its writer and their friends, so they do not repost it angrily, but nobody else sees it. Delete only clear scams and abuse.
5. Moderation settingsIn business.facebook.com: Settings, then Page settings, then Moderation. add a short list of blocked words (profanity, "scam", common spam phrases). Matching comments are hidden automatically.

22The first 30 days

Who: Ads operator · Time: 10 minutes a day, 30 minutes on review days

The hardest part of running ads is leaving them alone. Results in the first week jump around, and every edit restarts Meta's learning. This calendar says when to look, and when to act.

23Reading the numbers

Who: Ads operator · Time: 15 minutes a week

Set up the columns once

1. Ads Manager, the Columns buttonAbove the list of ads, click the button labelled Columns: Performance. Choose Customize columns.
2. Tick theseResults, Cost per result, Amount spent, Impressions, CPM, Link clicks, CTR (link click-through rate), Landing page views, 3-second video plays, ThruPlays, Frequency.
3. SaveTick Save as preset and name it CCS weekly. Choose it every time.

What each number tells you, and what to do

The ranges are rough guides from 2026 industry reports, not targets Meta promises. Your own first month becomes the real benchmark.

Breakdowns that answer specific questions

• The Breakdown button sits next to the Columns button. Click it, then By delivery, then choose one item.
• Breakdown, By delivery, Age and Gender: confirm the ads only reached women 45+. Anything else means the ad set was left on Advantage+ audience instead of the original options.
• Breakdown, By delivery, Placement: shows whether Reels, Stories or Feed produces the leads.
• Breakdown, By delivery, Region on the Away ad set: shows which states respond. Do not narrow to them until the test is over.

The weekly report

Every Monday, send the owner five lines:

1. Spend this week and in total.
2. Leads this week, from Meta and from our own count (section 9), and the gap as a percent.
3. Cost per lead for each ad set and for the best and worst ad.
4. Leads contacted, booked and paying, from the lead tracker.
5. One thing that changed, and one thing planned for next week.

24What a lead is worth, and when to scale or stop

Who: Owner · Time: 30 minutes a month

Every decision about money comes back to one comparison: what a lead costs against what a lead is worth.

So a lead that costs $160 breaks even in the first year. Months after the first year are extra. The 10% senior and military discount lowers each figure by a tenth. Once 30 or more leads have been followed up, replace "1 in 4" with the real share: in the lead tracker, count the leads with status "paying customer" and divide by all leads.

25Keeping the videos fresh

Who: Ads operator and video maker · Time: One new video every 2 to 3 weeks

• Film more scripts than you need at the first shoot (nine for six launch videos), so new videos are ready.
• Add a new video to the winning ad set every two to three weeks, on the day one test ends and the next begins. Adding an ad restarts learning, so never add one in the middle of a test.
• Switch off the weakest ad when you add the new one, so each ad set keeps three.
• Plan the seasonal scripts: Mother's Day (May 10, 2027), Memorial Day (May 31, 2027), Father's Day (June 20, 2027). Film each one about six weeks ahead.

26When something goes wrong

Who: Ads operator · Time: As needed

27Things that are easy to forget

Who: Owner · Time: 15 minutes

• Staff and family must mark test signups. Anyone testing the form must use first name TEST, and the developer deletes those rows. Otherwise they count as real leads.
• The Meta Ad Library is public. Anyone, including competitors, can see every CCS ad at facebook.com/ads/library (search Cemetery Cleaning Services to see what the public sees). Write nothing you would not want a competitor to read.
• Phone calls from the ad. The Page lists (206) 312-0001. Some people will call instead of filling in the form. Whoever answers should know about the offer and should write callers in the lead tracker with "Ad: phone".
• Opening-day wording. On December 1, change any "opening December 1" line to "now open". December 1 falls in the middle of test 2. Change it anyway, in both A and B on the same day, and note the date in the test log.
• Spending limit. When you raise budgets, raise the account spending limit too, or the ads stop suddenly when it is reached.
• Card expiry. If the card fails, the ads stop and the account can be restricted. Add a backup card.
• Who owns the videos. If a creator films them, get it in writing that CCS can use them in ads for as long as it likes.
• Grief. Some viewers will have lost someone very recently. Keep the tone gentle and never pushy. No countdown timers, no "last chance".
• Photos of other families' graves. Only with their permission, and no readable names unless they agree.
• Holiday prices. Ad prices rise from about November 20 to early December. Do not compare results from that week with a normal week.
• Service area wording. CCS covers King County north of Covington and southern Snohomish County, not all of both counties. Say it that way on end cards and in replies, and keep the Local ad set's town exclusions (section 17).
• Review the landing page with each new test. If the Local ad set wins, the page needs a local opening line. Ask the developer.
• Instagram profile. If the Instagram account is linked, give it a profile picture, a bio and a few posts too. People tap through to check.

28The master checklist

Who: Everyone · Time: Use it all the way through

Every box in this guide, in order, in one place. Tick here or in the sections. On the website the two stay in sync.

29Glossary

Who: Everyone · Time: Reference

30Sources

Who: Reference

• Meta Business Help Center: About the Meta Pixel; Set up the Meta Pixel; Test your events; About Conversions API; About deduplication; About Event Match Quality.
• Meta Business Help Center: Verify your domain; About Advantage+ audience; About the learning phase; About A/B testing; About placements; About ad fatigue.
• Meta Advertising Standards: Personal attributes. Meta Business Tools Terms.
• US Federal Trade Commission: Trade Regulation Rule on the Use of Consumer Reviews and Testimonials (effective October 21, 2024). FTC Guides Concerning Endorsements and Testimonials.
• CCS website build notes (ccs2_site/README.md): the Pixel placeholder, the Lead events, the intake form and database.
• CCS scripts v1 (Google Doc): the 25 UGC scripts.
• Benchmark ranges: 2026 industry reports compiled in the CCS Paid Media Field Guide (paidmedia.akroadside.com).
• Menus in Meta are renamed often. If a path in this guide has moved, Meta's help center search will find the new one.