Meta Ads Overview

Search ads reach people who are already looking. Meta ads on Facebook, Instagram, Messenger and Meta's partner apps reach people before they search, based on who they are, where they live and what they are interested in. For a new café in Kochi or a boutique in Jaipur, that can create demand that search alone cannot.

In this lesson you will learn how Meta's Ads Manager is structured, how to choose an objective, the main audience types, placements, how conversion tracking works with the Meta Pixel and Conversions API, and the rules that apply to sensitive ad categories.

Three levels: campaign, ad set, ad

Meta Ads Manager organises every campaign in three levels:

One campaign can hold several ad sets (for example, two audiences), and each ad set can hold several ads (for example, a reel and a carousel). That makes it easy to compare.

Choosing an objective

The objective tells Meta's delivery system what to optimise for. Ads Manager currently groups objectives into six simple choices: Awareness, Traffic, Engagement, Leads, App promotion and Sales. Meta renames and regroups objectives from time to time, so check the list in Ads Manager when you create a campaign.

The most common beginner mistake is choosing Traffic or Engagement when the business really wants enquiries or sales. Meta will then find people who click or like, which is not the same as people who buy. Choose the objective that matches the result you will actually measure:

• Want enquiries through a form or messages? Leads (or Engagement with a messaging destination for chats).
• Want online orders? Sales, with conversion tracking set up.
• Opening a new outlet and want local people to know about it? Awareness.

Audiences

• Location, age and other basics: for a local business, the radius around the shop is the most important setting.
• Detailed targeting: interests and behaviours that Meta infers. Useful as a hint, but Meta may show ads beyond these when it predicts better results, depending on your settings.
• Custom audiences: people who already know you, such as website visitors (from the Pixel), people who engaged with your Instagram account, or a customer list you upload. Upload only contacts who have agreed to hear from you; Meta matches them in hashed form, and the DPDP Act and Meta's terms still apply.
• Lookalike audiences: new people similar to a source audience such as your customers.
• Advantage+ audience: Meta's automated targeting, where your inputs act as suggestions.

Placements

Placements are where ads appear: Facebook and Instagram feeds, Stories, Reels, Messenger, Meta's Audience Network of partner apps, and more. Advantage+ placements (automatic) lets Meta spread the budget across placements where results are cheapest. Prepare creatives in vertical (9:16) and square formats so they look right everywhere; a landscape video squeezed into Reels looks careless.

Measuring results: Pixel and Conversions API

The Meta Pixel is a piece of JavaScript on your website that sends events such as page views, leads and purchases to Meta. The Conversions API sends the same events from your server, which is more reliable when browsers block scripts. Many sites use both, with an event ID so Meta can remove duplicates. Copy the base Pixel code from Events Manager; after it is installed, a page sends standard events like this:

<script>
  // after a successful enquiry (the base Pixel code must already be on the page)
  fbq('track', 'Lead', { content_name: 'festive-collection-enquiry' });
</script>

Lead is one of Meta's standard event names, like Purchase and Contact. As with GA4, do not put names, phone numbers or emails into event parameters, and tell visitors about tracking in your privacy notice. Also add UTM tags to your ad links so GA4 can compare Meta with other channels (see UTM Parameters & Campaign Links).

Reading ad set results

Two Meta-specific metrics to watch:

• Reach: how many different people saw the ad.
• Frequency: impressions ÷ reach, the average number of times each person saw it. When frequency climbs and results fall, the audience is tired of the creative.

ad_sets = {
    # name: (spend, impressions, reach, leads) - example numbers
    "15 km, women 22-45": (7_000, 52_000, 13_000, 35),
    "Lookalike of customers": (7_000, 61_000, 30_500, 28),
    "Instagram engagers": (3_000, 30_000, 5_000, 12),
}

for name, (spend, impressions, reach, leads) in ad_sets.items():
    frequency = impressions / reach
    cpl = spend / leads
    warn = "  <- high frequency, refresh creative" if frequency > 5 else ""
    print(f"{name:24} frequency {frequency:4.1f}  cost per lead ₹{cpl:4.0f}{warn}")

15 km, women 22-45       frequency  4.0  cost per lead ₹ 200
Lookalike of customers   frequency  2.0  cost per lead ₹ 250
Instagram engagers       frequency  6.0  cost per lead ₹ 250  <- high frequency, refresh creative

The frequency limit of 5 in this program is a rule of thumb for this exercise, not a Meta rule. Small, warm audiences such as engagers fill up quickly; plan new creatives every few weeks.

Rules and transparency

• Advertising standards. Every ad is reviewed against Meta's Advertising Standards. Misleading claims, before-and-after body images in some categories, and prohibited products get rejected, and repeated violations can restrict the ad account.
• Special ad categories. Ads about credit, employment, housing, and social issues, elections or politics must be declared as a special ad category. Targeting options are limited for them, and political and social issue ads in India need authorisation and a "Paid for by" disclaimer.
• Ad Library. Meta's public Ad Library shows ads currently running from any Page. It is a useful, legitimate way to study what competitors advertise.
• Boost vs Ads Manager. The "Boost" button on a post is a quick shortcut with fewer options. Ads Manager gives full control over objectives, audiences and tests.

Python: split a budget by results

A simple way to shift budget is to give more to ad sets with lower cost per lead, while keeping a minimum for testing:

results = {"15 km, women 22-45": 200, "Lookalike of customers": 250, "Instagram engagers": 250}
daily_budget = 1_500
minimum = 300

weights = {name: 1 / cpl for name, cpl in results.items()}
spare = daily_budget - minimum * len(results)
total_weight = sum(weights.values())
for name, w in weights.items():
    share = minimum + spare * w / total_weight
    print(f"{name:24} ₹{share:,.0f} per day")

15 km, women 22-45       ₹531 per day
Lookalike of customers   ₹485 per day
Instagram engagers       ₹485 per day

The three shares add up to ₹1,501 because of rounding; in Ads Manager you would round to whole amounts yourself. Make such changes gradually and only after enough results, because a few leads can swing cost per lead a lot.

Common mistakes

• Wrong objective, such as Engagement when you need leads.
• No Pixel or conversion tracking, so Meta optimises blind.
• One tired creative running for months at high frequency.
• Uploading phone lists without consent, which breaks Meta's terms and risks legal trouble.
• Undeclared special ad category for jobs, loans or political content, which leads to rejection.

Key points

• Meta campaigns have three levels: campaign (objective), ad set (audience, budget, placements) and ad (creative).
• Choose the objective that matches the result you measure: Awareness, Traffic, Engagement, Leads, App promotion or Sales.
• Audiences include location and detailed targeting, custom audiences, lookalikes and Advantage+ audience.
• The Meta Pixel and Conversions API send events such as Lead and Purchase; never send personal data in parameters.
• Watch reach, frequency and cost per result; declare special ad categories; use the Ad Library to research.

spend, impressions, reach, bookings = 4_000, 24_000, 8_000, 20
print(f"Frequency: {impressions / reach:.1f}")
print(f"Cost per booking: ₹{spend / bookings:.0f}")

Frequency: 3.0
Cost per booking: ₹200