Unilever's U-Studio — The In-Housing Wave and Retreat

In 2017, Unilever announced the most aggressive in-housing experiment in the history of consumer goods marketing. Six years later, the company's own executives conceded that what had been framed as a revolution was actually something more useful — a careful redrawing of the boundary between in-house capability and agency partnership. The story of U-Studio is the story of an industry that discovered, through expensive trial and error, that the question is never in-house or agency. It is which work belongs where.

The Situation

Unilever is a useful starting point for any serious discussion of marketing organisation because the company operates at a scale that exposes every structural weakness. By 2017, Unilever owned more than 400 brands sold in 190 countries, with roughly 2.5 billion consumers using a Unilever product every day. The marketing function was spending approximately EUR 7 billion per year on brand and marketing investment, with content production alone running into hundreds of millions. The sheer volume of creative output required — product shots, social posts, recipe videos, localised adaptations, promotional banners, in-store point-of-sale, regional variants of global campaigns — had become unmanageable under a traditional agency-of-record model.

Keith Weed, then Unilever's Chief Marketing and Communications Officer, had been publicly agitating for structural change for several years. At the 2016 Cannes Lions festival, Weed described the content demand curve as "broken" — the number of touchpoints had multiplied faster than the agency model could absorb, and the unit economics of traditional production had become absurd for high-volume, low-stakes content. A single Dove social post might cost the same to produce, under traditional agency rate cards, as a piece of work that would run for six months on television. Procurement pressure had squeezed agency margins but had not fixed the underlying mismatch between what the agencies were structured to do and what Unilever actually needed them to do most of the time.

The industry context was equally tense. The Association of National Advertisers had released its 2016 report on agency transparency, which exposed practices around media rebates and undisclosed principal-based trading that had already rattled the advertiser community. Procter & Gamble's Marc Pritchard had delivered his January 2017 speech at the Interactive Advertising Bureau conference calling for an end to the "crappy media supply chain" — a speech that effectively gave every other major advertiser permission to interrogate its own agency relationships. Against this backdrop, Unilever's decision to build its own content production capability looked less like a radical departure and more like an inevitable response to a broken model.

U-Studio was launched in 2017 as an in-house content production unit, initially embedded inside WPP's offices in London, Mumbai, and Sao Paulo. The embedded-studio structure was deliberate. Unilever wanted access to WPP's talent pool, its tools, and its production infrastructure without having to build everything from scratch. WPP, for its part, wanted to preserve the Unilever relationship even as the client began to insource work that had historically been agency revenue. The "agency village" model, as it was called, was a hybrid from the start — neither pure in-housing nor traditional agency, but a new configuration that tried to capture the benefits of both. WPP's leadership under Sir Martin Sorrell had been publicly resistant to client in-housing, warning investors through 2016 and 2017 that the trend threatened holding-group revenue. Sorrell's successor, Mark Read, took a more pragmatic position when he assumed the chief executive role in September 2018, arguing that WPP had to embrace the hybrid model or risk losing the relationship entirely. The embedded U-Studio arrangement was a direct outcome of that pragmatic pivot.

The stated ambition was striking. Keith Weed told Marketing Week in 2018 that Unilever planned to shift more than fifty per cent of its content creation in-house over the following three years. The rationale was articulated in three parts. First, cost: Unilever estimated it could reduce production costs by thirty per cent on high-volume content by cutting out agency margin and overhead. Second, speed: simple content that took three weeks through an agency could be turned around in three days in an embedded studio. Third, brand knowledge: an in-house team living with the brand every day would build institutional memory that rotating agency teams could not match. Each of these arguments was plausible. Together, they sounded irresistible.

The talent strategy for U-Studio is worth noting because it would later become a source of tension. The initial hires were drawn from two pools: experienced creatives from WPP agencies who were willing to move into an embedded client-side role, and younger producers and content designers hired directly from the market. The embedded WPP model was, in one sense, a way for Unilever to transfer agency talent to the client side without paying the full cost of recruiting from scratch. But it also meant that the in-house team inherited some of the assumptions and working practices of a traditional agency, even as it was being asked to operate on fundamentally different cost and speed parameters. The cultural mismatch between agency-trained creatives and the operational demands of high-volume in-house work would become one of U-Studio's quieter challenges.

What followed was a rapid expansion that, for a brief period, made U-Studio the most-watched experiment in global marketing organisation.

The Decision

By 2019, U-Studio had expanded to roughly twenty studios across more than thirty countries. The unit was producing thousands of content assets per month for Unilever brands including Dove, Axe, Hellmann's, Magnum, Knorr, and Persil. Internal Unilever statements at the time claimed that U-Studio had already delivered on the thirty per cent cost-reduction promise for the content it handled, and that turnaround times on routine work had fallen by more than fifty per cent. Industry publications treated the expansion as validation of the in-housing thesis. The World Federation of Advertisers' 2020 report on in-housing cited U-Studio as the most comprehensive build-out in the consumer goods sector.

But the numbers obscured a more complicated reality. U-Studio had never been asked to handle Unilever's most difficult creative work — the origination of brand platforms, the development of new campaign ideas, the strategic positioning decisions that shape a brand's direction for three to five years. From the start, those decisions had remained with Unilever's network of traditional agencies. Ogilvy continued to lead global creative on Dove. What was then AMV BBDO worked on Lynx and parts of the Home Care portfolio. MullenLowe handled significant work on Persil and Unilever Home Care brands in several markets. Wunderman Thompson held the global Dove relationship for parts of the digital and CRM mandate. The agencies had been pushed to defend their territory, but the origination mandate was never seriously transferred to U-Studio.

What U-Studio actually did, and did well, was the long tail. A new Dove Men+Care campaign conceived by Ogilvy would be followed by dozens of local market adaptations, hundreds of social cut-downs, thousands of product-shot variants, and an ongoing stream of always-on content to feed the brand's digital channels between major campaign moments. This was the work that U-Studio took over — and it is work that the agency model had never handled efficiently in the first place. The traditional agency billing structure, built around senior creative time and project-based fees, was fundamentally mismatched with the repetitive, high-volume, template-driven nature of always-on content. U-Studio's cost advantages were genuine, but they came from doing the right work in the right place, not from replacing agency creative origination.

The quiet rescoping began around 2020. Internally, Unilever stopped describing U-Studio as a replacement for agency creative and began describing it as a "capability layer" for execution. Externally, the fifty per cent in-housing ambition disappeared from executive speeches. By 2022, Marketing Week reported that U-Studio was still operating at significant scale but had been explicitly repositioned as a production and adaptation capability, not an origination engine. The brand campaigns that launched in 2021 and 2022 — Dove's evolution of its "Real Beauty" platform, the Hellmann's Super Bowl spots, the global Axe repositioning — had come through agency creative teams, not U-Studio.

Fernando Machado, who had led brand building at Unilever and then at Burger King before moving to Activision-Blizzard, was among the more candid voices about what had happened. In multiple public interviews and podcast appearances after leaving Unilever, Machado described the in-housing wave as "half right" — right that most content needs had outgrown the traditional agency model, wrong about what in-house teams could realistically replace. Machado's core argument was that origination requires a particular kind of creative culture that is difficult to sustain inside a client organisation. The internal pressures that make in-house teams efficient at execution — tight deadlines, embedded brand knowledge, cross-functional accountability — are often the same pressures that suppress the kind of risk-taking that brand-defining creative work requires.

Keith Weed, after his 2019 retirement from Unilever, became more reflective in his public comments. Speaking at events in 2021 and 2022, Weed described U-Studio as "a great production asset and an incomplete marketing revolution." His argument, in essence, was that the cost and speed benefits had been delivered, but the deeper ambition — to transform how Unilever developed brand ideas — had not. The company had learned, at considerable expense, that the capability map mattered more than the in-housing ratio.

The contrast with Procter & Gamble is instructive. P&G had pursued its own in-housing experiments, including a major build-out under Marc Pritchard's direction, but the P&G approach was from the start more cautious about origination. Pritchard's emphasis was on agency consolidation and fee-model reform rather than on replacing creative origination with in-house capability. P&G kept Saatchi and Saatchi on Pampers and Head and Shoulders, Grey on Gillette, and Wieden and Kennedy on Old Spice, treating these relationships as strategic rather than transactional. When the Unilever U-Studio experiment began to rescope in 2020, P&G's more conservative approach looked retrospectively wiser — not because in-housing was wrong, but because P&G had never over-claimed what it could do.

There was also a subtler organisational problem that only became visible in retrospect. U-Studio's embedded structure inside WPP buildings had been designed to capture the benefits of agency proximity, but it also blurred the reporting lines in ways that made accountability difficult. A creative working on a Dove social post might be formally employed by Unilever, sitting in a WPP office, managed day-to-day by a U-Studio lead, but influenced by the surrounding Ogilvy team culture. When the work was successful, attribution was easy to share. When it was weak, attribution was easy to deflect. The hybrid model that had been designed to capture the best of both sides sometimes captured the worst of neither — ambiguity about whose responsibility the output really was. Senior Unilever marketing leaders reportedly raised this concern in internal reviews during 2019 and 2020, and it became one of the drivers of the eventual rescoping. A cleaner separation between in-house production capability and external creative origination, the thinking went, would make accountability sharper on both sides.

Unilever's leadership transition is also part of the story. Alan Jope became CEO in January 2019, replacing Paul Polman, and inherited the U-Studio programme partway through its expansion. Jope's strategic focus was shaped by different priorities — the divestment of the tea business, the repositioning around purpose-driven brands, the ongoing debate about Unilever's headquarters structure — and he did not have Weed's personal investment in the in-housing thesis. When Conny Braams was appointed Chief Digital and Marketing Officer in 2020, succeeding Weed's role, the new leadership was more willing to rescope U-Studio without framing the change as a retreat. The organisational politics of the rescope were easier because the ambition had belonged to a predecessor generation. Subsequent leadership transitions, including Hein Schumacher replacing Jope as CEO in 2023 and Esi Eggleston Bracey joining as Chief Growth and Marketing Officer in 2024, have continued the more pragmatic capability-map approach rather than attempting to revive the original in-housing thesis.

The Data

The public numbers on U-Studio are imperfect — Unilever has never published a full operational breakdown — but enough has emerged through annual reports, executive speeches, and industry coverage to construct a reasonable picture.

Metric	2017 (launch)	2019 (peak expansion)	2022 (rescope)
Number of U-Studio locations	3	~20	~25
Countries served	5	30+	30+
Stated in-house share of content	~10%	~40%	~40% (but narrower scope)
Production cost reduction (claimed)	n/a	~30% on in-scope work	~30% on in-scope work
Agency-of-record relationships	~30 major	~25 major	~20 major

The agency-of-record count is approximate because Unilever's agency roster is distributed across brand, region, and capability, and the company has never published a comprehensive list. But the directional story is clear. U-Studio's expansion did not eliminate Unilever's agency network — it shifted the work profile. Agencies lost volume but retained origination. The in-house team grew but stopped short of replacing strategic creative work.

Unilever's marketing and brand investment, as reported in its annual reports, is another useful data point. The company's spend on brand and marketing investment held broadly steady as a percentage of turnover across the 2017-2022 period — roughly 14 to 15 per cent of revenue, which at Unilever's scale translates to approximately EUR 7 to 8 billion per year. The in-housing programme did not reduce marketing investment overall. What it did was change how that investment was deployed, with more of it flowing through internal cost centres and less through agency fee structures for high-volume content, while origination fees remained largely stable.

The industry-wide context is also worth noting. The World Federation of Advertisers' 2020 in-housing report surveyed 56 major multinational advertisers and found that 74 per cent had some form of in-house creative capability, up from 42 per cent in 2016. But the same report found that fewer than twenty per cent had moved origination in-house, and that most of those had reversed or narrowed that decision by 2020. Unilever's U-Studio was the largest and most-watched instance of a much broader pattern: in-housing worked for execution, struggled for origination, and produced the best results when organisations were honest about the distinction from the start.

One data point that deserves specific attention is the content volume itself. Unilever has not published exact figures for U-Studio's output, but industry estimates compiled by The Drum and Marketing Week in 2021 put the monthly asset production at somewhere between 15,000 and 25,000 individual pieces of content across the active studios. This is a volume that would have been effectively impossible under a traditional agency-of-record model at any defensible cost. The agency industry's billing structure assumes that each piece of work requires a certain amount of creative director attention, strategic input, and client-side review — an assumption that breaks down at the scale of a modern always-on content programme. U-Studio's value, in this sense, was not only that it was cheaper per asset than agency production. It was that it made a volume of content possible that the agency model could never have supplied on any economic basis. This is the part of the in-housing thesis that survived the rescoping and that continues to define U-Studio's role today.

The Marketing Organisation Lesson

The U-Studio case connects directly to F10-02's central argument that in-house versus agency is a false dichotomy. The question is not whether to in-house, but which work to in-house. U-Studio's genuine success lay in taking over the volume and adaptation work that agencies had never handled efficiently, while its failure to replace creative origination demonstrated that the boundary between in-house capability and agency expertise is not a simple cost curve — it reflects different kinds of creative work with different organisational requirements.

F10-07's argument that the marketing team is a system is equally relevant. Unilever did not fail in its in-housing experiment because U-Studio was badly built. It succeeded in its original operational goals — cost, speed, and volume — because the system was designed well for those outcomes. But the ambition to extend that system into origination was a category error. Origination is a different kind of work, with different feedback loops, different talent requirements, and different time horizons. A system optimised for fast, high-volume execution will tend to suppress the slower, riskier, more iterative processes that brand-defining creative work requires. Unilever's eventual rescope was not a retreat from in-housing. It was a recognition that two different systems were needed, and that trying to force both through the same organisational structure would degrade both.

This systems framing is useful because it reframes what the in-housing question is actually asking. When a marketing leader asks "should we in-house?", the question is usually interpreted as a single binary decision about a single marketing function. But the underlying reality is that marketing comprises many different kinds of work, each with its own optimal organisational form. Always-on social content is one kind of work. Brand campaign origination is a different kind of work. Strategic positioning research is a different kind of work again. Media planning, performance media buying, public relations, influencer activation, shopper marketing, and loyalty programme management are all different kinds of work with different structural requirements. A marketing organisation is not one system; it is a portfolio of systems, and the in-house versus agency question should be asked separately for each of them. Unilever's U-Studio was in effect a late and expensive discovery of this principle.

The implication for any marketing leader contemplating in-housing is not that it should be avoided, nor that it should be pursued as a matter of principle. It is that the capability map must be drawn explicitly, in detail, before any structural decision is made. Which kinds of work in our marketing portfolio require deep brand knowledge, fast turnaround, and high volume? Those are candidates for in-house capability. Which kinds of work require creative origination, outside perspective, and the ability to challenge internal orthodoxy? Those are candidates for agency relationships — and, as F10-06 will argue, specifically for agencies operating as interventionists rather than suppliers.

The F10-08 capstone argument is that a marketing organisation explicitly designs the boundary between in-house and agency rather than treating it as a cost-driven default. Unilever's U-Studio eventually arrived at this synthesis, but it took six years and considerable disruption to do so. The cost of learning by doing, at Unilever's scale, was significant. For smaller organisations, the lesson is that the capability map can be drawn in advance — and that the most important strategic decision in marketing organisation is not whether to have an in-house team, but what that team is actually for.

The synthesis

U-Studio is the clearest example in recent marketing history of an organisation that moved from an either-or position — agency, then in-house — to a both-and synthesis that worked. The early framing of the programme was binary. Weed's 2018 target of fifty per cent in-housing implied that the remaining fifty per cent was residual, a transitional state on the way to a more complete in-sourcing. That framing was wrong, and Unilever eventually abandoned it. The synthesis that emerged by 2022 was explicitly both-and: U-Studio for volume and adaptation, agencies for origination and creative challenge, each operating in the work they were structurally best suited to handle.

The The synthesis is not that Unilever failed in its ambition. It is that the ambition was mis-framed. The real question was never "how much can we in-house?" It was "what does each type of work actually require, and which structure delivers it best?" When Unilever allowed itself to ask the second question, the organisation it arrived at looked more stable, more defensible, and more effective than either pure option would have been. The agencies retained their creative origination mandate because that is what agencies, at their best, exist to do. U-Studio retained its production mandate because that is what an embedded in-house team can do more cheaply and faster than any agency. Neither structure was a compromise on the other. Both were operating at full capacity in the work they were designed for.

This is the shape of a marketing organisation: not a fifty-fifty split, not a compromise, but a deliberate design in which each component is doing the work it is best suited to do, with the boundary drawn explicitly and defended against drift. Unilever arrived at this synthesis expensively and slowly. The lesson is not that the journey was wrong — it is that the journey could have been shorter if the binary framing had been rejected from the start.

Sources

• Unilever Annual Report and Accounts, 2017-2023
• Marketing Week, "Unilever unveils plans for global in-house agency," 2018 and subsequent coverage through 2022
• Campaign, coverage of U-Studio expansion and rescoping, 2017-2022
• Keith Weed, public interviews and speeches at Cannes Lions and Marketing Week Masters festivals, 2016-2022
• World Federation of Advertisers, In-Housing Global Report, 2020
• Fernando Machado, interviews with Marketing Week, The Drum, and industry podcasts, 2020-2023
• Marc Ritson, Marketing Week columns on in-housing, 2017-2022
• The Drum, coverage of Unilever creative agency relationships, 2018-2022
• Association of National Advertisers, agency transparency study, 2016