David Dwek

May the 4th be with Shopify + Claude. Every merchant just got handed a superpower. The Claude Connector can write to your live store now. If you're using Sidekick heavily already, this is a blessing. Who this is for: - If you're running 100+ SKUs and do your own catalog work - If you're tired of doing bulk updates one product at a time - If you want KPI answers in chat instead of digging through admin reports - If you want to spec a discount, inventory move, or metafield change in plain English and have it land in the store - If you're done copy-pasting SKUs between spreadsheets and admin tabs Who this is not for: - If you're a single-product DTC store with under 50 SKUs, the admin panel is faster - If your team is not ready to handle approvals on writes, this needs a process - If your catalog is messy at the source, the connector will move faster than your data quality What you can do in chat(sample prompts): - Archive every product in the Winter 2025 collection - Set inventory for SKU LP-SAND-M to 50 at Brooklyn - Create a 15% off discount code WELCOME15 for first-time customers - Top 10 products by gross sales this quarter - Update the Linen Pant price to $79 across all variants Early learnings from my test: - Read paths like Get Product and Run Analytics auto-approve - Writes pause and show the plan before they fire - Raw GraphQL mutations need approval every single call - The agent will explain consequences before it acts. Switching a category triggered it to flag that 20 attached metafields would become irrelevant - The agent will also do exactly what you tell it to do even when it knows you're wrong. The merchant is the final guardrail A few patterns that made my prompts work better: - State the SKU or product ID directly when you have it - Be specific about scope. Tagged discontinued from 2024 beats old products - For analytics, name the metric and the time window I put together a reference doc with 15 prompts I ran on the live store. Comment CONNECTOR and I'll send it over.

26

17 Comments

When Dave Pickles and I started TTD, we wanted to create a platform that was more transparent and more aligned with buyers. We wanted to create closer alignment with brands and agencies. We were one of the few DSPs that didn’t focus on disintermediating the agencies. I stand by our choice to partner with agencies. I stand by our choice to be transparent, even though I think we at times have been too open (as I shared at Marketecture Media last week). Over the last year or two, things have gotten much harder for agencies. Lately it seems that there are those who want to wave the flag of transparency publicly, but run from it in practice as they arbitrage in the inefficiencies of programmatic. Or in the most ironic programmatic practice, principal-based buying. (More on this next week. I know…I’m such a tease.) It bothers me when leaders of non-transparent business models are critical of those of us who are setting the bar—especially when they advocate for moving dollars to more opaque platforms and transaction methods. As Terence Kawaja said to me last week as we sat backstage at Marketecture, the long arc of programmatic, like any market, will inevitably bend toward efficiency and competition. Those who succeed will be those who embrace transparency and offer clients competitive, objective value. Thankfully, almost all of our clients, including the world’s largest brands and agencies, share this point of view. TTD has not “failed" any audit ever. Ruben Schreurs, CEO of Ebiquity plc, assured me today that he was unaware (until after published) that his firm would be quoted or used in any articles about TTD. Relatedly, we will not disclose the bills of all of our clients and partners to one of them simply because they assert ambiguous audit rights.   Working with our clients, we always welcome objective and rigorous analysis of our processes, fees, and the value we deliver for clients.

1,448

56 Comments

Every sponsor wants a discount. Most publishers cave. 4 ways to stop getting low-balled on CPMs: 1. Build a high-value audience Not an easy fix, but it has to be said. 2. Target brands with real budgets Look at company size, LTV, AOV, and industry. A $500/month SaaS marketer can spend far more to acquire one customer than someone selling $90 jeans. 3. Focus on the right verticals B2B SaaS, finance, and enterprise tech pay premium CPMs. E-commerce and DTC brands typically don't. 4. Be confident in your pricing If you have a deep sense of insecurity (like most people), you'll discount too quickly. Acknowledge it. Then work on it.

4

2 Comments

Tech vs Service. Today, we manage two companies. Pontiac Intelligence: A Demand Side Platform to buy advertising space. A technology company. Clients pay a transparent licensing fee. Programmatic Mechanics: A trading desk model where we provide high quality traders and data scientists for organizations looking for personnel assistance in managing campaigns. Comes with dashboarding service across all channels. We spend a lot of time discussing the value of a service business versus technology. And, I just think the times are changing. It's not that technology is dead. It's just that really high quality service models might be able to retain clients for longer, and that might start to even the valuation playing field. Technology isn't dead. The ability to switch between technologies is easing. The moat is drying out. Every single technology companies moat is drying out.

30

4 Comments

Excited to share that Aperiam has invested in Bedrock Platform. Obvious statement of the century: programmatic advertising has a complexity problem. The infrastructure that was built to give media buyers control has, over time, done the opposite. Burying teams under layers of operational overhead, legacy tooling, and opaque supply chains. Bedrock is built on a simple but powerful thesis: what if you stripped away the bloat and rebuilt the stack around what actually matters... precision, transparency, and speed to execution. Their platform combines AI-driven campaign automation with curated, premium inventory across video, CTV, DOOH, and audio, and aims to be flexible and get smarter over time. Plus, the team (Shane, Austin, Ryan, James) has deep roots in the infrastructure layer of ad tech and understands the plumbing better than almost anyone. We believe the next generation of programmatic will be leaner, faster and AI-first. Bedrock is building that.

108

14 Comments

AI is no longer optional for publishers — it’s foundational. In today’s omnichannel world, the complexity of ad monetization has outpaced what humans can manage alone. That’s why AI isn’t just a nice-to-have — it’s a must-have. In this article, I explore how platforms can partner with publishers to deliver AI solutions that are built with them, not just for them. From automating ad ops to unlocking untapped inventory, the future of monetization lies in purposeful, integrated AI — and the next frontier is agentic AI. 📈 The goal? Less friction, more revenue, and smarter decisions at scale. How is your organization approaching AI in ad monetization? https://lnkd.in/gvKpGkkT

124

Businesses invest heavily in generating leads: through paid advertising, events, partnerships, and other hard-won channels. But what happens after the first meeting? For many companies, follow-up relies almost entirely on email sequences. But while email remains important, response rates continue to decline, especially with older or colder leads. So what about WhatsApp? We recently ran a re-engagement campaign for a client targeting leads who had stopped responding to email. By introducing a structured, compliant WhatsApp outreach strategy, we achieved a 10% positive reply rate from contacts who had previously gone quiet. If you’re investing in lead generation but not maximising re-engagement opportunities, it may be time to rethink your follow-up strategy. If you’re interested in exploring how WhatsApp can help you reconnect and convert more of your existing leads, feel free to get in touch.

11

The adtech world seems to be at war. Just as AI threatens the meagre market share the open web has. Imagine. We had an open adexchange. Every dollar tracked from advertiser to publisher, data and middlemen. All linked to verifiable companies, any of which advertisers or publishers can block. AI fairly and transparently looks at quality, fraud, viewability - open to innovative vendors. Private deals are easy, measurement by a well used conversion api fairly open to actors, cohort based ids for tracking and frequency via a clean room. If we had that trust would rise, take rates increase, performance up. We could save the open web and maybe journalism and society with it. I want that. Who's with me? Let's build it. A coalition of the willing. #AdTech #OpenWeb #DigitalAdvertising #Transparency #Innovation

79

29 Comments

U.S. TVOS platform revenue by company: 1) Roku OS (Roku) - $3.5B (25%) 2) Fire TV (Amazon) - $2.4B (17%) 3) Android TV (Alphabet) - $2.2B (15%) 4) Tizen Smart Hub (Samsung) - $1.5B (11%) 5) Apple TVOS (Apple) - $763M (5%) 6) WebOS (LG) - $759M (5%) 7) SmartCast (Vizio) - $738M (5%) 8) Xumo (Comcast/Charter) - $189M (1%) 9) Other - $2.4B (17%) https://screenwa.rs/a3142e

28

5 Comments

Consumer expectations around video have changed. For a long time, video strategy didn’t. For years, video marketing was built around exposure (impressions, views, completion rates). And if we’re being honest, most brands were comfortable there. Not because those metrics proved engagement, or much of anything, but because that’s what everyone else was measuring. It became a bit of a comfy blanket - familiar, easy to benchmark, and safe. But when you really think about it, the traditional video experience is a little strange. It’s like walking into a room full of people, giving your pitch, ignoring every raised hand, refusing to answer questions… and then leaving the room before anyone can respond. That’s essentially how video has worked for years. Meanwhile, consumer behavior kept evolving. People don’t want to be forced into anything anymore. They expect immediate access to what interests them in the moment. -The place they see. -The product that catches their eye. -The experience they want to explore. -No searching. -No friction. -Just access. And that’s where video is heading. Video isn’t just a storytelling canvas anymore, it’s the access point. A gateway where viewers can instantly explore the people, places, products, and experiences inside the content itself. When that happens, video stops being just a pretty screen and it becomes a living interface for discovery, engagement, and action. This shift has been quietly building for years because the future of video won’t just be about what people watch. It will be about what they can do the moment something captures their attention. #thefutureofvideo #Clicktivated #actionablevideo

5

🔐 Running Paid Search Without Security? That’s a Risk You Can’t Afford. At Mproven, we don’t just drive clicks — we protect what matters behind them. Your Google Ads account is a digital goldmine. If it gets hacked, you’re not just losing ad spend — you’re exposing data, damaging your brand, and losing trust. That’s why every Mproven campaign includes: ✅ Secure ad account architecture ✅ 2FA on all platforms ✅ Limited-access protocol for team logins ✅ Ongoing monitoring for account anomalies 💡 We're performance marketers who actually think like security pros — and that’s why brands trust us with their paid search. Want to grow and stay protected? #CyberSecurity #PaidSearch #Mproven #DigitalMarketing #GoogleAds #SecureGrowth #PPC

7

Question for the industry - are there any ways for DSP reporting to highlight how absolutely trash this video ad placement is?? Get your magnifying glass out and you can see a 16:9 video as a bottom of the page sticky banner... My assumption is DSP would report on the parent container size (which is probably a standard in-feed video unit) however the creative is then rendered substantially smaller so it doesn't get picked up. It's an interesting one for our PIQ technology, which is built to measure inventory quality. Yes, the ad density is lower. However, this is undoubtedly crap for the advertiser and therefore bad for the ecosystem - enshittification in practice. Based on this example we've started measuring parent container, rendered creative (when possible) and the delta between the two. We've also been building minimum size thresholds for creative formats. We'll be adding them in as signals in our Ad Experience Index which, of course, values low density but also recognises that people prefer free to paid content and well integrated, useful ads are necessary to fund that business model!

39

29 Comments

After helping several health & beauty brands launch on Amazon, here’s what we actually had to nail down before anyone even saw a listing: • Brand basics first: Get your Seller Central, Brand Registry, and trademark squared away right away; especially if you’re DTC-first. I’ve seen this take anywhere from days to months, so handling it in advance always helps with the launch timeline.    • Creative direction + best practices: The look, feel, product image order, A+ content, storefront layout… all need to be aligned. Pull inspiration from DTC, but layer in Amazon best practices + competitive research so everything is prepped together. This is a collaborative effort across teams, and honestly, this alignment alone can save months.    • FBA > FBM for most brands: Prime badge, Vine access, better data. Decide early who owns it. If Amazon is your first retail-like channel and you’ve only done in-house warehousing, there’s a steep learning curve for brands' supply chain teams. Starting with an experienced 3PL or having a supply chain expert on board makes the shipments smoother.    • Category quirks matter: In skincare, expiry dates must be clear on the outside of packaging. Hazmat or liquids require specific prep and pack rules. Teams need to know how to create shipments, case packs, or mixed units early on - and navigate Amazon’s internal inventory systems.    • Timeline sanity: All these layers take weeks to execute. Plan backwards from your target launch date and run it through a project management system regardless of launching yourself or working with an agency.     Boring? Maybe. But this is what separates a frantic midnight scramble from a calm, systemized launch. What’s one pre-launch step you wish you had known earlier? ---- Prima Mode helps health and brands launch on Amazon as an Amazon-Verified store partner, so if that’s been something top of mind, happy to chat!

24

2 Comments

🚨 Amazon quietly became the single biggest bidder on the open web and somehow the adtech industry reacted like a goldfish staring at a lava lamp. One buyer driving roughly 20.5% of publisher revenue on average, publishers getting quietly cut or rerouted, and trade press mostly busy moderating another “future of AI” panel beside an espresso machine fighting for its life. The wild part? The underlying data came from a report by Playwire that practically nobody outside publisher circles bothered to read. Meanwhile, the actual market structure shifted under everyone’s feet. Amazon isn’t “leaving” publishers. It’s rewiring the pipes through preferred SSPs while everyone politely calls it “supply path optimization” instead of what it looks like: one company quietly redrawing the plumbing diagram of the open web. Jayson Dubin & Stacy Willis Hi Michael Barrett. Hi Lori Goode. Hi Rajeev Goel. Hi Magnite, PubMatic, Index Exchange, and the entire ecosystem currently discovering that “diversified demand” maybe should not mean “one company controls a fifth of my revenue.” Tiny detail. Easy to miss between conference cocktails. Thanks for the quotes. Full report in comments

10

4 Comments

I think we are on the cusp of seeing a plethora of new “buying tools”. Each will have its own flavor or pov baked in. Some will continue to work with legacy DSP’s but the majority will not. Many will activate in other ways. This will change the ad landscape quite a bit. Yieldmo

34

4 Comments

Over and over again - we see great data on how OOH drives foot traffic. In NYC, LinkNYC’s massive 4000 screen network gets deep in the city, finding target audiences where they live work and play. In beauty, a brand can own the biggest city, and spark NYC’s trendsetting audience to go to retail and buy the product, as shown in this recent case example.

42

If your edge is your ability to get a tech team to integrate with you and no one else, it’s a vulnerable edge.

14

2 Comments