← The Attribution Brief

Why Affiliate Attribution Fails on South African Mobile (CGNAT)

2 July 2026 · 6 min read

If your affiliate or influencer program runs on South African mobile traffic, some of the partners driving your conversions are being under-credited every month — and it’s not your team’s fault. It’s the network.

The shared-IP problem

Most SA mobile subscribers don’t get their own public IP address. Networks like MTN and Vodacom put tens of thousands of subscribers behind a single public IP using carrier-grade NAT (CGNAT). From the outside, 50,000+ different people can look like one visitor — and the same person’s IP can change mid-session as they move between towers or the carrier rotates the pool.

Standard attribution leans heavily on the IP address (and cookies) to connect a click to a later conversion. On CGNAT mobile traffic, both signals fail: the IP isn’t unique and doesn’t persist, and mobile in-app browsers routinely drop third-party cookies. Here’s a plain-English explainer of CGNAT if you want the mechanics.

What it costs you

The result is attribution failure on exactly the traffic that matters most in South Africa — mobile, influencer-led, township-and-metro reach. Concretely:

  • Partners who genuinely drove a conversion don’t get credited, so you mispay them.
  • Your best mobile channels look worse than they are, so you under-invest in them.
  • CPA and ROAS are computed on a partial picture, distorting every optimisation decision.

None of this shows up as an error. It shows up as “mobile just doesn’t convert as well” — a quiet, expensive misread.

How you close the gap

The fix isn’t a better cookie. It’s to stop depending on the rotating IP and instead anchor the click to a cryptographic, device-bound token that survives the IP change between click and conversion. When the conversion arrives — even from a completely different IP — the token still resolves it to the partner who earned it.

That’s the layer AuraSync adds. Payouts stay FICA-gated and POPIA-compliant; no personal information is stored to make the match. (The signal-processing methods are proprietary — but the outcome is simple: the credit lands where it should.)

Don’t take our word for it — measure it

You don’t have to believe a claim. A short Attribution Audit places one tracking link on your mobile flow for two weeks and hands back a report of exactly how much your current attribution is missing, using your numbers. If the gap is real, you’ll see it in Rands.