Digital ad fraud could be costing SA businesses billions

Digital ad fraud could be costing SA businesses billions

Brands are losing billions. This is budget that companies could use to drive meaningful revenue and improve our economy. My company is analysing a live digital ad campaign for a brand portfolio belonging to a major global consumer brand. To date, we’ve examined almost 200 million ad impressions and have confirmed that 37% are fraudulent. Extrapolated, this represents a loss of close to 37% of the budget used to fund the campaigns. For the campaign we’re looking at, only 12% can be confirmed as legitimate – i.e. that were viewable by a human. The remainder sit in a grey zone, neither proven fake nor proven real, albeit with valuable direction markers. By our measure, that brand cannot demonstrate a genuine audience for close to nine in 10 of the impressions it has bought. As we optimise the campaign, confirmed fraud among the most recent 20 000 impressions falls to 27%. That is progress. It also means more than a quarter of impressions are still confirmed fraudulent, even as we actively clean up the campaign. None of this is unusual. Across the campaigns we have analysed for South African businesses, wastage in programmatic advertising commonly runs between 20 and 60%. I cannot claim our sample represents the whole industry and reliable industry-wide figures for SA are hard to come by. But these results are typical of what we see across over 100 campaigns totalling a billion impressions and, so far, the local picture has been worse, not better. By my estimate, SA’s digital advertising market is worth $2.1bn to $2.4bn a year (R35bn to R40bn), with programmatic accounting for 40 to 50%. Programmatic advertising is the automated buying and selling of digital ad space using software, algorithms and data, rather than manual negotiations. If the rates we observe held across that market, losses would run from R2.8bn to R12bn a year. South Africa is not uniquely exposed. Fraudlogix, a fraud and bot detection company, analysed 105.7 billion impressions collected throughout 2025 and found a global invalid traffic rate of 20.64%: roughly one in five impressions showing characteristics of fraudulent or non-human activity. Industry trackers measuring net unrecovered financial losses across all digital channels, including search, social, and display, put total global ad fraud losses in 2026 between $100.2 billion and $133 billion representing ~ 16% -22% of total digital spend. Mark Warner, the US senator, cited some of these figures in a recent letter to Andrew Ferguson, chair of the Federal Trade Commission. In it, Warner recalled warning a decade ago that bots were creating fake consumer traffic, “artificially driving up advertising cost in the same way human fraudsters can manipulate the price of a stock by creating artificial trading volume”. The mechanics are simple to describe. Operators stack several ads behind a single visible one. They serve ads into zero-by-zero pixel spaces, or on fake websites nobody visits. They generate ad calls never designed to produce a visible ad. They load background ads in apps while you play or sleep, and falsify traffic and download figures for “junk” apps. Advertisers can then be charged for impressions with no opportunity for human engagement. Volume is cheap to manufacture, too: one freelance marketplace lists more than 2 200 website traffic services, with some sellers offering thousands of “targeted” or “organic” visitors for as little as $5. This is the systematic diversion of significant portions of advertisers’ budgets by criminal means. We follow the ad after it is served. A JavaScript tag inserted into the ad logs data for its lifetime: where it appears, its geographic location, the internet service provider and the device. Those signals form a fingerprint for each impression. We classify an impression as confirmed fraudulent only when at least three negative indicators are present. Hagai Shechter, CEO of Fraudlogix, says ad fraud is now an incentive problem rather than a technological one and platforms should stop treating detection as a cost centre. He is half-right. The incentives are plainly broken: when parties across the supply chain are paid by volume, few ask whether that volume is human. For advertisers, the response is not complicated. Demand impression-level evidence of where your ads ran and who, or what, saw them. Hold media spend to the same audit standard as every other line in the accounts. No finance director would sign off on a supplier who could prove delivery on only 12% of an order. Media budgets deserve no less scrutiny. The ad fraud problem is as pervasive in SA as globally. Brands are losing billions. This is budget that companies could use to drive meaningful revenue and improve our economy.

Original Source

Read the full article at Citizen →

KhanList aggregates and links to publicly available news content. We do not host full articles from third-party sources. Always verify important information with original sources.