Unfiltered Media

Mark Marshall, chairman of global advertising and partnerships at NBCUniversal, joins Justin Lebbon to unpack how premium video competes for spend against social and short-form platforms — and why data, not just content, is now the battleground.

• NBCUniversal is betting that data (Performance Insights Hub) plus premium content is the only way to compete with social and short-form platforms.
• The Hub combines linear and digital on one plan with reach, frequency and first/third-party KPIs; it rolls out in Q4 and, via FreeWheel, is intended to become an industry-wide standard.
• Universal Ads, aimed at digital-first advertisers moving up from search and social, now generates tens of millions per quarter after 18 months from a zero base.
• Diversification is essential: BravoCon (30,000 tickets, sold out in 90 seconds) and the LA28 sponsorship-plus-media partnership are meaningful new revenue streams.
• Marshall argues advertisers moved to CTV too quickly — ~40% of US budget has shifted while ~89% of impressions still run on linear.
• The US measurement 'crisis' — including Nielsen's Gauge report — confuses the market; Marshall's answer is real first-party data agreed across the industry.

In this pre-Cannes episode of Unfiltered, Justin Lebbon and Ian Whittaker unpack a dramatic week of M&A and IPO activity. They examine what SpaceX's elevated valuation means for funding the AI and space "twin bets", why the public markets impose a discipline private players escape, and how that pressure could see AI giants come hunting for media's advertising revenues. They then turn to Fox's roughly $22bn acquisition of Roku, the market's sceptical reaction, and the wider wave of US-led media consolidation.

Highlights:

• SpaceX's valuation as a "virtuous circle": a high price justifies and fuels large-scale investment — and works in reverse if the price falls.
• Musk's super-voting B shares mean investors effectively cede control and ride his coattails.
• The combined SpaceX/OpenAI/Anthropic valuation sits north of $3.5tn — and to justify it, these companies must take revenue from somewhere, with advertising the obvious first target.
• Why incumbent media firms are "trapped": shareholders punish the big investments needed to fight back (the ITV digital-rollout share-price drop as precedent).
• Is this "performative capitalism"? Ian pushes back — he sees Musk as epitomising "dream big", with the US having outsourced much of its space policy to SpaceX.
• Fox isn't buying shows — it's buying the home screen and the interface between viewers and ad revenue, plus a sports negotiating "choke point".
• The market's 15%+ share-price fall signals fears of overpayment, execution risk and debt; the deal is cash-and-stock (~$14bn cash of ~$22bn, ~$96/share).
• A wave of US consolidation — Paramount/WBD, Disney/Hulu, Fox/Roku — versus a slower-moving Europe (Sky/ITV, RTL/Sky Deutschland), with many sellers quietly seeking an exit.

• SpaceX's elevated valuation is itself the funding mechanism — a virtuous circle where a high price makes large-scale investment look positive, and a falling price reverses it.
• Musk's super-voting shares mean investors are betting on the man and ceding control, with no real voice in the decisions.
• To justify combined valuations above $3.5tn, AI giants must grab revenue from elsewhere — and media advertising, a proven trillion-dollar model, is the first logical target.
• Incumbent media companies are trapped: the investment needed to respond hits cash flows and spooks shareholders, as ITV's digital-rollout share-price fall showed.
• Fox's Roku deal buys the home screen and the viewer-to-ad interface, not content — but a 15%+ share-price drop signals the market thinks Fox overpaid and is pricing execution risk and debt.
• US media is consolidating fast (Paramount/WBD, Disney/Hulu, Fox/Roku) while Europe moves slower, with many sellers quietly looking for an exit.

After 25 years in media — including a stint as CEO of PhD UK — Ali Reed made the jump from the holding-company world to independent agency Brainlabs. In this candid conversation with Justin Lebbon and Ian Whittaker, she explains why legacy structures, short-term KPIs and "shuffling spreadsheets" pushed her out, and how indies are building AI-native capabilities without decades of technical debt to retrofit.

Highlights:

• Why she left the holdco: Personal sacrifice (cancelled holidays, weekly commute from Malta) plus a professional frustration that real influence over the work had been replaced by "managing different parts of the system."
• The indie advantage: No legacy infrastructure to retrofit, fewer layers of approval and politics, and the freedom to build AI capabilities cleanly. Brainlabs has partnered with Claude and Notion at enterprise level and published over 700 skills across its ~1,100-person global team in a few months.
• Cortex and Cortex Labs: A centralised shared tech platform plus a decentralised mode where account directors and practitioners — not just the dev team — build and ship bespoke client tools in days.
• AI as force multiplier, not headcount cut: Brainlabs frames AI as a route to "more for the same," not "more for less." Examples include "Ralph," a growth agent that helps smaller pitches over the line, and "Proteus," a creative variation tool that helped deliver 3x revenue on Meta for one client.
• Brand vs. performance: Reed argues brand building isn't dead, but agencies have different "superpowers" — and clients need braver operating models that let an innovation/test-and-learn partner work alongside a stable global agency of record.
• The platform threat: With most new media spend flowing to a handful of platforms (much of it direct), Reed says agencies add value through creative quality, first-party signals, early product access and accumulated test-and-learn muscle memory.
• Her pitch question: Instead of asking agencies why they're great, she'd ask what they'd regret if they didn't win — a test of honesty over hyperbole.

• Reed left a holdco CEO role partly because real influence over the work had given way to 'shuffling spreadsheets' and managing the system rather than improving client outcomes.
• Independent agencies can build AI-native capabilities cleanly, without retrofitting onto decades of technical debt across acquired businesses.
• Brainlabs frames AI as a force multiplier for more output, clients and revenue — not a tool for cutting headcount; it wants clients who want 'more for the same,' not 'more for less.'
• Decentralised tooling (Cortex Labs) lets practitioners closest to the client problem build and ship bespoke tools in days, on the principle that 'innovation happens at the edge.'
• As platforms automate the buying layer, performance is won or lost at the creative and first-party data layers — and on speed to market.
• Clients should embrace braver operating models, letting an agile test-and-learn partner incubate ideas that a larger agency then scales, rather than choosing either/or.

Unfiltered returns from a spell on the sidelines with guest Erez Levin — speaker, writer and co-author of a major new paper on media quality — alongside hosts Justin Lebbon and Ian Whittaker. The conversation opens on the state of the ad market and a strong earnings backdrop, then digs into the core question: in a world of unlimited impressions, how do you know what you're buying, and why does quality matter?

Highlights:

• A two-tier ad market. Ian on the macro: 85%+ of S&P 500 companies beating earnings estimates, blended q1 growth pushing nearly 30% — the strongest quarter in around five years. Traditional advertisers growing mid-single digits, with the SME market still driving the platforms.
• The consumer signal is mixed. Weak US consumer confidence, but PepsiCo cutting prices up to 15% and McDonald's flagging pushback — yet recession fears have repeatedly failed to materialise across 2024, '25 and '26.
• The "quality trifecta." Erez frames marketing effectiveness as media quality, creative quality and audience/data quality — with the paper focused on media quality specifically.
• Quality is a range, not a binary. Why Erez prefers "quality" to "premium," built around attention (prominence of placement) and context (user receptiveness — e.g. a McDonald's ad before vs. after dinner).
• The race to the bottom. Billions going to muted videos in the bottom corner labelled as "in stream" ads, bought at the same price as full-screen sound-on inventory because buyers don't check.
• Finance vs. marketing. How marketing trained finance teams to expect ever-lower CPMs — and why that metric is now used against marketing's own case. Cost is not what matters; value is.
• Clock science vs. cloud science. Erez on treating marketing as deterministic when it's probabilistic, the move from multi-touch attribution toward econometrics and experimentation, and the death of the cookie-era illusion.
• Beyond averages. YouTube as the case study — the top 5% as high quality as the best TV, but "the exception, not the rule." The fix: pricing on a few qualitative dimensions (time of day, channel/show, geo) rather than channel-level averages.
• Transparency vs. quality. They're different things — high quality media can be opaque — but buyers increasingly need to validate what they're paying for.
• Will the big three move? Justin presses on three platforms taking ~90% of new spend with little incentive to change. Erez: only their incentives will move them; build a parallel quality path for the enterprise buyers who care.

• All impressions are not created equal — the paper's first principle and the foundation for assessing media quality.
• Quality isn't binary or the same as 'premium'; it's a range driven by attention (placement prominence) and context (user receptiveness).
• Buyers routinely overpay, paying the same CPM for muted, bottom-corner 'in stream' video as for full-screen sound-on ads — and often don't check.
• Marketing trained finance teams to chase ever-lower CPMs, and that metric is now used against marketing; the fix is to argue value over cost and account for long-term brand building.
• Channel-level averages hide huge quality variance (e.g. YouTube's top 5% vs. the rest); price on a few qualitative dimensions like time of day, show type and geo instead.
• The big three platforms won't move first — only enough buyer dollars demanding quality will change their incentives, so build a parallel quality path for enterprise marketers.

In this episode of Unfiltered, Justin Lebbon and Ian Whittaker examine whether ad tech is facing its reckoning. From a finance lens, Ian explains why the market has turned skeptical on the sector — and why the biggest platforms, not the independent DSPs, increasingly hold the power.

• The Trade Desk's fall: Decent Q1 numbers, yet the stock is down ~70% over the period — a story of collapsing multiples, perceived risk, and doubts about future growth.
• Platforms own the pipes: Amazon and Google are taking DSP share while smaller players get squeezed; a handful of companies control up to 90% of new media spend.
• The Standard Oil parallel: Ian argues dominance comes not from sheer size but from controlling the choke points — just as Rockefeller controlled the route to market via railroad deals.
• Differentiation survives: Players with proprietary data and strong client relationships can defend value; commodity middlemen sitting in the gap are most at risk.
• Agencies move back in: As TV money shifts online, proprietary trading and rebates prop up agency margins — but the deeper problem is that media buying and planning is now priced as a commodity.
• AI as the efficiency tool: It has the potential to overturn ad tech economics, driving fewer steps, lower cost and pressure on fees.
• What premium publishers must do: Simplify processes, push money toward working media, and build or white-label their own controlled solutions to protect CPMs.

• The Trade Desk's ~70% share-price decline is driven less by weak numbers than by collapsing multiples, higher perceived risk, and lost faith in future growth.
• Amazon and Google are taking DSP share and squeezing smaller players, who lack the scale and balance sheets to compete.
• Ad tech's power dynamic mirrors Standard Oil: control of the choke points, not sheer size, is what matters.
• Ad tech players with proprietary data or strong client relationships can defend their value; commodity middlemen are most exposed to consolidation.
• Agency margins increasingly rely on proprietary trading and rebates because core media buying and planning is now priced as a commodity.
• Premium publishers should simplify their processes and build or white-label controlled solutions to keep money flowing to working media and protect CPMs.

This week Ian's away, so Justin Lebbon is joined by Sean Wright, chief insights and analytics officer at Guideline, for a data-led tour of where brand budgets are actually flowing. The conversation moves from Meta's scandal-proof growth to the heating DSP wars and the awkward economics of shifting money out of linear and into streaming.

• Meta keeps powering through: Scandals tracked back to 2009 show little impact — revenue up ~21% on average in the quarter after, with agency ad spend holding around 17%.
• A slowdown that isn't about morals: Wright's hypothesis is that Meta's "AI everything" tools are driving up costs and underdelivering ROAS for sophisticated brands, while SMBs love the point-click simplicity.
• The top-three squeeze: Of every £3 spent, two go to the major platforms, leaving thousands of publishers fighting for scraps.
• DSP wars: Four players (Google/DV360, Trade Desk, Yahoo, Amazon) now hold ~85% share, up from ~75%. Amazon has jumped from under 10% to just under 20% in ~18 months; long-tail DSPs (MIQ, StackAdapt, etc.) are losing ground.
• Trade Desk feud: Leaked Holdco memos and public sparring over take rates and transparency — but Guideline's Q1 data shows Trade Desk roughly holding, with ~13% global growth versus low single digits in the US.
• Streaming vs linear: Consumption hasn't left linear at the rate dollars did. Only ~25c of every US linear dollar lands back on streaming (~15p in the UK, ~20c CAD in Canada), partly due to streaming CPMs pricing 5x above cheap linear.
• Pricing nuance: The big three are near line-priced globally, while UK streaming looks underpriced versus US/Canada equivalents — an opportunity to grow rates or stay ultra-competitive.
• OTT in freefall: Video/OTT CPMs have fallen almost 40% in 18 months, driven by an influx of ad inventory, programmatic shift, and Amazon's low-price entry — with FAST/BVOD around 85% transacted programmatically.

• Meta's scandals — tracked since 2009 — have barely dented growth: ~21% average revenue rise in the quarter after, agency spend holding ~17%.
• The recent slowdown likely stems from Meta's AI tools raising costs without delivering ROAS for sophisticated brands, even as SMBs embrace the simplicity.
• Four DSPs now control ~85% of market share; Amazon surged from under 10% to nearly 20% in about 18 months while long-tail DSPs lose ground.
• Despite leaked Holdco memos and public sparring, Trade Desk is roughly holding — ~13% global growth but only low single digits in the US.
• Streaming isn't replacing linear dollars: only ~25c of every US linear dollar returns to streaming, partly because streaming CPMs run ~5x cheap linear.
• Video/OTT CPMs have dropped nearly 40% in 18 months as inventory floods in and buying shifts programmatic — UK streaming remains notably underpriced versus US/Canada.

After a week's break, the hosts return for results season and find the big platforms posting a spectacular year — alongside eye-watering AI investment that the markets, for now, are happy to wave through. Ian walks through the numbers and the Wall Street reaction; Justin presses on what it all means for the advertising and media industry.

In this episode:

• Alphabet's results: Search ad revenues up 19% year on year (vs 17% the prior year), YouTube up 11%, total ad growth 13%, and cloud revenues up a "massive" 63%. Far from cannibalising search, there's a case AI is accelerating it.
• The cloud CapEx story: AI investment is now showing up in cloud growth across the board — Google Cloud +63%, AWS +28% (its fastest growth in 15 quarters), Azure +40%.
• Regional splits: US ad growth up 23% (vs 17% prior year), EMEA at 12%, APAC at 22% — an unusually wide US/EMEA gap worth watching.
• YouTube vs Search: Once the faster-growing engine, YouTube has now trended below search for several quarters.
• ~$200bn in AI spend — and markets didn't blink: Investors don't punish spending per se; they punish spending without a visible return. Diversified revenue streams (Alphabet, Amazon, Microsoft) buy more tolerance.
• Meta's single pillar: Growth around 23%, but advertising is ~90–98% of revenue, leaving it more exposed. Q2 guidance — possibly deliberately conservative — weighed on the share price and reflected a deeper concern about reliance on one engine.
• How Meta is growing: Pricing up ~12% and impressions up ~19%, fuelled by more (AI-driven) content. WhatsApp scaling from ~1m to ~10m conversations a week.
• Where the growth really comes from: SMBs (≈85% of the market) and likely Chinese advertisers. Large advertisers are decelerating — guideline data cited shows US Meta growth falling from 27% ('23–'24) to 14% ('24–'25), UK 15% to 10%, Canada ~10% to 4.5%.
• Is the money leaving? Probably at the margins — but not obviously to traditional media. More likely diversified to other online platforms like TikTok (a Bloomberg report cited ~60% revenue growth expected for 2025).
• The measurement angle: Advertisers running geo-testing and switching off spend to test incrementality — echoing Google's Latin America episode. A friendly disagreement on whether moral/regulatory pressure (EU breach claims) is pushing brands to dig deeper.
• Amazon: Advertising up 22% ex-currency, AWS accelerating to nearly 30%. The hosts argue advertising and AWS — not retail — drive the profit. Roughly ~$200bn annualised AI spend; shares up ~4% post-close.
• Coming up: Guests in the next few weeks, including Guideline, plus a proper measurement discussion.

As always: this is most definitely not investment advice.

• AI hasn't hurt Google search — it may be helping: search ad revenue accelerated to +19% YoY, with cloud up 63%.
• Markets tolerate huge AI CapEx (~$200bn at both Alphabet and Amazon annualised) when growth visibly follows the spend.
• Meta's vulnerability isn't its growth (~23%) but its reliance on one engine — advertising at ~90–98% of revenue.
• Meta's growth is increasingly powered by SMBs and likely Chinese advertisers; large-advertiser growth is decelerating across the US, UK and Canada.
• Spend leaving Meta isn't obviously returning to traditional media — TikTok and other online platforms are the more likely beneficiaries.
• Advertisers are increasingly using geo-testing and switch-offs to prove incrementality, with moral/regulatory pressure prompting deeper scrutiny.

This week on Unfiltered, Justin Lebbon and Ian Whittaker dig into three of the spiciest stories in B2B media. They start with NBCU's unusually pointed attack on Nielsen — not on methodology, but on the claim that its measurement is devaluing media companies and impacting shareholder value. They then reframe retail media's surge toward $190bn as a capital-markets story driven by high-margin promises to investors, before closing on Havas and Publicis Q1 results and why North America remains the growth engine.

Highlights

• NBCU vs Nielsen: Why the line of attack has shifted from measurement methodology to shareholder loss, SEC sensitivity, and the litigious nature of US class actions.
• The power of "the gauge": How a simple, trusted chart shapes how CEOs, CFOs and CMOs allocate spend — and the YouTube "is it TV?" parallel.
• Retail media as a capital-markets story: Walmart's high-margin (70–80%) ad revenues, the rerating logic, and the promises retail CEOs and CFOs now have to keep.
• Incrementality warning: The risk of siloed, closed-loop ROI that simply discounts to existing buyers — and the "vicious circle" of diverting brand budgets into retail media.
• Two universes for agencies: Publicis (~4.7%) and Havas growth led by North America; AsiaPac and China strength; share prices lagging amid AI concerns.
• The case for agencies: Healthy ~14% margins, the danger of self-commoditization on price, and how AI plus a renewed focus on knowledge and creative could lift margins.

• NBCU's attack on Nielsen targets shareholder value, not just methodology — a potentially powerful weapon given US litigation and SEC sensitivity around perceived share-price impact.
• 'The gauge' matters because CEOs and CFOs trust a simple chart, and their view of the media landscape filters down into how CMOs allocate spend.
• Retail media (heading to ~$190bn by 2026) is best understood as a capital-markets story: ultra-high-margin (70–80%) revenue that drove rerating promises retailers must now keep.
• Diverting brand budgets into retail media risks a vicious circle — weakening the brand makes advertisers more reliant on retail media to get noticed.
• Brands should demand proof of incrementality rather than accept closed-loop ROI that discounts to buyers who would have purchased anyway.
• Publicis (~4.7%) and Havas show solid organic growth led by North America, but share prices lag on AI-driven worries about the agency business model.

In this week's Unfiltered, Justin Lebbon and Ian Whittaker open with Accenture's test of a new agency model and what it means for the traditional cost-plus economics that AI is now making unsustainable. They argue the real threat to the holdcos isn't just pricing — it's that their profits remain concentrated in media, leaving them without the balance sheet and revenue diversification to absorb a pricing reset the way Accenture can subsidise from its consulting base.

From there the conversation turns macro: US inflation, consumer sentiment, the reliability of official data, and the looming risk of European jet-fuel shortages tied to the Middle East conflict. The pair then break down Bill Ackman's $64bn Universal Music deal — why music rights are now treated as a financial asset class rather than a creative trophy — before closing on the regulatory squeeze facing social platforms and the broadcaster pushback against YouTube.

Highlights:

• Why the cost-plus agency model is "unsustainable" in the age of AI, and the MiFID II analogy for how a pricing reset rewards diversified players.
• Accenture's "secret weapon": infrastructure, scope across the client's business, and a balance sheet to absorb performance-linked volatility — three things holdcos largely lack.
• The case for repositioning advertising from commodity to investment, and agency services as a premium product.
• US inflation at 3.3%, a 21%+ rise in gasoline prices, record-low April consumer sentiment, and downward-revised February payrolls.
• The jet-fuel risk: UK over 40% dependent on Strait-derived jet fuel, with Vietnam already rationing.
• Why Q1 results season and the November midterms act as a "hard stop" on the macro picture.
• The Universal Music deal: limited cash, a US listing move, a 20x EBITDA valuation, and Vincent Bolloré as the swing vote.
• Music rights reframed as long-duration, bond-like income assets — and the shift from creative bracket to financial asset class.
• Governments increasingly seeing it as a "vote winner" to go after tech platforms — a Standard Oil-style populist turn.
• Broadcasters (Channel 4, ITV, TF1) pushing for fairer YouTube rates and TV-style regulation, plus the under-asked question: YouTube vs TikTok, not YouTube vs TV.

• The cost-plus agency model is unsustainable under AI; Accenture's move to a subscription-like, outcomes-focused service is the genuinely interesting shift.
• Holdcos have deep infrastructure, relationships and talent, but their media-concentrated profits leave them vulnerable to a pricing reset they may lack the balance sheet to absorb.
• Macro signals are mixed: US inflation at 3.3%, gasoline up 21%+, sentiment at record lows in April, but jobs data positive — with jet-fuel shortages a real European risk if the Strait stays disrupted.
• Ackman's $64bn Universal Music deal is far from done; it hinges on Vincent Bolloré being convinced a US listing delivers value, with limited cash on the table.
• Music rights are now treated as a financial asset class — predictable, inflation-linked, bond-like income — rather than a creative trophy.
• The biggest long-term threat to tech platforms may be regulatory: governments increasingly see going after Big Tech as a vote winner.

Recorded after the Future TV Advertising Forum in Sydney and a trip to New Zealand, Justin Lebbon and Ian Whittaker dig into how broadcasters change the narrative by collaborating and building what buyers actually want — and why belief in the product matters as much as the numbers.

• Broadcasters on the front foot: Future TV Sydney showed how to lead with positive product framing, outcomes platforms and collaboration rather than dwelling on declines.
• Faith in the product: Drawing on the newspaper industry's mistakes, Ian argues losing belief in your product is fatal — broadcasters hold 70+ years of skill selling premium video to mass audiences.
• Out-of-home's NZ surge: A collaborative measurement solution with buyer buy-in helped OOH go "gangbusters," picking up digital and retail media budgets — prompting TV players like Channel Nine to diversify into the space.
• The "premium" problem: If cheap delivers the outcome, is premium still premium? The hosts debate whether the industry should reframe around "quality," environment and long-term brand effects rather than immediate outcomes alone.
• Accenture's back-to-basics model: A retainer-plus-percentage structure with no kickbacks, no principal media and no opaque deals, trialled in Australia — and whether transparency-led pricing can reset the value of agencies and advertising.
• Measurement as power: Nielsen's delayed February gauge report and the VAB's "interference in markets" accusation underscore that whoever sets the currency sets the terms of trade. Justin makes the case for buyer-backed JICs like Barb.
• AI hype meets geopolitical risk: With ~$700bn in tech AI spend looming, IPO pressure on OpenAI and Anthropic, and Middle East funding under a cloud, the Paramount–Warner deal and the wider M&A outlook may be at risk.