On the surface, MediaPost selected Kinesso as our media agency of the year based on a reboot, because following years of failed attempts and organizational false starts to find ways of monetizing parent Interpublic’s $2 billion acquisition of Big Data giant Acxiom in 2018 as a media service, IPG Mediabrands consolidated those data and technology assets under a different kind of Kinesso brand -- one that was inherently media neutral and focused 100% on developing the best internal data and technology to power the services of its branded media service agencies: Initiative, UM and Mediahub.
That’s the surface story. Beneath it is a much more nuanced story about the role engineering should play in advancing the art and science of data and technology for agencies and their clients, and to do it in the most media-neutral way possible.
To be clear, Interpublic continues to look at how to use Mediabrands to develop ways of monetizing its costly Acxiom investment. Late last year, it announced a proprietary deal between its Magna unit and OpenAP to integrate its "OpenID" consumer identifier data with Acxiom’s consumer-profile data.
Ironically, that is how Kinesso began life in 2019, when a year after acquiring Acxiom, Interpublic launched it as “a marketing intelligence engine powered by Acxiom” and began integrating it with the agency trading desk turned “addressable media activation unit” Cadreon to effectively figure out new ways to upsell Acxiom data to clients.
A year later, Interpublic rebranded Cadreon as Matterkind with the explicit mandate of leveraging Acxiom data.
The exact mission and branding of both Kinesso and Matterkind never quite resonated with the outside world, and it is questionable how much it did inside Interpublic itself. So when the holding company announced Kinesso’s reboot in September 2023 -- combining the best data and technology capabilities of the old Kinesso, Matterkind and Reprise -- it described it in the most media-neutral way possible: an "ingredient brand."
“To use an automotive analogy, we might be the common components that are invisible to the client in many instances, like the engine or the chassis. But the details that people see and touch and interact with will be the agency brands." That's how Kinesso’s new CEO Jarrod Martin described its mission.
This is not to say that Kinesso won’t be leveraging Acxiom data, if and when it’s appropriate for the agencies they serve, or their clients. It’s just that it’s not its main goal.
And that is one of the beneath-the-surface reasons why MediaPost selected it as our media agency of the year, because in doing so, Kinesso is setting a path to fix one of two big problems identified as the major transparency problems in the advertising industry, according to the Association of National Advertisers’ 2023 transparency reports: "misaligned incentives."
By rebooting Kinesso as a media-neutral technology unit, it no longer has an incentive to upsell or remarket Acxiom -- per se -- unless its data is the best solution for the problem Kinesso is solving for.
In fact, Kinesso’s business model is based entirely on IPG Mediabrands’ P&L -- not on any direct fees or commissions -- so it is inherently media-neutral and its incentives are aligned 100% around making Initiative, UM and Mediahub perform better for their clients.
That’s the first beneath-the-surface reason we selected Kinesso this year. The other reason has to do with the second major issue identified by the ANA’s transparency study: "information asymmetry."
That’s a fancy term used by economists to describe what happens when traders compete in a marketplace where one or more sides have more information about the intrinsic value of what they are trading than other sides.
Unlike Wall Street, where there is an abundance of regulated open-market data on the underlying value of commodities and equities, Madison Avenue has a long history of operating with asymmetrical information.
Long before digital media -- and especially programmatic marketplaces and so-called walled gardens -- exacerbated the problem, media suppliers often held most of the cards.
Whether they were rate cards and proprietary subscriber databases controlled by big print publishers, or TV networks requiring agencies “register” their upfront advertising budgets in order to trade in the marketplace, agencies historically were at an information disadvantage.
But as the ANA’s reports revealed last year, programmatic trading was akin to putting information asymmetry on steroids.
The second reason we picked Kinesso as our media agency of the year is that it has figured out a way to make the media-buying marketplace more symmetrical, shifting from programmatic trading to "algorithmic trading."
To be clear, much of what Kinesso has done so far is more of a skunkworks approach to developing these solutions, but at least it has the roadmap.
And during a series of proprietary meetings with the editors of MediaPost, it showed us some of them. Not surprisingly, we cannot show you the details of what Kinesso’s team presented to us, but we can give you a good example of how one of them -- algorithmic trading -- works, and why it makes the digital media-buying marketplace more symmetrical.
“It’s analogous to what we saw with stock trading 15 years ago,” explains Dan Chapman, global chief strategy officer of Kinesso, referring to high-speed trading systems that gave a competitive advantage to Wall Street traders who could afford to invest in the technology in order to leverage data insights in seconds -- even nanoseconds -- faster than the next trader.
Chapman says the roots of Madison Avenue’s algorithm trading began in the mid-1990s with the advent of biddable media like paid search, and have been evolving ever since.
The problem is that most of the science and engineering has been on the supply side inside walled gardens like Google, Meta and Amazon, while the demand side had to rely either on their benevolence for optimizing media buys, or on best guesses for what was driving them.
“Google ranks your paid-search bid based on 23,000 variables,” Chapman continues, citing elements from the price a search advertiser is willing to pay to the speed with which their website loads to the probability that their bid will actually yield a sale or conversion.
As good as media agencies have become at optimizing media trading, Chapman says the information disparity has clearly been in favor of the supply side -- which is why Kinesso has turned to algorithmic trading, which utilizes a scientific approach of rapid testing and experimentation to deduce the underlying logic of the algorithms that big digital advertising platforms use to price and allocate audience inventory.
Chapman says the approach is the most logical way to counter the information asymmetry between the supply and demand sides -- especially as the platforms incorporate more advanced forms of machine learning and AI to optimize their yield.
To illustrate how Kinesso is analyzing the algorithmic logic of big digital suppliers, he cites some recent experiments the agency conducted to understand Google’s Performance Max algorithm, which optimizes and allocates all formats of Google’s ad inventory -- search, product listings, display ads, video, etc. -- to an advertiser’s campaign based on what is most likely to yield a conversion.
Google doesn’t make the algorithm powering it public, but based on Kinesso’s experimentation and analytics, the team has at least deduced its logic.
“We think it uses a genetic algorithm based on Darwin’s theory of evolution,” Chapman explains, adding: "We think Performance Max operates that way."
Chapman likens the shift to algorithmic trading to an "arms race," with buyers and sellers utilizing high-speed experimentation to deduce what the other side is doing.
“We think this is where the future of media agencies will be existing,” he says.
Time will tell whether Kinesso’s approach becomes an industry standard. But at least it has a roadmap for getting there and is working fast to develop it, which is why we selected Kinesso as our media agency of the year.