Some time ago, I heard some gossip that a former colleague of mine had been accused of a serious crime. This person had been rather difficult to work with, to put it mildly, so it was very tempting to believe the worst about them; but the spurious nature of the allegations and their source – as well as a purported detail that cast doubt on their veracity (a court hearing supposedly scheduled at a court that would have been closed for repairs) meant that I eventually dismissed the claims against my former colleague as probably false, and thought little more of it.
In the aftermath, I saw that my former colleague had started their own personal website and begun to publish on various sites what looked at first glance like LinkedIn-style insight articles: professional commentary, promotional articles in business blogs, travel and culture reviews, and the like. I was intrigued – had my colleague, whom I could easily imagine dismissing such material as pointless nonsense – suddenly decided to embrace the ‘personal branding’ trend?
However, there was something very off about the articles. They all had the strong whiff of AI writing: most were in a tell-tale style of prose that sounds lofty without providing much real insight (the kind of writing style you’ll recognise if you’ve ever written or graded student essays), and in several cases, the articles even stated directly that their contents were works of fiction – exactly what happens if someone publishes the output of a chatbot without proofreading it first. (I got a particularly good laugh out of a floridly expressed, clearly fictitious account of my colleague’s culinary adventures in an exotic European city, steeped in overwrought nostalgia.)
What on earth was going on?
A little digging through the network of websites provided the answer: the personal branding site and associated ‘publications’ were the work of a content farm, likely as part of an agency in the reputation management industry. By searching Google and DuckDuckGo for links to content on the sites I’d found, I discovered that whoever provided this service had used the exact same playbook for multiple customers:
- A personal website and domain
- Accounts on other social media platforms and content sharing sites, most seeded with a small amount of published material, probably AI-assisted.
- Promotional articles (puff pieces) establishing their users as authorities in their industry (and walking bastions of culture, to boot).
That was when I realised: the transparently flimsy execution of this reputation management network had provided me with a rare and fascinating insight into who uses such services and invited a few general conclusions to be drawn – about the customers as a group, but also about whether the service actually achieved its intended effect, by making negative content about its users less prominent in search results.
Strong caveats apply, of course: this was a small sample of 21 identifiable individuals, probably unrepresentative of the target market for such services, and all conclusions are based on web search results which (as we all know) are not necessarily accurate sources of information; criminal claims about individual behaviour were often reported as allegations, but did not necessarily proceed to arrests, convictions or other legal sanction.
Who uses reputation management?
Of the 21 unique individuals represented in the content network that I examined this week, men were over-represented (90%) and the vast majority of users (86%) were based in the US. Most tended to be middle-aged professionals or business owners – I would loosely describe them as mid-career individuals with both a lot to lose and the means to engage a service to mitigate that risk.
Based primarily on media reports, 67% of users had allegations of a criminal nature against them (charges or arrests reported in online media), although I could only find confirmation of sentencing or convictions for 43% of those with allegations against them. The alleged crimes spanned a range of severity, from traffic incidents (DUIs) at one end of the scale to financial crimes (fraud/embezzlement) and sexual offences at the other. The dates on which allegations were reported spanned the last 20 years, with most (33%) having been reported in 2024.
For those users without allegations of unlawful behaviour against them, I could only speculate on why they had engaged the reputation management service. In those cases, I suspect the decision was motivated by personal associations with other individuals or organisations accused of criminal behaviour, or because of unflattering online reports of conduct that didn’t rise to the level of criminality (for example, an executive who was portrayed as greedy in the media over a dispute with their employer, or a professional who lost their license to practise).
Did the service improve users’ online reputation?
Since I couldn’t assess the state of users’ online presence prior to their use of reputation management, I decided to measure the effectiveness of the content network by looking at whether a user’s front-page Google and DuckDuckGo search results contained links to reputable media reports or official sites containing negative content about them. I used the names of users and their location as search terms, and did not look through any search results that Google had hidden at the request of a user (e.g. DMCA takedowns).
Overall, among all users, 71% had unflattering search results on the front page of Google and 52% had such results on the front page of DuckDuckGo. However, among those users with allegations against them, the equivalent rates were 93% for Google and 64% for DuckDuckGo – indicating that the content network was largely ineffective at reducing the prominence of negative reporting about allegations against individuals, even allowing for Google’s efforts to let people remove search results about them.
Conclusions
This kind of online reputation management should be considered a last resort. My limited findings here suggest that this strategy only has a very small effect on reducing the prominence of negative content associated with a person. But, if identified as the work of of a professional service, the high proportion of users with criminal allegations in this sample suggests that a service user could become ‘tarred by association’ through their engagement of a reputation management agency. Perhaps the main caveat of this particular analysis is that a more competent agency or approach to reputation management might achieve better results for its customers.
There are wider questions about the way individuals with problematic histories are and should be treated, and of course, in societies that adhere to the rule of law, people are presumed innocent until proven guilty. The recent dismissal by the BBC of Scott Mills, a radio presenter with historical allegations that were investigated and resulted in no charges being brought, is a case in point. Individuals should also have the ‘right to be forgotten’ – balanced against a public interest in the reporting of serious crimes.
But on the basis of this analysis, I would strongly advise against the use of such services: they likely won’t achieve their intended aims, and if poorly executed using a template approach, could end up inappropriately associating their users with others that have highly questionable backgrounds.