How to track conversions in paid social campaigns

Define a conversion that matters to the business

A conversion is a defined target action, not any activity after an ad impression. For a service, it might be a successfully submitted enquiry; for a shop, a confirmed purchase. Button clicks, form opens and submitted contact details are separate events. Combining them inflates results without showing how many people enquired.

Before starting, agree on the name of the event, the condition for its activation and the method of verification. It is better to confirm the enquiry after successfully receiving the data, and not when pressing a button. A test opening of the thank you page does not in itself prove that the contact has been delivered to the team.

Separate website forms, Direct messages and platform lead forms

These scenarios have different measurement points. You can check the submission of the form on the website. In Direct, it is important to distinguish the beginning of a dialogue from a relevant request. An embedded form can provide a contact that still needs to be qualified. Don't compare them just by price if the value of the result is different.

Record each enquiry's source and contact method in your spreadsheet or CRM. Add a quality status: relevant, duplicate, spam, unreachable or sold. For destination selection, see our analysis of the site, Direct and forms.

Check events before spending

Test the visitor journey on a phone: open the page, submit the form and confirm receipt. Also test a failed submission and repeated clicks. A failed submission must not count as a conversion, and reloading must not turn one enquiry into several leads.

• Check that the right event fires at the right time.
• Match the test enquiry to its record in the receiving system.
• Exclude test enquiries from business reporting.
• Do not submit message text, phone number or name in arbitrary analytics parameters.

Do not confuse the platform report with all sales

Advertising platform and sales system may show different numbers due to attribution rules, registration time, repeat enquiries and availability of signals. Start the check with the same period, time zone, and event definition. Don't declare one source "wrong" without looking at what it actually counts.

To assess channel quality, link lead qualification to sales. If this is unavailable, state the limitation. A platform event alone does not prove that a contact was delivered or a product paid for.

Check data sufficiency, not just cheap conversions

One inexpensive enquiry does not prove the stability of the campaign. Compare results over comparable time periods and consider the sales cycle. For an expensive product, the solution may come later than the ad conversion. Before changing the campaign, check whether the period has not ended before the team has had time to process the enquiry.

Don't increase the budget just because the cost of the event has decreased. Check its quality first. Perhaps the text attracts curiosity, but not the intention to buy. Or the form became simpler, but there were more random contacts.

What to agree with the contractor

Ask for a description of events, test results, attribution rules, and a sample report with qualifications. If several methods of transmission of the same event are used, the elimination of duplication is checked separately; the specific implementation depends on the chosen system. Don't consider an integration complete without testing.

Read on how to build analytics and how to calculate cost per lead. A launch plan for Facebook ads must take into account not only the ad, but also this path of obtaining the result.

Before the next step

Record your starting data and decision criteria. Calculation examples are illustrative unless explicitly linked to a case study. Results from one project do not guarantee results for another business.