The conversation that led to Private By Design happening was not, in hindsight, a particularly unusual one. Kevin was on the phone with a publisher he had known for years - a well-run site covering a specialist subject, loyal readership, decent traffic, the kind of property that any advertiser with a relevant product should have wanted to be on. The publisher had just been rejected, for the second time, by one of the larger ad networks.
The reason given was vague. Something about content category. Something about brand safety thresholds. The actual audience - real people with documented interest in the subject matter - was apparently not part of the calculation.
That conversation happens constantly. It is not exceptional. It is the default experience of a large proportion of independent publishers operating in specialist niches.
The networks were built for scale, not fit
The major programmatic advertising networks were designed to operate at enormous scale. That architecture has advantages, but it also has a built-in bias: everything that cannot be processed at scale gets filtered out or deprioritised. Niche publishers, specialist content, audiences that require contextual understanding rather than demographic proxies - these things do not fit neatly into the categories the networks use, so they get rejected or assigned floor prices that bear no relationship to the actual value of the audience.
The result is a system where a publisher with 400,000 highly engaged monthly readers in a defined specialist vertical earns less per impression than a content farm with five million low-engagement page views. The CPM model, as implemented by the major networks, has never been particularly good at pricing audience quality. It is reasonably good at pricing audience quantity.
The advertiser side of the same problem
The advertisers experiencing this problem rarely know it is happening. They are told they can reach their target audiences through programmatic channels. What they are actually buying, in many cases, is a probabilistic approximation of that audience, assembled from behavioural data of varying quality, spread across inventory that may or may not be relevant to their brand.
The specialist publisher who demonstrably has the audience they want - because the site is literally about the subject the advertiser is selling into - is invisible to them, because that publisher is not in the network.
This is not a trivial inefficiency. For advertisers in specialist categories - B2B technology, outdoor equipment, legal services, financial products, health and lifestyle - the quality of audience alignment on a well-run niche publisher can be significantly higher than what the programmatic ecosystem delivers. The CPM may be higher in absolute terms. The cost per meaningful engagement is typically much lower.
What we actually do
We work in the space between publishers who cannot get into mainstream networks and advertisers who would benefit from reaching them. That means finding publishers with the right audience quality and compliance posture, and connecting them with advertisers who have been looking for exactly that audience without knowing where to find it.
It also means working with publishers whose content falls into categories the major networks will not touch - not because the content is problematic, but because their content policies are written for the median case and do not accommodate the full range of legitimate publishing. Satire, commentary, legal information, certain health topics, political analysis - categories where the content is entirely legitimate but the automated classification systems fail.
We are not a programmatic platform. We do not run an auction. We make considered decisions about which publishers belong in our network and which advertisers are a good fit for them, and we manage those relationships directly.
The underlying principle is straightforward: publishers who have built something genuinely valuable should be able to earn from it, and advertisers who have a relevant product should be able to reach the people most likely to want it. The infrastructure that was supposed to make that happen efficiently has not actually delivered it for a large portion of the market. We exist to fill that gap.