Much of the past week’s rage at Facebook appears to hinge on the idea that wrongly obtained user data was used to influence the 2016 U.S. election – specifically, in favor of President Donald Trump. After all, dire warnings about mass data gathering have circulated for at least a decade. The public seemed to take little notice until the technology appeared to send politics haywire by using what a whistleblower described as an “arsenal of weapons” to influence voters.
But a series of reports suggest that, however real Facebook’s abuse of its users’ trust, Cambridge Analytica itself was hardly the master manipulator that it claimed to be. In a New York Times report from more than a year ago, Cambridge Analytica executives admitted that the “psychographics” techniques it promoted, supposedly able to profile voters’ deepest emotions, were not used in the Trump campaign. More recently, a political tech executive told the trade publication AdExchanger that the technology CA actually used was fairly standard, and that “Facebook or about any commercial [data management platform] can do that better even if their employees want you to lose.”
So, while psychographics has proven effective in influencing buying behavior in experiments, it’s still unclear whether or how the approach would work in an election. One political micro-targetting expert interviewed by The Verge questioned whether knowing a voter’s mindset based on Facebook “likes” could really have a dramatic influence amid the “overwhelming wave of data going into people’s head” during a political campaign.
Get Data Sheet, Fortune’s technology newsletter.
And when Cambridge Analytica made its most ambitious claims, it sometimes fell on its face. According to a new Mother Jones report, the company told the Ted Cruz campaign that a powerful software tool named Ripon could help it target voters, but that tool didn’t exist. It also told the Ben Carson campaign that it was adept in TV advertising, but turned out to be inept. Carson staffers reportedly thought it was possible the company was a sham. Even Trump wasn’t a lock for the firm, which onetime Trump campaign chair (and now indicted money launderer) Paul Manafort once described, according to Mother Jones, as “just full of shit, right?”
The disconnect between Cambridge Analytica’s image and its abilities has been repeatedly pinned on now-suspended CEO Alexander Nix. One former colleague described Nix to Mother Jones as an opportunist, whose sales pitch often amounted to “’Can I sell this to you and work out the details afterward?’” In 2016, the opportunity Nix saw was in Republican politics, where a dearth of political-tech players left an opening. His tendency to oversell may have also proven his undoing when a documentary crew taped him suggesting his company could deploy entrapment tactics to smear clients’ opponents.
It seems fitting, then, that Facebook has been the focus of anger after the discovery of its flawed data policies. But the fact that Cambridge Analytica was little more than a digital marketing firm with a posh British accent shouldn’t defuse anxiety about the impacts of digital profiling. YouTube has been shown, for instance, to algorithmically push viewers towards extreme content, and online propaganda has been tied to a rise in tribalism. Whether or not that dynamic can be steered to the benefit of a particular candidate, the risk to democracy itself is obvious.