Learn · How-To

How to Measure Social Media ROI Including Dark Social

Measuring the return on organic social is the single most-cited problem in marketing, named by over 70 percent of marketers surveyed globally in 2026. It is genuinely hard, and it is harder in South Africa than in most markets, because so much sharing happens in WhatsApp — which strips referrer data and lands in your analytics as direct traffic. This guide will not give you perfect attribution, because it does not exist. It will give you a measurement setup you can defend in a board meeting.

1. 1

   Define what return means before you measure anything

   Decide up front whether social is accountable for pipeline and revenue, or for reach and brand awareness. Both are legitimate; measuring one while being judged on the other is what makes social reporting feel like an argument. Write the definition down and get whoever signs the budget to agree to it before you build any tracking.

2. 2

   Build a UTM taxonomy and write it down

   Agree a fixed convention for utm_source, utm_medium, utm_campaign and utm_content — all lowercase, no spaces, hyphens not underscores, and a controlled vocabulary for each field. Source is the platform, medium distinguishes organic social from paid social, campaign matches your campaign name everywhere, and content identifies the specific post or creative. Inconsistent casing alone will split one campaign into four rows in your reports.

3. 3

   Enforce the taxonomy with one shared builder

   Put the taxonomy in a shared spreadsheet with dropdown-constrained fields that assembles the URL for you, and require every link to come from it. Taxonomies do not fail because people disagree with them; they fail because someone builds a link by hand at five o'clock on a Friday. Removing that possibility is the whole job.

4. 4

   Understand why your analytics under-reports social

   A large share of social-driven traffic arrives with no referrer and is recorded as direct. This is dark social — links copied and pasted into WhatsApp, sent in DMs, shared in Slack, or opened inside an app in-app browser. In a WhatsApp-dominant market this is not an edge case; it is likely a substantial fraction of your real social traffic sitting in the wrong bucket.

5. 5

   Add self-reported attribution to your forms

   Add a required How did you hear about us field to every lead form, with a short controlled list of options rather than a free-text box. This is the single highest-value measurement change most businesses can make, because it captures exactly what click-based tracking cannot. Compare it against your analytics attribution monthly — the gap between the two is roughly the size of your dark social problem.

6. 6

   Handle the in-app browser problem

   Links opened inside TikTok, Instagram or Facebook load in an in-app browser with its own storage, so a visitor who returns later in a real browser looks like a brand new user. Expect inflated user counts and depressed conversion rates from social sources, and check that your consent banner and forms actually function inside those in-app browsers — they frequently do not.

7. 7

   Track the few platform metrics that correlate with intent

   Ignore impressions and follower counts. Track saves and shares, which indicate the content was worth keeping or passing on, and profile-to-website taps, which indicate someone chose to leave the platform for you. These are the on-platform metrics most closely associated with commercial intent, and they are comparable across time even when reach fluctuates.

8. 8

   Report one blended view, with its limitations stated

   Present three numbers side by side: last-click attribution from your analytics, self-reported attribution from your forms, and branded search volume over time as a proxy for awareness. State plainly that the three will not agree and explain why. A report that shows the disagreement and accounts for it is far more credible than one precise number that quietly excludes every channel you cannot track.