In-housing isn’t a new idea. It’s been discussed seriously before, particularly during the last major shift in digital media between 2017 and 2020. That period coincided with growing concern around transparency in the media buying ecosystem and prompted some brands to really consider whether they fully owned everything in digital media.
At the same time, the other driver was digital performance channels that grew in their stature inside retail organisations. Search, Shopping, and Paid Social had become larger budget lines, and retailers began to see that they delivered serious commercial outcomes, and that meant there needed to be greater accountability for media performance. Digital media was no longer something that could sit under the radar within the eCommerce or marketing department, as decisions made in these channels had direct impacts on revenue, margin, profit, stock, and growth. Therefore, in-housing emerged as a response to those shifts, but it did not happen en masse.
At that point, optimisations relied heavily on people, where capability was driven by years of experience inside platforms, where performance was influenced by how well individuals could respond to media volatility and extract marginal gains. In that context, in-house control and ownership came with real potential downsides, particularly when compared to agency teams that had accumulated their expertise through exposure to many retail accounts and associated edge cases.
Those dynamics meant that even where activity moved in-house, dependency often remained with external agency support continuing quietly, shadowing work, and things rarely moved entirely to the client side internally. What appeared to be a false start in the industry was, in my view, a reflection of where expertise sat.
Today, the mechanics of digital media operate differently. Deployment and optimisation of media does not go hand in hand with how many people you have in your team, and equally, it should not sit with just one or two experienced practitioners. The explicit knowledge is now accumulating continuously elsewhere. It's growing in the AI systems that learn, adapt, and operate in and across channels.
This change is now real, and it shapes how media is run day to day and how decisions are made. The media platforms are responsible for the majority of execution, the role of people should now shift to different types of work, moving away from constant intervention and more towards direction, constraint-setting and a closer interpretation of what outcomes mean for a retailer's business objectives. The most logical and appropriate place for this to happen is within the confines of the business, not outsourced to an agency.
As a result, the economics of media management change as well. Large teams deploying and managing media are no longer a requirement to achieve growth. The justification for outsourcing activity has diminished as people and expertise are no longer the constraints. In the last two years, I have seen a demonstrable shift in where the dependency now sits.…from explicit knowledge in media practitioners to AI-driven technology.
From this, I don’t see retailers attempting to replicate agencies in-house. What is changing is the balance between execution and ownership. With AI-led technology now carrying more of the operational load, questions around measurement, accountability, and decision-making naturally move closer to the retail side.
In-housing is not happening because it is similar to rhetoric of the past; it's happening because ownership is moving to where capability now sits, and with the right strategic leadership, more value can be extracted using the right approach.
This is why I think the current moment is different, as the conditions that previously made in-housing risky have gone with automation maturing and AI-led optimisation embedded in the core of performance media.