Tip 1: stop boosting, start structuring
The most common Facebook ads history we inherit: a couple of years of boosted posts, decent reach, warm feelings, and no idea whether any of it sold anything. Boosting posts is easy. So is guessing. Only one of them is a strategy.
Boosting was not wrong, it just was not built for the job most businesses want it to do. A proper campaign has the parts boosting skips: a conversion objective, a defined audience, creative variants, a landing page with one job, and reporting that connects spend to enquiries or sales.
If you take one tip from this post: run campaigns in Ads Manager with a conversion objective, and let the Boost button retire with dignity.
Tip 2: treat creative as the targeting
Meta’s delivery system decides who sees your ad largely based on who responds to it. That means the creative (the hook, the image, the first three seconds of video) is doing the audience selection. An ad that speaks to everyone gets shown to no one in particular.
One ad concept is not a test; it is a guess with a daily budget. Real creative testing means genuinely different angles, not the same image with three background colors:
- Different problems. Price angle vs quality angle vs speed angle.
- Different formats. Static, short video, carousel, plain-text “ugly” ad.
- Different proof. Customer voice, before/after, numbers, demonstration.
Give each round enough budget to produce a real signal, kill the losers without sentiment, and feed what you learned into the next round. The learning compounds; individual ads do not.
Tip 3: broad usually beats clever
Detailed interest stacking, the old craft of layering “small business owners” with “likes golf”, mostly lost its edge. Meta’s Advantage+ products find buyers from conversion signals better than manual interest guesses in most accounts we see.
- Go broad by default, and let the creative and the conversion event steer delivery.
- Keep one genuine constraint: geography. Serve where you actually sell.
- Retargeting still earns its keep for considered purchases, but it needs enough traffic to matter. A retargeting audience of 400 people is a rounding error, not a campaign.
Tip 4: feed the machine honest signals
The algorithm optimizes toward whatever event you tell it to. Choose the event closest to money that still fires often enough to learn from, for most service businesses that is a lead form or enquiry, for stores it is purchases, not add-to-carts.
And make sure the signal survives the journey. Browser-only tracking loses a chunk of conversions to blockers and privacy features; the Conversions API sends events from your server as well, so Meta learns from a fuller picture. If lead quality varies wildly, pass the good ones back as a separate event, the machine cheerfully optimizes toward junk leads if junk is what it can see.
Meta optimizes toward what it can measure, not what you meant. Broken or shallow signals are instructions to spend badly.
Tip 5: budget by math, and know when to stop
Work backwards: what a customer is worth, what you can pay per lead or sale, and roughly fifty conversions a week for the account to learn well. If your budget cannot buy meaningful data at your real cost per result, better to concentrate it (one campaign, one audience) than to sprinkle it across five.
- Do not spend more when the landing page leaks. Clicks that land on a slow, vague page are a rental payment on confusion. Fix the page first.
- Do not spend when follow-up is broken. Leads that wait a day go cold. Cheaper leads do not fix that; faster follow-up does.
- Do not scale winners violently. Big overnight budget jumps reset learning. Move in steps, watch cost per result for a few days between them.
- Do consider whether Facebook is even the channel. Urgent, high-intent needs (burst pipe, locked out) live on Google search. Meta creates demand; it is slower at harvesting it.
If you would rather someone did all of this for you (the structure, the testing cadence, the signal plumbing) that is precisely what our Facebook Ads management covers.
Frequently asked questions
Why is boosting posts not enough?
Boosting optimizes for engagement (likes, comments, reach) not for enquiries or sales. It also skips the structure that makes ads improvable: conversion objectives, creative variants, defined audiences and proper reporting. It is fine for visibility on a good post; it is not a customer acquisition system.
How much should a small business spend on Facebook ads?
Enough to generate meaningful conversion data at your real cost per result, as a rule of thumb, aim for a budget that can produce dozens of conversions a week on the event you optimize for. If that is out of reach, concentrate spend on one campaign and one offer rather than spreading it thin.
How many ad creatives should I test?
Test genuinely different angles rather than many near-duplicates, three to five distinct concepts per round beats fifteen recolored versions of one idea. Let each round gather enough results to judge, keep the learnings, and feed them into the next round of creative.
Do detailed targeting and interest audiences still work?
They still exist, but broad targeting with strong creative and a clean conversion signal outperforms manual interest stacking in most accounts now. Meta’s delivery system selects the audience from who responds; your creative and conversion event are the real targeting levers.
What is the Conversions API and do I need it?
It sends conversion events to Meta from your server rather than relying only on the browser pixel, which loses events to ad blockers and privacy features. If you spend meaningful money on Meta, yes, fuller data means the algorithm learns from more of your actual results.
When are Facebook ads the wrong channel?
When demand is urgent and high-intent (emergency trades, same-day services) search captures existing demand better. Also when the landing page or follow-up process is broken: ads amplify what you have, so fixing the page and the lead handling comes before spending more on traffic.