Dorothy Tse: We are so thrilled to have, hosting all of you tonight for Quantcast’s first Girl Geek Dinner. My name is Dorothy Tse and I lead product management for our Advertising Solutions here. Hopefully, you had a chance to mingle a little bit and do a little bit more networking. I certainly did. It’s just so fantastic to get to know some of you here and your like-minded interest as women all together in a single room. So it’s really fantastic.
We have a number of exciting talks to share with you today. And up first is Esther Hsu, talking about delivering ads with machine learning. And then we’ll have Malvika Mathur, speaking about what her experience has been like transitioning from a corporation to a startup. And then we’ll have Brittni Gustaf, talking about what it’s like to hack into the customer experience, and then followed by Somer Simpson to talk about how a small team can impact an entire industry. And this is the GDPR piece. And then we have last but not least, Disha Gosalia, speaking about her experiences of how she navigated her career as a shy engineer.We’ll close with some Q&A for all of the speakers and continue with more networking and drinks, if you’re still up for it. All right.
One of the things I’d like to do first is to give you a quick intro on Quantcast. Here at Quantcast we believe that AI will fundamentally change every company, industry, and customer experience. Well, that’s not something to fear because in the 21st century AI is more designed around how to compliment and boost human learning rather than replace it. We’re on a mission, as a company to help brands grow in this AI era. The way that we do this is we help brands and marketers make sense of all the data that’s out there, to understand and make smarter decisions faster. We started in 2006 as a company, as an audience measurement platform, helping online publishers understand their audiences as well as their web traffic — this is called our Quantcast Measure product.
And over the years, our technology and our business has grown and we now measure over a 100 million web destinations across the globe and some of those web destinations include what you see here on the screen. Our technology track things such as site visits, keyword searches, content categories, just to name a few. We process over 30 petabytes of data in a single day. That’s kind of hard to quantify so it’s trying figure out what is a good way to give a visualization of that — that’s essentially, think of 600 million four-drawer file cabinets filled with content. It’s a lot of data, but many people say they have a lot of data as well.
What we’re particularly good at is our ability to make sense of this data. We’ve been working on Q. It is the world’s largest AI audience behavior platform for the open internet. We utilize the measure platform data to help us drive up predictive models as well as AI optimization to score audiences in real time. As a result of that, there are several capabilities that we provide as a suite of solutions for brands and marketers to help them with growing their business.
First, we offer real-time audience insights. And these help to uncover who our marketers’ target customers are as well as what motivates them and how to influence them. We also provide predictive targeting and our predictive targeting allows us to target the right audiences at any point in their user journey, even before they’re in market. And then we also have comprehensive measurement. This is our ability to, in real-time, share audience level and campaign level insights that inform optimization decisions and decisions about how they want to better market to their target customers. This is really exciting stuff for our 700-plus employee base here. We’re a global company. We span over 10 different countries and have 20-plus offices across the globe. We are hiring, so definitely talk to us at any point. First up is to thank you so much for listening and again thank you all for attending today’s Girl Geek Dinner. First up is Esther Hsu. Thank you.
Esther Hsu: Hi, everybody. My name is Esther and I am a staff software engineer on the targeting team here Quantcast. Today, I’m going to talk to you about how we how we use machine learning here at Quantcast specifically in the context of our ad targeting product. This will not be a lesson in machine learning and I’m sorry if you were expecting that.
What do we use machine learning for? We use to reach a specific audience. Each of our clients is running a display advertising campaign and they all have a specific goal in mind. That goal can be anything from raising brand awareness or driving certain actions. Like clicks or purchasing products or anything like that. We use machine learning to help our clients target the right people at the right time in order to accomplish that goal.
For you to understand what it actually means for us to target people, you need to understand what real-time bidding is. Real-time bidding or RTB is a mechanism to deliver ads by auction on an impression by impression basis — impression meaning, like, ad view. With real-time bidding, we can target people at an individual level and we can also buy impressions at individual level. Hopefully this illustration I’m going to go through is going to clarify what that actually means.
Let’s say you’re this little gray stick figure and you go to a website. A lot of things actually happen before you see an ad on website. And the first thing that happens is that a request is fired to an inventory supplier. And what this inventory supplier is is, there are several inventory suppliers out there, some examples are Google, AppNexus, PubMatic, but what happens there is basically a milliseconds-long auction, where the inventory supplier asks all of their bidding partners, how much are you willing to pay to show this person an ad? So what happens is they would send out bid requests to all their bidding partners, requesting a bid. And Quantcast is one of those people doing the bidding.
This a blind auction. Meaning that none of the participants know what the other people are bidding, but everybody sends back their bids and just like any other auction, the highest bid wins. Then the inventory supplier will then choose that ad and that’s the ad that you end up seeing. I know all of you were surprised that this happens because it happens in literally 10 milliseconds so you probably have no idea.
How do we actually use machine learning to do this well? The part that machine learning has in this is that we need machine learning to understand how valuable each one of these bid opportunities are. We use machine learning to basically come up with an optimal price. What this means is that we train machine learning models and we use them in real time to understand how we should price each bid.
What does this model training actually look like and what are these models trained off of? Like I said before, all of our clients have a specific goal in mind with their advertising campaigns. And usually in the most common case, they are trying to drive a certain action, which, a lot of times, was represented by site visit or a certain page. For example, a shoe supplier or shoe company might want to drive shoe purchases, in which case they would choose the thank you page or the shopping carts, which indicates that someone actually bought the shoe or expressed interest in the shoe. Where an insurance company might choose the request a quote page and in any case the client will tag the page and in that way we can then label our data.
Our data, like Dorothy was saying, is coming from our Measure network, which is made up of more than a 100 million sites, and from that, we have this really rich data set of user behavior and really interesting things, that we can actually then label as converters and non-converters or the baseline. And then this is very simplified, but then we run a supervised learning algorithm and we produce a model. And that model will then tell us what does someone who converts actually look like.
We process about 30 petabytes of data a day. A lot of that is because of model training. We built infrastructure to train thousands of models a day, process again lots of petabytes of data. And that way we have up to date models for all of our clients at all times.
Now I’m going to go over an example what a model actually looks like. We train a lot of different models, but this is just like a very old example with a curated set of features. But basically this is an old model that we trained for an online dating service, who was a client of ours. You can see that the green coefficients correspond to features that mean that you’re more likely to be interested in an online dating service and red ones mean you’re less likely to.
You can see that a lot of these are actually very intuitive and make a lot of sense. For example, you’re looking for online dating, you’re probably going to register for online dating. And if you’re looking for baby care, you’re probably too sleep deprived or too busy to care about dating. But something more a little less intuitive like, fantasy sports, does that mean you’re single? I don’t know. And if you like books, maybe you would rather meet people in different ways. I don’t know. But The point is … Thank you.
The point is that even if you’re an expert in your product or your market, machine learning is going to pick up on all these signals, event that no one would normally be able to find. And normally, models have millions of features. This is just like a very curated set.
How do we use this model in real-time bidding? When we get this bid request from the inventory supplier, we have to retrieve the user data that we have for this particular bid request, futurize it, and then basically score it against the model. And from that, we can calculate a basically a number that tells us how likely is it that this person is going to be a converter or someone who’s of interest to our client. And then based on that, we can calculate a bid. And I realize that’s very simplified, but on very, very simplified terms, the bid is calculated from both the score and also several different control signals that we have, which indicate how much budget the client needs to spend and things like that. But very simply, if you’re more valuable, they’ll bid higher. And also this entire process again, happens in less that 10 milliseconds. And we do this for about a million bid requests per second. That’s kind of like an overview of what happens right now.
Wwhat are we working on in general or what are we continuing to work on? Scalability, obviously, always an issue for any engineer. How do we make sure that we can do this for more clients, more data, more complex models or just more bids? And then also ad tech is a very dynamic industry. It’s relatively young. Things that our clients care about from one year might be different the next and because of that we have to adapt quickly. We have to always be updating our models to be optimizing for the things that our clients care about. And even besides that, if our clients care about multiple things, how do we make it so that we can optimize for different goals, balance those against all the constraints that we have as well. And that’s it.
Dorothy Tse: Thank you Esther. That was a great primer on machine learning at Quantcast. So next up we have Malvika Mathur, who’s going to talk to us about her experiences moving from a large company to a smaller one.
Malvika Mathur: Hi guys. Before I start off, I just want to say I’m really nervous so if you don’t get me, that’s not my fault. All right. So I mean working in the tech industry is hard. Like long working hours, keeping up with the latest technology, all that weight that you gain from eating the free food. I mean I did.
Hi, I’m Malvika Mathur, and I’m a senior software engineer here in the Data Platform team at Quantcast. And today, I’m sort of going to talk to you about my journey of transitioning from a big company corporation to a start up. And I’m hoping by the end of this talk, you guys can take away some pointers on what you can do to evaluate the right work environment for you and that can be even within the same company that you’re at right now or somewhere else.
Where was I before this? I joined Quantcast January of 2017, but before that I was working in Microsoft for five years in the India headquarters at Hyderabad. And I joined as a 21-year old, right out of college. And I was like, “Damn it! That’s it. I’m done. I’ve accomplished everything that I need to.” So happy with myself, but the 21-year old in me was really naïve as well. So the first years with Microsoft were really great. They had this program where they give you the opportunity to go between … to go in different teams and different business units and sort of get a feel of what it is to work in these different roles. I got the opportunity to work as a developer. I got the opportunity to work as a tester. I got the opportunity to work as a program manager. But then I decided to work as a developer, continuing that because I really like problem solving.
I joined my new team and I’m there a few months and then my then-manager comes by he was like, “Hey, we have this project. Would you like to join?” I’m like, “Yeah. Sure. Secret project. Why not?” So the task for us was sort of like reinvent the entire calibration process for the entire company. Like, “Okay. How do we do that?” And the other thing that we had to do was we had to deliver this in a really, really short time. That meant for the next three months, we were working nights and weekends and everything. It was super exhausting. But the good thing about that was that is forced me to have like a really steep learning curve.
For the next three months, I was working with great engineers. I was working on the latest cloud technology that Microsoft had to offer. It’s like, “Awesome. This is great.” But then, early 2016, my husband and I decided to relocate to San Francisco. I was like, “Well, okay. Microsoft has offices here. There are teams here. I’m just going to stick and go to one of those teams.” I was in talks with recruiters and figuring what I need to do next and then I decided to talk to one of my mentors and he asked me something really important. Something I never thought I’d ask myself. He asked me, “Why do you want to stay?” I was like, “Why is that even a question? I mean like it’s my dream company. The pay is great. All my friends are here. I like the work. Why would I want to move?” But then he asked me again, “Why do you want to stay?” And I thought about it. It turns out the answers for both these questions are not the same. I thought about what I’ve done so far in Microsoft. I thought about if I move to a team here, what would it mean for me? And I realize that it’s going to sort of slow down my growth trajectory, and it’s something that’s really important to me. I mean it’s great to be learning new technologies, but I realize that as a developer, that’s not all I wanted to do. I don’t want to just go in and write code. I want to do something more. Contribute more in the work that I do. Suddenly, life out of Microsoft sort of became an option. Since I was moving to the Bay Area, working at a start up was suddenly on my shortlist.
I started looking for jobs. And looking for jobs is hard — it is exhausting. And I realize that subconsciously that was one of the reasons I didn’t want to move out — I was in a stable job, I was comfortable, I have my friends around. I don’t want to move because of that. But in the whole process of not looking for a new job, I ended up ignoring the whole process of what’s right for me and my career at that point. So I started to evaluate that and I started to sort of like give that a lot of focus when I was interviewing in all these companies. They’re asking me questions, but I also made sure that I was asking these guys the right questions as well. Because I wasn’t that girl anymore who joined a big company, who was excited with any project. I wanted to sort of do more things. And I wanted to make sure that wherever I went I got those things.
Whenever I go and talk to these people, I started checking on like, “Hey, what’s your technology stack, am I going to learn something out of that?” Right? What are the sort of projects the team is working on right now? What are the projects they’re going to work on later? What’s a big problem that the team is trying to solve for the company or the industry that they work in? And as I started asking these questions, I realized that I am sort of leaning towards working in a start up environment. I think that’s something that’s really important. Whenever you’re trying to find a place that you want to work at, it’s really important to sort of know what challenges you and what excites you to work there. And that’s how I ended up at Quantcast.
I joined Quantcast January of 2017. The last year and a half has been a rollercoaster. I like roller coasters, but it’s one of those Six Flags Magic Mountain types. Well, initially when I joined, I had this really bad habit of just comparing everything that’s done here with how I used to do it in Microsoft. I do that sometimes still, but I try not to as much. And the more I compare, I realize that there’s a pattern. And the pattern is that in these big companies, particularly like what I was doing in Microsoft, work is more divided. Responsibilities are divided. Teams are more siloed. You know exactly what you’re supposed to do. When I go in as a developer, it would be like, “Hey, these are your things to do today.” And you just do that and walk out and that’s it. But here, I was involved in stuff from ground zero. Like I was there at the conceptualization of ideas and while we’re building the feature or while we’re doing the system. And I’ll see it through. And then in the end, I’ll be responsible for taking care of it when it’s in our production.
Another thing was the technology stack. Microsoft was all .NET, here it’s all open source. I mean the first day I walked in, these guys gave me a MacBook. I had never worked on a MacBook before. My first few weeks here were so frustrating. And then after that, they’re just like, “Hey, we have some systems here that are written in Ruby, Java, Python. You own them now.” Right. It was challenging, scary, challenging, but in a good way. So while I was ramping up and figuring out all these differences, I realize that the biggest takeaway is that it doesn’t matter how a company operates or what technology stack they have. The biggest thing that matters is your appetite for learning and where you can get that in a work environment.
In different stages of our career, we have different needs. And it’s really important to cater to those needs. When I started off, everything inside in me, I didn’t care, it’s big company, awesome. But then as I grew up, I realize that I wanted something more specific. I want to do certain things and I tried to find the right fit for me.
I think one thing all of us should do when you go back today is try to figure out why you are where you are and what would help you make the right career move in the way that you want to go to. And that’s something that could be within the company that you’re at right now or outside. Thank you.
Dorothy Tse: Thank you Malvika. That’s fantastic. I’m sure there’s a few of us in the room who can resonate with a story like that, going from a large company, a small one, doing all the comparisons. Pros and cons. So thank you for that. So next up, we have we have Brittni Gustaf. She’s going to speak to us about hacking the customer experience.
Brittni Gustaf: All right. Hello. So I’m Brittni Gustaf and I’m a senior software engineer here at Quantcast. And I’m on the Measure team. So you guys have been hearing a lot about advertising, but I’m in the other side of the company. So I’ve been here for just over four years now and since I’ve started, we have lot of changes on the Measure team, especially on how we go about creating products and features, how we design them and then how we actually implement them. And it’s improved for a lot since that time. So I’m going to kind of get in to why, like how we’ve made those improvements and, yeah, bring my leader’s knowledge to you guys, I guess.
Okay. Measure has kind of been neglected since I started. We’re not the side of the company that brings in the money. We provide the data that provides … the side that brings them the money, but it’s really hard to quantify features that we’re doing it and for it to actually having an impact on your company or not. Because we are focusing so much on what will make the company grow, Measure kind of got passed to the side. We didn’t have a lot of people looking into what features we should have to continue to improve Measure. What we did instead is we kind of came up with what we thought people would want. We didn’t really ask them very much what they wanted, we just kind of, we’re like, “This would be cool and probably would help. Let’s create it.” As you can probably guess, that didn’t work super well.
We spent a lot of engineering time creating products that no one actually really wanted. And that was pretty disappointing when you spend all this time as an engineer and you’re like, “Wow, I made this.” And it was like, “Wow, no one wants that.” You’re like, “Okay.” The company, soon after I started was like, “All right. We’re doing this wrong. We need to change something up.” We did a reorg. And I remember the day after we announced that we had to reorg really well, we had our full day meeting with our new leader of Measure. His name is Sam. And one of the things that stuck out most to me during that meeting was he started telling a story about his acquaintance that he had previously. Don’t quote this to him by the way, I am doing this from my memory, from a long time ago, but I think I have it pretty well because it was so kind of terrifying to me.
What he told us was like, he had an acquaintance and what this acquaintance did was he went and created, spent the least amount of effort he could to create a prototype and then went out to find customers for his prototype, selling it as a product. And he would be like, “Here. Look at this awesome product we have.” And the customer will be like, “Wow. That’s really cool. It would be really awesome if it had this feature.” And he’d be like, “Okay.” He would spend the bare minimum amount of time implementing that feature into his prototype and then he’d keep going back to clients and being like, “Look at this awesome product we have.” Until he got enough investment into his product to then actually create the product. And I was like … While he’s telling me the story, I’m just like, “What is happening? Like this is super sketchy, are we going to be lying to our clients here and telling them we have products we don’t actually have?” Well, he quickly assured me that that was not what our goal was, but that we should have a client-first sort of approach to things, where we create, we spend minimal effort, create a prototype, show it to the client and get feedback before we waste all of this time on it. And that was this mindset that led us to the Measure Hackathon.
How the Measure Hackathon works and how it’s totally different from other hackathons is that we would actually get all of our … Well, we try to get a diverse and key clients into the actual office and we just brought them in and in the morning, we spend three hours with them, just asking them questions about what they do in their job and how could it be improved. And then trying to come up with ideas for how our software could improve it. So after these three hours working with them, you’ll probably recognize a lot of these from one of our earlier slides because they’re the big ones, right? After this three hours, we actually … they went and got on a bus and went to go do fun clienty stuff and the engineers got stuck in the office for 24 hours to try and put this idea into an actual functioning-ish prototype. That is at least demo-able.
This creates two different experiences. So yeah, the clients come here. They’re like, “Oh, here are my problems. All right. Cool. You guys work on that and we’re going to go up and get literally, wined and dined and party it up until 24 hours later.” Which they’ll come back and then hopefully we have solved all of their problems. On the other hand, you have the engineering experience. Not quite as glamorous, as you can see. You get a really creative with what kind of seating you’re going to sit in. I love this because it’s like, how many different seats can you try, but you need to be comfortable for 24 hours and it takes a lot of work to be comfortable for 24 hours programming. And then you also gain a very unhealthy dependency on caffeine so that you can function throughout the entire time.
All right. I don’t know who I’m kidding. We all love hackathons and we all know it. Luckily for us, the company literally butters us up and they give us tons of stuff while we’re doing our hackathon. They’re like, “We love you guys for working. Here, have all your favorite things.” And I feel like they literally catered this food to me. Like sushi, they literally bring in sushi chefs and they make sushi for us. They give us acai bowls, which are my favorite thing in the world. Pizza, it’s like amazing. And they also give us tons of other stuff, like swag. I was going to wear my sweatshirt because it’s really nice, but it’s way too hot. And we get massages.
One of the best things that we get is that you really get to know people who work in your organization that you don’t really work with. It’s a lot of bonding when you’re trying to solve these problems really quickly and you’re all working in the same code base for 24 hours on the same thing.
We also like to take breaks, keep the brain lubricated. We’ve gone midnight drinking, which is always a lot of fun. So the client wins, they get wined and dined. The engineers win, the company’s trying to butter us up a ton, and the company wins, as well, because we’ve had a lot of really successful features come out of this.
Here we have something that came out of our first Measure hackathon, actually. Oh, I forgot to tell you how you win. So the clients, they get a hypothetical amount of money that they can spend on each prototype, so each project, and whoever gets the most money, we actually try and turn into a functioning feature on the website. This is our first one. And it’s one of the most successful products now at Quantcast. People really like it. Yeah. And then if you guys are interested in learning more about how you should prototype things, be customer first, you can check out our blog post. We have the last two years up there and it has some cool videos that give you a full feel of the entire thing. Not 24 hours on, just five minutes, but yeah. Thank you.
Dorothy Tse: Thank you so much, Brittni. I can attest to the comparison of the hackathons here at Quantcast to the hackathons at Facebook. And I’d much prefer the hackathons at Quantcast. So up next, we’ve got Somer Simpson to talk to us about how a small team can impact the entire industry.
Somer Simpson: Thank you. You guys still doing good?
Audience Member: Yeah.
Somer Simpson: All right. Awesome. So we have a pretty significant success at Quantcast. We’ve been on a lot of industry news over the past, it’s only been two months since May. And that’s like the marketing story. That’s like the cleaned up version. What I wanted to do is kind of give you guys the story behind the story and that’s really the interesting part. And also because I’m like super proud of my team because they seriously seriously kicked ass with a really really complicated problem. We’re going to talk about GDPR. Raise your hand if you’ve heard GDPR. Excellent. That makes my job easier. Cool. And I hope you really paid attention, as well, to the first presentations because it gave you a really nice clean overview of what the ad tech sort of industry looks like. So you kind of know the area that we’re going to have to play with.
So what is GDPR? I’ll give you the short version. So it is the General Data Protection Regulation. I have other words that I plug in for those those letters occasionally, but that works. So anyway, this is a law that was passed by EU regulators a couple of years ago. It went into effect on May 25th of this year, which was an incredibly fun day for me. It’s actually an addition to a previous law called the ePrivacy Directive that was passed a number of years ago. And basically, all together what it does is it says that companies who access users’ devices and they set cookies and they collect data on individuals and they process that data, they have to have consent from users to be able to do that.
When ePrivacy Directive was passed, there was no sort of like definition of what does that consent mean. Like what do we actually have to do? So it’s open to interpretation. All across Europe you see all these banners that pop up on everybody’s sites that say things like, “If you continue navigating the site … “ And often it’s like really tiny, down in the corner. “If you continue navigating the site, that means you consent to us using your data for any reason that we want and we’re not going to tell you why or how. And by the way, you’re first-born is ours too.”
What GDPR did, when they realized what was happening, it actually says, “Okay guys. All right. So first, here’s some rules around what consent means. It has to be unambiguous. You actually have to have somebody like click and take an action that says yes or no, I can send.” Everybody in ad tech, just about anybody on the web sets cookies. Everybody does some level of tracking. You go to a website now, everybody’s got plugin, browser plugins that show you just how many cookies are being set. We’re being tracked a lot.
If you can imagine, if every single one of those companies individually had to ask a user if they can consented or not, now when you go to a website, you’ve got 50 pop-ups happening in front of you, asking you for your individual consent. That wasn’t going to work.
We knew that disruption was going to be inevitable. But this is tied to revenue. Disruption is not an option, but neither is business as usual. We have to respect consumer privacy. We’re all consumers. We value our data, we value our privacy, and it’s important that the companies we work for and the companies that you work with do that as well. We set out on a project to deal with how we were going to deal with GDPR because we had significant business in the EU and it was important to us, being a privacy-first company to begin with, but to also address this and stay up to date with the clients.
We have been working with IAB Europe, Interactive Advertising Bureau, IAB Europe. They have had this working group going for a number of years. That started out mostly with the lawyers talking and trying to debate and understand and figure out what their thoughts were on it. And we kept getting closer and closer down to the wire of May 25th. And then last minute, they pull the engineers in and they’re like, “Hey guys, we need a technical solution for this. You’ve got three months. Go.” Yeah. That was mildly entertaining. We knew that we needed a solution, if that’s the solution to every problem.
What we did was we, at Quantcast made a bet on an industry solution. This was the only way that we were going to be able to prevent major fragmentation in the marketplace and still, as an ecosystem, be able to work together and at the same time, not just be compliant with the law, but actually really respect consumer privacy and listen to what their preferences were and actually honored those signals. A number of companies work together. A lot of competitors. Not only were we trying to solve a common problem that we all had, but everybody was coming to the table with their own agenda. It was a lot like herding cats. When I got pulled into these conversations, I’m a little blunt. I have a little problem with patience sometimes.
I went through two of these calls, actually three. They were happening weekly. Same set of people and we just talked about the same thing every single week over and over again and never made any progress. We had four potential solutions that had been proposed and we were basically debating like the most … The most ridiculous minutiae of each one. And trying to figure out which one we were going to do. Everybody got impatient. We’re like, “Okay. Fine. This Friday we’re going to put it to a vote.” And I’m like, “No. None of these work. They’re all awful for some reason.”
I went back to my team — and, which Brittni was on our team — and pulled everybody together and was like, “Okay. Here’s the situation.” And I explained everything, pulled our chief privacy officer in so she can answer the legal questions because I don’t have a legal degree and we basically had three days to come up with a better solution than what had already been proposed. And then bring that back to the group and hopefully they buy the idea and then we go from there.
This is the team. We had chief privacy officer. We had one incredibly busy designer because they always are. We had an engineering lead, who also had four other teams that he was having to manage at the time and then we had four engineers, who also had other work that they were responsible for and none of them had a legal degree, but had to still be able to understand and operate on that level. We had … Well, the first day was me talking so really three days. But four days to get … Until a group vote was going to happen on this proposed solution. We had one moment of inspiration where one of our chief engineers who we talked to about the problem was out jogging one day and just had this moment of inspiration and came back in and he’s like, “I have an idea.” We pulled everybody together. We had three days to figure this thing out and then an hour to visit the idea and basically change the future of everyone in this working group.
The way we approach this, and these are the … It’s kind of the things that I think were what really drove our success, other than the fact that we had a incredible team of super, super smart people. We went consumer vote first. The problem with the working group is they were so tied up in their own agendas and figuring out what they wanted. They were forgetting that it’s all about consumer privacy. And that’s what we had to solve first. We did that, realized none of them were doing discovery and talking to consumers or even talking to very many publishers. I very quickly went out, picked up the phone, started calling people, getting input, and and came up with a framework of what was being asked for. We added the context of that so we all got a crash course in the law and what our interpretation of it was, and then we understood the unique sort of positions of the other companies because they’re the stakeholders, you got to convince them. And then we prototyped this thing in just a number of days. Just so that we could have a proof of concept. Because talking about something and the architecture of something is not quite as valuable and strong as actually showing someone that this will work.
That’s what we did. We present it to the team and it was like dead silence and then the leader of the group said, “Do you guys want to vote?” We’re like, “Sure.” And then we all voted and ours won by landslide. I think we had one person who voted against it. The dude from the Daily Mail.
What we created was this industry-wide standard, and which allows all the system to talk to each other in the same language. It’s open source, which is something that we absolutely demanded because we didn’t want to get into arguments over who owned what IP. It was publisher-centric and it was consumer-centric.
The outcome: today, a little over two months since the 25th, we have a new industry. Consent management platforms. This is this whole new thing. We’ve got a 113 that have launched on the on the IUD framework, which is what we ended up calling this thing. We’ve got 400 registered vendors and growing every single day. You are a part of this framework and talking to each other and sharing consent. 19% of the top 10,000 US and UK sites now have an IAB compliant solution in place on their site. 45.3% of the tools, actually they have some sort of CMP in place, 45% of those are IAB compliant and Quantcast has 69% market share of all of those consent solutions in market. That’s it.
Dorothy Tse: Thank you so much, Somer. That was awesome. Somer and her team have such great impact on this organization in ways that I wasn’t aware. I didn’t realize what happened. And the biggest impact for me is that I was able to hire one of my most senior recent hires because of the leadership industry impact that Somer and her team did. So they … he was very aware of the work that we were doing and liked it so much that he joined the company. Thank you for that.
Next, but not least, we have Disha Gosalia, speaking to us about her experiences navigating being a shy engineer. Thanks.
Disha Gosalia: All right. Last talk and given the topic of my talk, I should start from a point of vulnerability or I can’t wait for this to get over so as you guys so we can all get back to our mingling.
I run customer support and operations here at Quantcast. And why do I qualify and why am I here talking about this? Growing up in India, when you’re somebody who’s a straight A student, or academically focused, you’re kind of placed at a pedestal and you always make your parents proud and so it doesn’t matter if you’re a loner out there. I never realized that I was a shy, loner kid.
Imagine my surprise when, after I completed my software engineering degree, Computer Science, and went for my first job as a software engineer, in my first half-yearly performance review, when your boss goes through all the great 10 things you did. But that one area of development that you always think about.
He actually asked me, “So are you an introvert? I never see you walking around the desk of your colleagues or chatting up with them and you actually don’t even talk much in team meetings.” And I’m like, “Hmm. Am I supposed to talk much in team meetings? Well, I’m new. Should I not be listening more?” But that was honestly the first time I realized that my personality didn’t have a part to play in my career.
Fast forward several years. Now, as I parent really sensitive kids who are often called shy and quiet, I grapple with this thought on a daily basis — like how do I raise confident young adults who can accept themselves as what they are but at the same time also has this growth mindset. And so today, I’m going to go through some of my learnings as I’ve navigated my career and hopefully as I share my story, you guys can pick up some tidbits here.
One of my first experiences when I became a new manager, I attended a new manager leadership training. And the instructor actually asked me and actually the class to write down your word cloud. What it meant was what are qualities that you look in a leader that you want yourself emulate. And when I wrote that down, how this helped me is I kind of became sure sure of what I wanted to be, where I want to go. And I stopped actually feeling bad about traits that I saw in other people that I didn’t actually have. And so I think this helped me because the first step for me was to understand what I wanted to be and then everything else became easier. I just had to go get it.
As Gandhi says, “You need to be the change you want to be, but then you need to understand what that change is.” Before I talk about personality inventory, I will share this story. There was a academic incident that was a big learning point for me.
I was in a really big meeting with my colleagues, my boss was there, my boss’s boss was there, and we were discussing this solution, an implementation solution, a complex solution and the person presenting the solution kept going on and on and I didn’t necessarily agree with that idea, but being who I was, I decided not to really call her out in front of everybody and just decided to kind of go one on one later and talk to her about why I thought this was not a great idea. When I did that, she actually accused of being indecisive.
She said, “Why did you agree with me in the first place?” And I was really taken aback. I’m like, “Really? Did I even agree with you?”
It actually gave me a couple of sleepless nights. And at that point what I didn’t realize, which I realized a little later, was that it wasn’t that she was accusing me, it was that my lack of speaking up or lack of objection in the meeting was actually taken as agreement by her, and it was only because we have had different ways of processing information.
Fast forward in the same manager training, they made me take this Myers-Briggs personality test or there’s the Myers-Briggs, the Enneagram kind of same type of personality test and my original thought with these was, these kind of pigeon hole you into specific categories and it’s like, “Do I have to choose between being a compassionate person, like Mother Theresa or being a leader like Martin Luther King. Why can’t I be both?” But being a good student that was, I went with the flow and what I understood was this wasn’t labeling me in a particular bracket, but it was really understanding how I communicated and how can I become a better communicator with my co-workers and team mates and kind of others in my circle of influence. But that’s what it is. That’s basically all this personality inventory is.
Going back to that example, this person, the way she processed was she would talk and think while she’s talking while how I process was like think and then talk. Like I would have these long awkward pauses but she would keep going on and on, and what I realized in actually going through this process was I need to just find a pause and then ask clarifying questions and that’s kind of how to better communicate with her.
Now to contrast that is the growth mindset. I read this really great statement that’s made a big impact on me about this contrast theory. The growth mindset actually tells you that, do you accept yourself the way you are or do you actually try to be more, more than what you are and constantly evolve and constantly grow?
Bear with me for a minute. I want to actually give you guys an example that I read that, again, made a lot of sense to me. And this is about the metal industry and how do they rate the hardness of metals. They rate them from a scale of 1 to 10 from a hardness perspective. A diamond is a 10 and a tin is one. A copper is a three. Tin is the softest and copper is three. Now tin and copper are not found in the same vicinity at all. They’re like found in a completely different vicinity. Somebody decided to take tin and copper and combine them. You would actually think that that would be an average so its hardness would be a two, but no, combining tin and copper gives you bronze, which is a six.
This is what happens, and it’s called the contrast theory. It creates this unique magical combination. And that’s how personality traits are. I mean you could be way over here as an introverted shy person or you could be way over here. Aggressive, type A sales guy. I work with sales guys a lot in this job so I can pull on that a little bit. But if you combine and while you are right here, try to get a little bit of this side, you can be unstoppable. You could be an engineer surrounded by a lot of shy engineers. Try to get more communicator, public speaking skills or even skills to make other people feel special and it will just be going places.
This is … what Sheryl Sandberg says in her book Lean In, always sit at table. Don’t take side seats. And it was really important for somebody like me, who had a very soft voice when I was in meetings and if I wanted to say something, I would think and by the time, sometimes the time’s already gone to speak. But when you’re in that center stage, people can actually see your body language that you want to say and can actually give you a way in. I started showing up in some important meetings where I would know there are a lot of people before time so I could get the right seat.
What this also did was when somebody disagreed with you, they actually had to look in your eyes and do that. I hate conflict. I don’t like that one bit, but when I think of some of the biggest innovative solutions, the breakthroughs I’ve been part of. They’ve usually been through a lot of intense intervention, conflict, and I’ve learned to put myself in those situations. Put your ideas out there, let it be beaten up and you will learn something through it.
Beth Comstock was a leader I truly admire. She’s the ex Vice Chair and CMO at GE, where I was previously before Quantcast, said this:
“Conflict is a primary engine of creativity and innovation.”
And I’ve learned to accept that, however hard that is. Kind of let that in once in awhile. One other principle that I grew up with was, you do your karma and don’t worry about the results. Other way of actually putting that is you can actually outwork anybody else, you can out prepare anybody else and that’s kind of what I try to do. I try to be double prepared and triple prepared when I know it’s kind of my chance to do things that are uncomfortable.
I use to get really flustered when I would be put in a position by someone or in a spot by someone where I have to give quick responses or make decisions quickly. And what I learned … It was actually a mentor of mine who helped me through this and coach me through this is, you know it’s okay to ask for more time.
It’s okay to say that, “I’m going to need 24 hours. I need to sleep through this. I need to think through this,” and there’s no shame in doing that. Don’t let anybody else put you on the spot and make you give answers that you’re not ready to give.
Let’s bring it all together. Find out what you want and just go for that. Always take a seat at the table, not the side seats. Always be prepared, but if you’re not, there’s no shame in asking for more time. Find a Yang to your Yin.
This is something again, I’ve done when I hosted large events or large meetings, find somebody who is … who can compliment your quiet type of personality. Somebody who’s upbeat and funny and loud. It just makes things easier, and I don’t go to a social gathering where there are too many strangers without my husband who was a talker. There’s like no awkward silent moments. But he sometimes forgets when he’s talking that I’m even around. But that’s a different topic.
Lastly, I think if you remember that all human beings are really at core, alike. And we all like to be respected and we want to perform in our jobs and want to be heard and and listened to. So I think when you remember that, I think everything else is just smaller. That’s it. Thank you.
Dorothy Tse: Thank you so much Disha for sharing your story. So right now it’s time for a Q&A from all of our speakers. Any topics you may want to ask any of us. So if you’re interested and have some questions, you can come over to my right side here and ask a question and all of our speakers will come up and answer any questions that you may have. Thank you.
Audience Member: This is kind of just general to everybody, but as a female in tech and in engineering and in product, what do you feel like is your biggest struggle and how do you succeed in this role when we are kind of overwhelmed with males in our community?
Somer Simpson: It’s funny when I first stood up here. I mean I’ve been in tech since 1994 and I’m looking out in the audience and I’m like, “You know what, finally I’m looking at what tech should look like.” But having been in tech that long, I learned a long time ago to just not differentiate. Not even admit or acknowledge that there’s a difference and just be myself and speak my mind and be a part of the conversation. Just don’t take no for an answer.
Disha Gosalia: Yeah. I mean I’ll just add in. I think Somer’s really right, what I’ll add is also you know, I’ve always had good women role models, who helped me like when I got first child. How to navigate that and just kind of go through things. So it’s important to obviously not see yourself as different from a man, but then we are different. So definitely try to find somebody that you can follow and who’s ready to like guide you through some life changes.
Esther Hsu: I will admit that it actually took me awhile to realize what a problem it was for women nowadays, and once I did, it was actually looking pretty discouraging. Like you notice all these differences from you to all the people around you and you kind of automatically see it as a detriment. And I think, for me, what made the biggest difference was just having mentors and people who I really look up to — men or women — who really point out all my strengths, and I’ve realized that all my strengths are the things that made me different. As cliche as that sound. It’s like when I started it’s like I hate hearing that too. But it’s like it’s so true. Everything that makes me different that people might see as feminine qualities are what make me a better engineer and a better communicator and a better leader.
Brittni Gustaf: So the thing that probably held me back to the beginning was like the imposter syndrome. I’m sure you guys have heard of it. And I really struggled with that at the beginning. I still struggle with it sometimes now, but at the beginning it was so bad.
It took me a really long time to realize that there are a lot of people who have very strong opinions and they voice them as fact — but it’s not.
You sit there and I always at the beginning, I was like, “I’m just completely wrong.” Like, “I don’t think this is right at all, but obviously this guy knows what he’s talking about. He’s so sure of himself.” It took me a really long time to realize that if I’m confident on something, you have to actually bring it up. And then a lot of the times, there are other people in the room who will also be like, “Yeah.”
Malvika Mathur: I feel like whatever I want to say has already been said. But I’ve caught myself in situations where I’m the only female in a team of people or in a meeting and I realized that nothing is bigger than logic. If you have solid points and if you know exactly what you’re saying, it doesn’t matter what gender you are or whatever else. It just like you have a good point and a good point always wins. That’s about it. Just note that. Sorry.
Dorothy Tse: I didn’t realize you were all going around. But one thing I will add is that as a female leader, I try to embrace it. I embrace the fact that we’re bringing a different perspective in a very male-dominated industry and that is an asset to a company. The different ideas that come from a woman’s brain and the types of perspectives that are brought sometimes are very unique and different. So I encourage myself and certainly others to think about just embracing that diversity.
Audience Member: Awesome. Thank you ladies.
Somer Simpson: Just one other thing to add. I think part of my sort of struggling in the journey was, I was fighting more to be queer in the work environment. So being a woman in the work environment kind of like took a side stage.
Audience Member: All right. Thank you ladies. You guys are all amazing.
Sheryn: Hi. I’m Sheryn and I’m a co-founder of a startup. We’re also only females. So it’s really great to see you guys up here. I think all of your stories complimented each other and it’s very nice. It’s a novel of stories that’s set in front of us. My question’s more specific in terms of the hackathon. I’m a UX designer, researcher, and when you talked about the hackathon, it seemed very developer focused. So I was wondering if that’s part of the culture here that the designers are also part of the hackathon or is it very engineer focused? Because you keep saying the users first and we’re the ones that were kind of super obsessed. We made our whole lives about the users. So how does that work here?
Brittni Gustaf: Yeah. I should have probably clarified that better. So it’s not just engineers, it’s definitely product managers and UX as well and all of the designers and we actually sometimes, we like get disappointed if you don’t have the designer on your team in a hackathon because having a designer is like a huge asset because things that look nice and work well for the user, tend to win really well even in your simple prototype. So, yeah, they’re a huge portion of it and also a huge portion here at Quantcast at working with the product managers to make sure the design is what customers can understand. And we’ve learned that the hard way because we use to have tests where we would have people run through our stuff and that was just so painful to be like, “Just scroll down. What do you want? It’s just down a little bit.”
And like people are trying to get to certain paid and they’re clicking everything but the button they should be clicking. So yeah, that was the struggle we had and we’ve become a lot better at that by having both product and design create clickable prototypes and then have the user use it and then get feedback and then make improvements. Which has been really awesome and it’s really improved our products so far.
Sheryn: Thank you.
Somer Simpson: I was going to say that worked so well in the hackathons that we’ve actually reorged our groups to have dedicated teams to each product, that’s made up of a product manager, engineers and assigned UX person and a product marketing manager.
Audience Member: Hey there. The question I have is more specific to Somer’s story, but if it makes you think of stories that you want to share because of what I asked, go for it. My question was about your decision to say, “Wow! All these ideas suck. I’m going to come up with my own and present it and hope for the best.” What was going through your mind when you made that move? What other steps did you take to increase the chances that they’d go for it?
Somer Simpson: The options that were on the table, at the surface, all of them were great ideas. But once you scratch, pull off the surface, they all had problems. For instance, one idea was a centralized registry to store people’s permissions, but that would be one company building a massive database that might hold the trillions of records necessary to do it but all that data would be in the hands of one company. That was bad. And then we had one that was like this pure, what they called daisy bit, which was we just pass this information around.
What we ended up doing was we took kind of like the best of the solutions that fit everybody’s needs and were like not quite so controversial and created what we initially called a hybrid solution, but I mean it wasn’t completely my decision. I mean I walked back into the room with the team that’s like, “All these ideas suck.” But it was the team that actually really got together in understanding, in breaking down each solution, took the best out of each and came up with the right thing that ended up working well for everyone.
Somer Simpson: Brittni, you want to give your side of it?
Brittni Gustaf: I think that the other side of things is that it’s really important to bring in other people and get outside perspectives. That was one of the things that disappointed me most about the GDPR implementation is that IAB and all of these people who are meeting for so long, trying to come up with a solution to this. And all it took was pulling in more people with more ideas to be able to get the best one. But we spent so much time not getting there because we weren’t pulling in everybody needed and getting all the different diverse perspectives to be able to come up with the idea that was best.
I feel like we got a later start than we should have because if it had been … If the correct people had been pulled in sooner then we would have not have such a stressful time trying to get this done before the law was in place.
Audience Member: All right. Thank you so much.
Dorothy Tse: Ladies, thank you so much — and some gentlemen too. Thank you so much for attending the Girl Geek Dinner and we just want to emphasize also that at Quantcast, our greatest asset is our employees and there’s a bunch of folks around the room that are wearing Quantcast clothes as well as all of us up here and we would love to talk to you further about Quantcast. Thank you. We’re hiring!