S2:E5 – Advertising at Google (Sissie Hsiao)

Subscribe: Apple Podcasts | Spotify | Android

Google’s main source of revenue is digital advertising. Sissie Hsiao, VP of Display, Video and App Advertising at Google, explained what digital advertising is and its workflow. We also talked about the role of Machine Learning in advertising and tackling invalid activity. Sissie then explained her path to a leadership role.

@techwomenshow

Sissie Hsiao

Sponsors

5mm-Ad-Image-SMALL

Transcript

ES:

I’m Edaena Salinas, software engineer and host of The Women in Tech Show, a podcast about what we work on, not what it feels like to be a woman in tech. For more information about the show go to wit.fm.

Google’s main source of revenue is digital advertising. Sissie Hsiao, VP of Display, Video and App Advertising at Google, explained what digital advertising is and its workflow. We also talked about the role of Machine Learning in advertising and tackling invalid activity. Sissie then explained her path to a leadership role.

Sissie welcome to The Women in Tech Show. 

SH: 

Thanks, I’m so excited to be here and thank you for inviting me to this. 

ES: 

It’s great to have you on the show. I’m excited as well.  

First, I want begin talking about technology. This is something that has been a part of your life since you were a kid. One example is gaming, which you have been doing since you were six years old. What was your introduction to gaming? 

SH: 

Yeah, I have really fond memories of gaming and I’m embarrassed to say that I still am a gamer. I Identify myself as one. My first recollection was watching some kids play, you know, the first versions of Mario on a Nintendo console. I think that was around when I was six. I begged my parents for one and ever since then, I grew up playing video games and that really influenced my decision or actually last minute decision to go into computer science. I thought that using code to create like a world and a story and characters was just so fascinating and very creative and said that’s been a defining part of why I chose this field. I always thought I would essentially make video games someday in my career I didn’t get around to that, but I did the next best thing which is to build products and businesses that support game developers.  

Right now I build a lot of essentially a business around helping app developers, not just game developers, lots of developers, but we have a lot of clients who are app developers, who are game developers, help them find relevant users who would love their games. An also help them monetize their games through advertising. I’m really happy that even though you know I didn’t get around to making games in my career, I’ve been able to build great products and services in support of that vertical. 

ES: 

You mentioned that gaming influenced you to pursue a career in computer science and technology. You also mentioned that it was sort of a last minute decision, what was the other field you were considering in studying? 

SH: 

My father is a professor in the field of finance, so, it was senior year in high school, I was thinking about becoming a business major for my undergrad, and one day I woke up and I just had an epiphany, which was like, why can’t I go into computer science instead and worry about getting an MBA later, was the plan. The irony is I never got around to getting the MBA, but you know the process was very much a snap decision and again prior to that I had never really written any code. I think lots of girls growing up in my generation were not exposed to coding in the areas of computer science. You know, going through regular public schools here in the US, and so it was a bit of a last minute decision and I have to say a really very very good one because it ended up being a field that really much suits my passions, which is just, I love that technology is so open ended and it’s so dynamic, you know I can think of all the advancements in the past ten, twenty years. Most of the very interesting advancements to how you know how we live, behave, work, all of those things, influencing humanity, being advanced in the areas of technology.  

Technology is something that I feel like is such a powerful tool that you know can exponentially improve everything around how humans operate and work every day, bring it to my job right now. Our mission using technology is to help keep the Internet free. An open through advertising through making sure that advertisers can find their users. They’re making sure that publishers who create content can monetize their content so that it can be free. 

For users and for users to find relevant services, but also to have access to free content. But it’s just one way that  it’s really powerful thing that there’s a free Internet that everybody has. You know equal unfair access to content in a very open way. An again that couldn’t have been made possible without, the technology advancements that we’ve all been building through our careers. 

ES: 

Let’s talk a bit more about digital advertising first. For people not very familiar with how advertising works, I want to get a general idea of the workflow and some of the pieces involved. I know it’s it’s a very broad field and there are many use cases. Can you give an example scenario of how advertising works from releasing an ad to seeing the outcome sort of? What are the pieces in this technology and in this process? 

SH: 

Yeah, sure. Well there’s firstly what we call the ecosystem and then I can talk about the steps in how an ad works like the life of an ad. 

Really the ecosystem, it is three primary constituents. There’s one group of people, I’ll call them advertisers. Really they’re businesses and businesses what they want to do is they want to find customers. So whether that’s your local restaurant that wants to find customers or a travel site that wants to find customers, there’s a bunch of customers. 

SH: 

Roughly they’re businesses, and what they want to do is buy ads. The second constituent is obviously the user. This is the user that the business wants to attract. This is the user that’s using the Internet every day. Surfing websites using apps, etc.  

And then the third constituent we don’t talk a lot about is what we call the publisher, and the publisher is typically you can think of it as small as a blogger on a website all the way up to the New York Times. A very, very big publishing house, an in apps land. It’s people that make games, but essentially, people who create content, and typically because that content is freely accessed the way that the content creators are compensated is often through putting ads on their content in order to drive some revenue for, you know, the services that they’re offering, which is free content free access to services and apps and games and so forth. So really, there’s a symbiotic relationship between advertisers, users and publishers. The publishers get paid for their space roughly. 

And advertisers are essentially trying to attract users through advertising. So it’s a very well balanced and elegantly designed ecosystem and within that an advertiser, let’s just say I am a developer of new yoga app and I want to make sure that people who love yoga, who might subscribe to my yoga service in the app, have access and know about this app because finding a service and finding your users is actually really hard. You want to attract the right users to your app, not the ones who might not be interested, but you also need to evangelize your app and get it out to people. So you might go to one of the tools that we built something called App Install Campaigns and you might say: I have a yoga app, I want to find users, I’m willing to pay X dollars per install. And in our case we’ve used machine learning to essentially automate all of the processes of what happens after that fact. So there’s a lot of parts of ads that involves, like where do I want the ads to show? What do I want that to look like and say how much I should I pay for this ad if I’m trying to drive a $1 install and only 1% of the users will click, another, you know small percentage of those will actually install. How much should I pay and that involves a lot of machine learning and prediction models and so forth. 

And also prediction models to determine, for example, who I should not to show the ad to. There are many users in the world who probably are not going to be in the market for a yoga app and it would be wasting the advertisers dollars in showing those users the ad. So process of buying that ad and then figuring out all the pieces and so forth is done through the tools like ours that we provide to advertisers. And that’s what my team builds everyday is how to make those ads really affective. Very beautiful. Very appropriate, very relevant to users, and also making sure that we’re hitting the business objectives that the advertisers are telling us. 

ES: 

And you’ve been this space at Google since 2008, which is about 12 years. What do you find interesting about working in advertising? 

SH: 

Yeah, I think the most interesting thing is that you know digital advertising, which is roughly advertising that lives on the Internet. I mean, prior to the Internet, you know I’m old enough to remember that you know advertising was, you know 32 second commercials in between breaks on TV that you were watching some show for 30 minutes. But digital advertising arose as part of the rise of Google and search and models for finding things on the Internet. That might be things that you need. And what’s fascinating is that it is definitely a very dynamic space. So, for example, the product that I talked about, which is app campaigns, we took a very different approach to buying advertising which was basically taking what was prior to that, a very manual process where the advertiser has to essentially manually input what ads they want, where they want the ads to show how much they want to pay for each of the ads, which is very manual, very grunt work involved. Hiring lots of people to push lots of buttons and look at lots of dashboards all the time. And interpret data and then you know two knobs to moving that to a machine learning model and of course machine learning is a huge area of technology. Applying it to advertising and building one of the largest revenue driving implementations of machine learning is really interesting. That’s been one technology change that you know when we’ve applied it to advertising, which is a perfect problem to apply it to because it’s very data driven. It’s very well modeled in terms of the math in what the outcomes you want, but just building that at a massive scale has been extremely interesting, and seeing the results be, demonstrably better has been very gratifying. 

SoI would say that digital advertising, even though it feels like it’s been around forever, it hasn’t. And the dynamic part of the business lends itself to technical innovation like the machine learning I talked about, but also Business Innovation. I think there’s new models to be introduced in digital advertising that my team continues to innovate on. 

That’s what keeps it interesting. It’s not a solved problem. There’s lots of areas that can be improved and improved in non incremental ways. It improved in much more dynamic and larger steps like the machine learning one that I talked about. 

ES: 

And Google’s main source of revenue is advertising. I was going to ask you about characteristics of how Google approaches it. I’m getting a really good idea from you. There’s machine learning involved, a lot of automation. I was looking at how to set up app ad campaigns and I did see what you’re describing where it just says you don’t need to design the individual ad, just provide resources, provided text and behind the scenes we’re seeing through this automation, the tools are able to experiment and provide different formats of vets and then measure what works. I’m getting the sense of an A/B testing approach and all this is done on behalf of the person that wants to publish the ad, right? 

SH: 

Yeah, exactly. As you can imagine there’s many layers of machine learning algorithms in the background, really dynamically trying to figure out how to. 

Place ads, and yeah, we’re very proud of this product. I think it’s just a demonstration of how technology can be a new technology that is now becoming. You know, even more advanced can be applied to. I’ll call it a space that had one model of solution which is more manually tuned advertising and really completely rethinking it and reshaping it. 

ES: 

When we develop software, we think a lot about our end user, but also we think about malicious users and how the tools and technologies that we build can be misused. In digital advertising one of the scenarios that I saw was in the form of invalid activity and I did see that there are also tools that Google provides to stop invalid activity. Can you give some context around this scenario? And why it’s important to look at it in advertising? 

SH: 

Yeah, sure, I think that invalid activity or a class of spam and fraud and abuse of the systems is an issue that certainly, digital advertising has as well as many other technical systems. I think in the early days of Google, for example, we needed to make sure that people weren’t spamming our search rankings through different techniques to make their websites show up higher, for example. 

I think in digital advertising because there is a monetary trade of value because the ads themselves have value and the clicks have value and the conversions the purchases have value, there’s a business that’s willing to value those with a certain dollar amount, there’s always going to be people who then figure out how to trick or provide implementations of the systems that then are more abusive. So you brought up invalid traffic. That’s a technique to pretend like there’s real traffic where there isn’t. Maybe through bots or other things like that, which then causes a payment to be issued incorrectly to non real humans. So those are just examples. 

I think that it’s an area that we actually have a lot of systems. Internally we have a whole team of people who focuses on this topic and make sure that when we do buy ads on behalf of advertisers that we’re buying good traffic. That of course relies on lots of again machine learning. Other techniques really smart engineers who are constantly trying to measure where the flow of bad traffic might come from, and what techniques are being used to spoof and then to detect and not for example, render our ads on that kind of inventory. 

ES: 

Exactly, and like you’re mentioning, this is an on going process, kind of analogous to spam email, where attackers or users find new ways to create different kind of spam email and then we constantly, behind the scenes, have systems learning about this new ways and sort of tackle them and improve these tools right? 

SH: 

Exactly, yeah, it’s never fully done because there’s always more to do, and there’s always more creative techniques that are being used by the spammers. 

ES: 

Your current role is VP of Display, Video and App Advertising at Google. What are some important qualities for this role? 

SH: 

My role that I took on over a year ago now, our mission is to keep the Internet open and free while balancing the needs of users, publishers and advertisers. And we take that really seriously to heart. And so I think just as any leader you have to have a mission, number one and number 2 really believe in it and make sure that all of your goals and your team are aligned behind that. 

Specific to Display Video and Apps Advertising, it’s a really broad set of customers. So our display customers are very different than our apps, customers and so forth. So thinking about what kind of products and services are needed to, for example, on the app side, helping app developers find users through app installs. There’s a lot of commonalities to, for example, a display advertiser who’s trying to find a user to convert on their website. There are commonalities, and so finding efficiencies in the team to align behind where they are common goals between the different customer sets. But I think the role itself requires a lot of understanding of the different customers of the technologies. The pivots that we’re making on the technology side between the different products and really embracing that our job is to keep that content free and accessible. Which means that you know on the different surface is we have to be really creative about applying our tools, so I’ll give an example. We also service third party video publishers. So a company like Disney for example, might use this for ad surveying into they’re video content and that requires very different technology, very different business models. But ultimately our goal is to make that content ads enabled so that it can be freely made available to consumers. 

ES: 

Vice president at a big tech companies of very high leadership role. Was leadership something that you aspired to throughout your career or did this come along as you were in your trajectory? 

SH: 

I am a first born child of three, so I don’t know if there’s a birth order thing, but I feel that like most of my career, why I went into computer science was because like I said, I just thought it was very creative that you could create services and things and features and games and there was just so much opportunity.  

And so throughout my career most of it has been like what’s the next amazing tool or feature or product that I can build like what problem can I solve and how can I use technology to solve it and how can I use technology specially in a novel way to solve it? Which I think is sort of what I would put my stamp on and then the fun is, you know, finding a lot of great people, smart people to build things together. That team and that sense of shared ownership and the sense of building a common goal towards a common goal with a common kind of culture and operating model. And like kind of that spiritual alignment between people. Hey, where this is the goal, we’re going to go solve this. We’re going to launch this. We’re going to take this mountain is a really like powerful feeling. I think overtime as I grew my career at Google, building lots of great products with lots of great teams, I now see my role slightly different, right? I’m not the one building the product personally, I’m not writing the you know specs, or the PRDS, but what I’m doing is sort of a couple of layers removed from that, and so I like to say that my product is the team now, which is like I’m still designing a product and implementing a product. It’s just that the output of it is the structure and organization and goals and alignment, but like all the teams under me have and in my creating an environment where they’re successful where their goals are clear where they’re not stepping on each others toes where they’re challenging themselves to innovate and be bold and take risk. And by setting the culture right so that you know the right values and performance are rewarded. 

If I do my job right, then we should see lots of very successful technology product features, solutions being built within my organization, and so I think of my drop still as building. But building the team is super important and I think people call that leadership. I call it like my product is my team because that to me is the best way of kind of taking what I used to do, which is being passionate about building products and solutions. And now like as a leader, being directly removed from that but still driving. So having a ton of personal satisfaction when I can see that my team is achieving milestones and getting great product out to market. I think as a leader you know I’m here to serve them. I’m here to make the environment work for them so that they can do what I did in my career, which I found to be very satisfying personally. 

ES: 

That’s an interesting way of putting it, how your focus is on getting great smart people to work together, building great environment, making sure people know what they’re doing, why they’re doing rewarding performance accordingly.  

We are recording this episode as a lot of people are still working from home because of the COVID-19 pandemic. What have been some changes or thoughts that you’ve had throughout this time in how to keep nurturing that culture, keep managing the teams. 

SH: 

The pandemic has been a huge reset for a lot of thing. Just even mentally, personally, like what are the values and what do you take as a priority in your life and how do we think about the priorities at work like what is the bare essentials, the minimum but still, what is the core of what we need to do? 

I think it’s brought out a lot of humanity as well  in terms of just the empathy that we have to show to each other. As you know, co-workers with different in situations at home. it’s brought us in a way closer together to our remote colleagues, who typically would in the past not be in the room. It’s made us be way more laser focused on prioritization. What needs to get done, what doesn’t have to get done, what can be dropped and more explicit in our communications with each other? But that’s you sort of internal facing in terms of operations in the people and how we’re running the teams.  

I think that the pandemic will change the world, you know yet to be clear exactly what behavioral changes will stick versus not stick. I’ll give an example, during the first month or two of everybody working from home sheltering in place, we saw a huge rise in use of mobile games, which of course, you know was translated in parts of our business. But the change in behavior was very noticeable. Very different than prior patterns, and increasingly we’re seeing more usage of to get delivery or local services brought to the home. I think that you know when everything is over and we go back to work. It’s going to be a fascinating thing in terms of what has changed and what will stick in terms of the behaviors afterwards. And I think that it’s times like this where again technology is I already see a bunch of startups being pitched with a COVID lens. I’ll call it on what you know what life might look like after and I get really excited about the creativity of that and like what opportunities are being created right at this moment that you know technology can go help solve for. You know, once you know this phase is over. 

ES: 

Yes, that’s interesting is about what can become a habit because I know if you do certain things after a certain period of time, it might be a habit. Maybe you stick with, for example, doing virtual workouts. Maybe you drop getting the groceries through your phone. So yes, I definitely agree that we will see some changes in this.  

Well Sissy. Thank you for coming on the show it’s been great chatting with you. 

SH: 

Thanks so much for having me and especially during this really interesting time and I really enjoyed our discussion. So thank you!