When Garrett Gan looked ahead at his career after college, bringing transparency, simplicity, and organization to the way ad vendors find and communicate with their market wasn’t even a faint possibility.
Gan started out in mechanical engineering, switching majors partway through college to graduate with a degree in public health on the premise that he wanted to be a doctor. But instead of heading off to med school, he entered the job market in 2008 when it was in a terrifying and dizzying free fall and took a job selling online ads for the Orange County Register. It was after a few years in advertising, at his job at Quantcast, that he discovered a complex industry in dire need of innovation.
Gan’s agility and open-minded approach is infectious and unique. Instead of hitting on a single startup idea and running with it until it either took off or fell flat, Gan started gathering data on what wasn’t working in the ad vendor space and sketched out multiple ways he might solve each of those problems.
Once he started working on a specific product, Gan kept his method for tackling the problems flexible, altering his vision for his company completely when it became clear the original concept (though still useful and profitable) would be a risky and enormous investment that only the most optimistic billion-dollar companies might dare dive into. This flexible approach has clearly paid off.
The final product, Thalamus, has been an enormous hit, servicing 85% of AdAge’s Top Digital Agencies List by Media Spend within a year of launch.
I recently sat down with Gan for an interview, peppering him with questions about Thalamus, that first product, the unique challenges in an informational marketplace, and much more.
How did the idea for Thalamus form?
Prior to Thalamus, I was at a company called Quantcast, and then left to join a mobile startup. We were trying to build a mobile targeting data set, to layer on top of real-time bidding ad exchanges, through a free product we’d offer to app developers. Things didn’t really pan out for the company, and it was right after that I partnered with a friend I met at a mobile meetup years back to start an idea I had.
Well, no. We were actually working on a completely different product.
What was this earlier project, if you don’t mind sharing?
Sure, we were building a sales force for agencies where these different agency media planners could send, what they call, “requests for proposals.” It was kind of like a workflow automation tool. They could go in and they could send these RFPs to different ad networks and different publishers that sold advertising. They could also keep track of their workflow, generate these things called Insertion Orders to execute campaigns, and collaborate with different ad partners on placements. Then you could traffic these placements into a third-party ad server and pull in reporting from the server. It was a unified dashboard for media buying that enabled collaboration with different ad partners on the actual campaign, launching a whole campaign all within one platform.
We actually had some traction there. We got the mobile agency for Johnson & Johnson to use it and they sent about $10 million in requests for proposals, but we realized that there were a lot of issues with that. There were so many individual customizations that these different agencies needed. They had different workflows that they followed, they needed different integrations. People wanted many specific things, like different channels — we were just focusing strictly on digital, but they said they wanted radio and print. We needed to do integration with Nielsen and things like that.
By the time we collected the list of all these different things we needed to do, we realized we would’ve needed a massive team, and even then. There are some companies who are trying to tackle this right now, like AppNexus and the Rubicon Project, which are some larger players in the space. And even for them with a lot of money in the bank — one of them being a billion dollar company — they’re having some issues tackling this kind of automation of buying advertising across every single vendor.
Once you realized that you needed to create the Thalamus that exists today, how did you validate it?
I always knew there was a need in the space — one singular directory or database of every single vendor, all their capabilities, all their contacts, all their marketing collateral, all their ad specifications, all of this network data, and the ability to search, filter, find the right product for you, and reach out to them.
I knew generally that was lacking in the space, but I didn’t really know how much of a pain point it was. I just had to keep talking to anyone who would take a meeting with me and running new ideas by them. We only built stuff that we were 80-90% sure that people needed. Validating the idea was having the belief in the back of my head that this is the way that the industry should work, and getting feedback, getting opinions of anyone who would listen, and eventually getting to the point where now it’s starting to become a main thing in the industry, and we’re starting to get a name.
This is starting to become a very crucial media planning tool for people in the industry. In the future — in the next couple of years — I could definitely see it becoming this de facto standard of advertising data where people’s behaviors have shifted. So instead of going back and forth over email to get one document or seeing all this different information, they could come to one place and be able to see everything they need about different partners and reach out to them.
What’s been the biggest surprise you’ve encountered in the founding role?
I’ve been in the industry for a while, so I was surprised at the sheer amount of learning that had to take place prior to starting something like Thalamus. The more people in advertising I talked to about their needs, the more I understood how little I actually knew. It took me a while, probably at least a year, or year and a half, to really understand every little nuance of need within the industry — which matters a lot when you’re building a product.
I think I’m at the point right now where I can see 90% of the angles of what people need and what kind of data is important for them, how the product should work. The journey can be really long because it just depends on the industry you’re in, but sometimes there’s a massive amount of information to learn.
Thalamus now provides service for 85% of AdAge’s Top Digital Agencies List by Media Spend, which is some incredible growth. How have you managed to grow so fast while still growing responsibly?
We have a moderation system in the backend. We rebuilt the engine. When someone signs up, they have to verify their email so when edits come through, we’re actually able to see all the different data that’s being entered into the database. We make sure that all of the information is from representatives from the company or actual brands and agencies, and that’s how we’re able to keep the data and the reviews accurate.
In terms of growth, it’s pretty hectic sometimes, but at this point we know we’re building something that is absolutely crucial and the next couple of years will change the way people transact in advertising.
I’m intrigued with the name choice. The thalamus is a really interesting and important part of the brain that isn’t talked about much in lay science news or everyday conversation. Were you a part of choosing the name?
I came up with the name, yeah. The first startup (the sales force) we thought of it as a centralized media buying dashboard, like an operating system. The first thing that popped into my head was, what are some portions of the brain? We thought of thalamus and the cortex. Cortex was taken, but thalamus.co was available. It was analogous because the thalamus is where all the sensory information flows to before being processed in the prefrontal cortex.
Would Thalamus be considered a marketplace business?
Yeah, I guess you could say it’s a marketplace business. It’s more like an informational market — we’re not actually taking a cut of the transaction.
We wanted to keep the barriers really low. We made it free, and the product is unique so nothing else like this exists. That’s why, when people find out about this, they immediately email it to all their colleagues, and that’s why we’ve been able to grow so quickly on the media-buying side. Also, since the database is fueled by ad partners inputting their data, we need to keep it free for them as well; we can’t charge them to input data, so we kept it free.
On the frontend we made the editing process very simple, very streamlined but, obviously, with the moderation process on the backend. Then we are going to do the insight business model where, as an informational marketplace, we are able to track what marketers are looking at and what files they’re downloading, and what companies they’re looking at, what filters they’re using, what keywords they’re typing in.
Essentially, what we’ll be able to do in the future is, for the ad partners, the people using our paid insights product, we can tell them exactly what marketers are looking at on their profiles, downloading their media kits or case studies, which ones are using which filters, or typing in certain keywords. They’re able to see all the marketers that have expressed interest and have intent as to their offerings, and they can just harvest that demand.
What’s the biggest challenge of running an informational marketplace like Thalamus?
For most marketplace businesses the challenge is getting the data seeded, getting each side of the marketplace using it. It’s getting enough supply of data, to the point where it’s useful for the marketers. It’s swinging back and forth, convincing both sides, until you’re at the point of having enough information to where it starts getting really useful for both sides of the marketplace where the supply side wants to have their information in because they know a lot of marketers are using it and the marketers are using it because they know there’s a lot of data in there that could be useful for them.
Initially scaling the dataset was probably the hardest part, as any marketplace company will tell you. If you go to the site, you can see very specific details on what kind of data points, and what kind of metrics, and what kind of taxonomy these different media buyers need. That was another challenge, figuring out the exact things they need. The ad verification or the campaign measurements partners field that they need, that that part is filled in, or the targeting portion, or the publisher partners, or inventory partners. All those kinds of information, if you were not within the industry or you didn’t know that much about the industry, it would’ve been impossible to know that this is the information that people look for.
How did you find out exactly what each type of vendor would want to know, what kind of information you needed?
Actually, when I was at Quantcast, I was already researching a lot of partners. I was thinking about this idea — not even building a company around it — but I was already building my own database of all these different ad partners a long time ago. I already knew after looking at different media kits exactly what these ad partners wanted to sell or list. Then I just came up with a standardized format of all the different data points that cover every single ad partner.
What’s next for Thalamus? You guys have tackled quite a lot in a short time.
I think in the next year we’ll be starting to partner with agencies to get them to mandate that their partners input data. Then it’s just scaling the data set and building out more features. A lot of the features have already been built out. There’s probably going to be a bunch more, but I think a huge portion of the foundation has already been built up.
Looking to grow? Learn how other successful founders grew their business.
Featured image credit. Other images provided by Thalamus.