The Maze: MMA South Africa has opened a proposed retail-media measurement framework for industry review. The draft tries to solve a familiar problem: two networks can report the same return on ad spend while using different sales definitions, attribution windows and levels of modelling. Its answer is less glamorous than a new ad format, but more useful. Make the basis travel with the number. The South African version also confronts the market's physical-store weight, cash transactions and dependence on loyalty IDs for linking advertising exposure to offline purchases.
The draft creates a reporting floor, not a winner's badge. Version 0.3 covers on-site, off-site, direct-marketing and in-store retail media. It uses two working tiers, Standard and Advanced, plus Not Applicable where a metric genuinely does not apply. Core delivery is the exception. Any network selling paid media should report impressions, clicks or click-through rate, spend and reach/frequency at Standard level. That does not certify performance. The tier names, challenge process and transition needs remain open, and the framework is neither law nor an adopted standard.
A return figure must bring its receipts. The draft's most practical rule is portable disclosure. Attribution windows, gross-versus-net sales, VAT treatment, discounts, extrapolation and loyalty-linkage rates should sit beside the result they qualify, not in a methodology appendix. A retailer may model outcomes beyond the customers it can identify, but it should show both the modelled share and the method. That matters where cash baskets can be invisible to identity matching. Without the base, weighting and sales definition, the return is impossible to compare.
Attributed sales are not automatically incremental sales. A sale can receive credit because it happened after an ad exposure without proving that the ad caused it. The South African draft renames simple before-and-after movement as directional sales lift and reserves incrementality for stronger designs with a credible counterfactual, bias control and enough signal to separate effect from noise. That distinction follows broader IAB guidance, which maps experiments, model-based counterfactuals, econometric methods and hybrid proxies to different business questions. For brands, this is the difference between asking "which network claimed the sale?" and "which spend created sales that would not otherwise exist?"
Physical stores make the standard harder—and more relevant. The in-store spine moves from traffic to opportunity to see, verified ad playout and, eventually, verified impressions. Sales measurement then depends on point-of-sale data, loyalty linkage and transparent treatment of unmatched cash purchases. Store tests need comparable control locations and aligned timing. Mature networks can make inventory easier to trust and compare. Less mature ones face more reporting work and clearer exposure of their gaps. Brands gain a procurement checklist before moving budget.
Why it matters: South African retailers and brands now have a concrete draft to challenge rather than another promise of "closed-loop" measurement. The consultation was announced for 60 days from September 14, although the official microsite still says the comment channel and closing date are pending confirmation. That inconsistency is exactly why transparent definitions matter. The real test is not whether every network reports the same return. It is whether buyers can see what was observed, what was modelled, which sales were counted and when a claimed lift is actually causal.
