‘There’s two ways to build these platforms’: OpenAI’s ads boss David Dugan on going the other way

OpenAI's global ads boss David Dugan said the company is building its ad platform rapidly by embracing agency holding companies and technology partners rather than controlling all client relationships directly, aiming to create a durable home for ad dollars. The ad pilot, launched 19 weeks ago, has already introduced cost-per-click bidding and a Conversions API, with most volume now running through CPC.

‘There’s two ways to build these platforms’: OpenAI’s ads boss David Dugan on going the other way OpenAI’s ad product moves fast. Nineteen weeks or four months https://digiday.com/marketing/being-very-careful-weeks-after-unveiling-ad-plans-openai-works-to-control-the-message/ in and it’s already struck commercial partnerships, built measurement tools https://digiday.com/marketing/openai-builds-tool-to-track-whether-chatgpt-ads-convert/ and launched an ads manager https://digiday.com/marketing/openai-has-quietly-launched-its-ads-manager-as-it-races-to-build-out-its-ads-business/ . Ask about what’s underneath those moves — independent ad verification, how trading deals evolve — and the answer slows right down. That gap between what’s shipped and what’s still vague is the real story of OpenAI’s ad pilot so far. Because those are the answers that will decide whether this becomes a durable home for ad dollars, not just attention. That’s why Cannes Lions could go down as one of the more consequential weeks for the fledgling ads business in hindsight. Global ads boss David Dugan https://digiday.com/marketing/who-is-openais-global-head-of-ads-david-dugan/ was everywhere, listening to creators in villas, hosting CMOs between the SPA Villa Belle Plage and the Majestic and meeting holdco bosses over dinner. Eventually those interactions will turn into moves. Until then, what little details he said matters just as much, if not more, than what he didn’t. This conversation has been lightly edited for clarity. What’s next for the ad pilot? We started with CPM bidding at launch — a natural beginning, and there was active feedback from the market — but demand for CPC, cost per click https://digiday.com/marketing/openai-turns-on-cost-per-click-ads-inside-chatgpt/ , came next. It was 10 weeks from launch to launching CPC, and now the majority of volume on the platform runs through CPC. Next we’re thinking about conversions https://digiday.com/marketing/openai-turns-on-cost-per-action-ads-inside-chatgpt/ : we’ve launched the Conversions API https://digiday.com/marketing/openais-chatgpt-ads-get-their-first-measurement-partner-in-liveramp/ , and we’ll be looking at optimizing for conversion. There’s this trajectory of every X weeks, every X months, we roll out more features that are based on feedback we’re getting from the market. Why are you building the ads proposition so quickly? I’d attribute that to two things. One: we’re an AI-first company, so we’ve got engineers that use OpenAI’s products to build our product. We have Codex inside, and we’ve got the enterprise platform, so that’s advantageous for us in speed, which you’d expect. Two: we don’t have legacy infrastructure behind us, everything is forward facing, so we’re more nimble. We publish a roadmap when we talk to clients, showing them features of what will come in 30, 60 and 90 days. We’ve gotten great feedback from the market regarding our pace, so that’s what we’re focused on. The basics. We’re very collaborative with our agency and technology partners https://digiday.com/marketing/pitch-deck-how-chatgpt-ads-are-being-sold-to-criteo-advertisers/ . I think this is a very purposeful way we launched. Does OpenAI see itself as a walled garden? There’s two ways to build these platforms. One: you try to control all those client direct relationships and keep all those private and closed. Or two: what we’ve done; embrace the agency holdcos, embrace technology partners, and do it together as an industry. That’s helped us move faster, but it’s also given us insights from these partners. They have lots of diversity of vertical industries, geographies, and different size clients. Enterprise client versus torso mid-market clients might be different, so the insights we’re getting are extremely helpful for our product and engineering teams, and help us lay out what needs to come next on the roadmap. You hired The Trade Desk’s Samantha Jacobson as vp of partnerships. Will you continue to work with multiple adtech https://digiday.com/marketing/ad-tech-is-lining-up-behind-openai-its-been-here-before/ partners, or are there plans to build more internal adtech? adtech https://digiday.com/marketing/ad-tech-is-lining-up-behind-openai-its-been-here-before/ partners, or are there plans to build more internal adtech? These are two separate things. One: we’re aiming to recruit the best talent in the industry, and I think diversity of thought, bringing people in from different backgrounds, is extremely important for the business on the partner side. Different advertisers have different preferences about what partners they work with, and we want to meet them where they are. Some advertisers work with one agency exclusively, others might have multiple. Some work with an agency and a technology partner. We’re trying to meet them where they are, which means we need to have a range of partners. We’re focused on getting an established product that’s driving advertiser performance, which stays true to the user principles. Our ads experience is very rooted in our principles and getting the market to understand how we’re going to roll it out via the roadmap going forward. We’ll iterate from those core principles, and think about new tools, probably eventually new surfaces, but where we started, is a very clear execution against those principles. Do you want to automate the entirety of ad creation on the platform? Today, we don’t have automated creative https://digiday.com/marketing/openai-moves-to-automate-ad-creative/ execution as part of the platform. Different cohorts of advertisers have very different needs and expectations. SMBs in the long tail would be very curious about tools to help them generate creative to run ads in the simplest way possible. That’s dramatically different from a large enterprise that’s invested in their brands over decades and has very strict creative controls about how they serve creative, whether it’s text, image or video. They’re going to be extremely focused on control of that creative message with their agencies, and we understand those differences. As we think about new tools for creative, we will think about those extreme differences between small, medium and large businesses. There’s such dramatic differences that there’s no such thing as one size fits all. You have to give the market choice and control in those systems. How do you see the commercial model with agencies, and those trading arrangements, evolving as the ads business scales? In terms of how agencies use us, there’s a framework which is how do you build the work inside an agency, how do you create the work, and then how do you build reach. We have some agency partners using our enterprise tools to help run their in-house operations — How do you give knowledge workers inside an agency tools to do their job better, more efficiently, to build client services? Then using the ads product to generate user reach. In the future, there’s a really interesting flywheel in terms of taking those ads results and feeding them back into the organization to get smarter on creative optimization. We’re not thinking about that from a commercial or pricing standpoint, but we think connectivity is fascinating. Clearly, for every agency, it’s no longer a question of will they use AI tools to power their businesses. Everyone’s made that decision. Now it’s how you use those tools and differentiate versus the market. You’re encouraging test budgets for now, but where do you see the majority of spend eventually coming from? The starting point is curiosity, interest and passion from advertisers, whether it’s CMOs or agencies, about this new channel. It starts from that user behavior shift; there’s a way people use ChatGPT that has evolved for everyday task completion. When somebody opens the ChatGPT app, it’s for a purpose: a task you want to complete. You’re not scrolling — we are not focused on time spent per day — this is not an entertainment app. You’re going in for productivity and to conclude a task. That makes this channel so different, because we’re not chasing time. This is not about attention minutes in the app, no consumer wants to spend more time on their phone. When it comes to channel selection, the starting point for advertisers is seeing this consumer behavior shift, and needing to be in this space to learn how user behavior is evolving. We’re the first platform at scale offering that opportunity, and I think advertisers will decide on their own how they’re going to budget for that, and where they source it from. We’re less focused on where it comes from, but the starting point for advertisers’ is the user experience, and to see this broad shift in consumer behavior. How are you balancing ads without damaging the user experience? We started with the principles about being super focused on user trust. We have this incredible asset, 900 million plus weekly active users that love this app, who come in for a very specific purpose. Everything we’re doing with ads is grounded in how we ensure ads improve, increase, and are additive to that experience. That’s the foundation. We are very watchful on user metrics to ensure that as we introduce ads in a new market, that those metrics stay strong. That’s the first and most important thing for advertiser ROI. Of course, we’re focused on it, but it’s secondary to user metrics. Is there a path for publishers to get a cut of ad revenue, or is that siloed from the ads business? We’re focused on the ads product as we have today. The pricing model we have for bidding in the auction, CPM, CPC, eventually getting to optimization — I’m purely focused on that experience right now, so that publishers hasn’t been a conversation that’s come up in my Cannes week with advertisers, I’ve been totally focused on the core experience they have, and the roadmap for for the core ads product. More in Marketing How the Chicago Bulls retooled their sponsorship business to meet CMO data demands https://digiday.com/marketing/how-the-chicago-bulls-retooled-their-sponsorship-business-to-meet-cmo-data-demands/ As sports sponsorship spending rises, CMOs are looking for hard evidence to justify their deals. Teams and franchise owners are responding, but there’s competition. OpenAI ads boss David Dugan on third-party measurement: ‘it’s a natural step’ https://digiday.com/marketing/openai-ads-boss-david-dugan-on-third-party-measurement-its-a-natural-step/ Since OpenAI refuses to share chat data, as it goes against its principles, external proof may be the only way to win advertiser trust. Cannes Briefing: The Cannes confessional https://digiday.com/marketing/the-expense-of-cannes-is-getting-out-of-control/ Cannes is soft win. Attendees showed up and so did everyone else and the showing up is the whole transaction.