Hunter Harris: Hey everybody and welcome to Build and Break Through. I'm your host, Hunter and it's just me again today. I've been asked several times recently and many times over the years to talk about GDPR C CPA compliance. It's something that's come up a lot in my client calls and something that's come up in interviews pretty regularly. but it's something that I haven't about on the podcast very much, and something that I think Could be extremely valuable to some of the founders listening here in part because I feel like compliance is often misunderstood by a lot of people. It's viewed as kind of a a large burden, but at the same time, appropriately complying can be a very significant moat for your company. can unlock sales, it can protect ⁓ business, it can open up a ⁓ new streams. ⁓ there's a lot of advantages to making sure that you're appropriately addressing compliance. So ⁓ GDPR is general protection regulation and C CPA is the California Consumer Privacy Act. ⁓ CCPA is largely based on GDPR, but it essentially states that you as a person own your own PII, PII being personally identifiable information. And understanding PII is essential to understanding these regulations. So what is PII? you have a lot of information about you, right? Talking about me, I have a name, I have a phone number, I have multiple email addresses, I have a phone number, I have credit card number, I have an IP address, I have a social security number. there's a cloud of information that represents my identity. So if we look at each piece of this, a a picture begins to emerge. So If we look at just name, for example, Hunter Harris, ⁓ how many Hunter Harris are that there in the world? quite a few, I will say. a pretty significant number, which means that the name Harris by itself is not very identifiable. I have an address, but how many people live at this particular household? I have a phone number, but if you call the phone number, how many people pick up? So by itself, a name is not especially descriptive, a phone number is not especially descriptive, an address is not especially descriptive, especially if it's a commercial address. However, when you start bringing pieces of data together, you can data together to get a personally identifiable piece of information, which is to say your name plus your phone number plus your address, there is probably only one person that that represents, right? And so that's now personally identifiable. Relatedly, IP address as an example. IP address by itself, probably not especially identifiable. It may represent a dynamic DNS. but when you combine IP address with other pieces of information like a name, there's probably only one Hunter Harris at my IP address, for example. The exception here is your email address. Now I may have multiple email addresses, but each of those email addresses points directly to me. I think there's probably some like a hello at corporate the emails that aren't necessarily personally identifiable, but ⁓ ⁓ a solid entry point into the world of PII. GDPR and regulation as a whole. Why is it important? Well, first of all, what you have to understand about GDPR C CPA is it ripples. is to say, if I a business and I'm noncompliant, then anyone that I do business with is also noncompliant. And if you're small venues and you have small clients, this probably doesn't matter. Relatedly, if you're not doing business in the EU, probably doesn't matter. But if, for example, your business wants to do business with or Microsoft or Google, those are large multinational corporations that do want to do business in the EU, which means that if they do business with you and you're non-compliant, they will become non-compliant. as a founder, the way you should be thinking about compliance is not ⁓ in terms of compliance itself, in terms of the sales that can be unlocked. via the compliance motion. There is a another flip side there, which is risk. if you are non-compliant and you successfully sign enterprise clients, those customers are churn risks. Therefore you should consider compliance in terms of making sure that you can continue to profit from that business arrangement. And also if you choose to do business in the EU and you don't comply and you're above a certain level, there is the of fines. ⁓ ⁓ although generally speaking, the actual risk of fines, even though the number is relatively high, the like risk adjusted value there is much less than the opportunity cost of the sales that you might lose. one advantage to addressing GDPR relatively early is GDPR CCPA compliance kind of forces you to nail things like data warehousing and ⁓ data policies. Doing those Early means that you have a really strong data foundation for future work. can make your engineering team faster, ⁓ can make some of your AI usage a little bit easier to use. So it's definitely something to consider. My general advice is don't worry too much about compliance unless you're going ⁓ after or if customers compliance as a as a sticking point for your deals. So now talking about the project as a whole, first off was the first phase, which was investigation. This was a period of time in which I sat down with my product peer, and we and we really made sure that we understood the data foundations of the company we were at at this time. It was a long, arduous process of ⁓ literally we at every table and every database across a series of microservices. ⁓ looked at the fields there, pulled data back, and tried to establish where PII was and how much PII exposure we had, and whether or not we were actually compliant with the previous redaction and removal requests that we had received through the GDPR process. we did discover that we were totally non-compliant. we had a system that no-opt, which is to say it It was a system that checked the wrong record for the wrong information and then did nothing with it. So fully non-compliant. We also had PII scattered all over, which is to say this was clearly going to be a very large project, not something that we could just trivial trivially knock out because PII was everywhere and how the organization was structured, all the across of the engineering organization would would need to be involved in this in some kind. This meant that we needed get started ASAP to de-risk project and to make sure that we were ahead of any potential slowdowns due to things like deadlines, etc., on all of the of engineering organizations. So ⁓ hopping into this, principal goal was to, from an engineering angle, de-risk this as much as possible, which means build an MVP, prove that it works ASAP. So I went, I over the course of I believe it was about a week, I built a package that allowed us to cover a few core capabilities. First of all, it allowed us to in-code document PII. This is an extremely large advantage because by just documenting the PII, that that it was very explicitly written in the code base, easy for AI to pick up. Additionally, that meant that as things changed, we could very clearly see what was documented and what was not. by documenting PII, we could hook it into systems that would interact with a PII based on that documentation. That meant that the future engineers that would be using this would not have a lot of work to do in terms of getting this set up. All they had to do was go assess their PII. ⁓ And then document it, and the system would take care of it for them, essentially. So by doing this, this was a lot of work for me, ⁓ me to pull ⁓ risk ⁓ delivery in the future into today and resolve it. So once I had this software package allowed us to cover these activities, which is to say documentation, which retrieval, if you will. anonymization and then pulling the data back and storing it for return to the customer. Essentially right to know kind of activities. We then a workflow that would incorporate this. That way we could ⁓ take inside a given system, plug in a piece of PII, test whether anonymization could work, test whether retrieval could work, and then ⁓ continue on. And then we built up one more layer after that in terms of outside of this individual system, have a system for kicking off the PII process, run the PII process, and then have a system for monitoring the PII process at the end. this was especially important because part of compliance is being able to send an email. So that meant that we with our microservices architecture, we had to complete a series of activities one after the other. And then wrap it all in a bow and send an email out out to the requester at the end of the process. so that was a ton of work that really my for a significant period of time, think it was a month or two. we probably could have done it quicker, but ⁓ one of the goals here was to this risk while at the same time not letting any of our existing deadlines slip. ⁓ which was extremely difficult but manageable by being very smart about your head count If you stay really crisp on your priorities and you manage your time well, you can definitely slot in side priorities and still stay on your committed execution I did want to mention this was at the point where we hit a couple of a couple of the curveballs. First of all, we had some fairly critical stakeholder misalignment as part of this. there was one particular stakeholder that was repeatedly trying to kill this project, despite not compromising on ⁓ priorities, and the dollar amount associated with the sales that we were expecting to get. We were we were trying to land a very large enterprise client, and this was contingent on our ability to prove. our GDPR C C PA compliance the other curveball here discovering that our data provider could only synchronous flows and given our microservices we had to have asynchronous flows that is to say they would data from us and we would have to go across a series of systems to pull the data back and give it to them, which would just take too long for a web request. So things would time out. So this was moment where we were like, uh-oh, we've discovered that there's a huge risk factor here. There's additional work that we need to do. And luckily, and I'm emphasis on luck, the way that I had structured this meant that the additional work that we needed to do was just a few hours. But if had this in a very different way, ⁓ If we hadn't had this documentation flow and this workflow-based flow, would have been a significantly larger lift that might have ended up taking us an additional month or two. so after that point, after these curveballs and after getting the MVP out, we kind of realized that we were at a point where we could actually move towards shipping this thing. that is to say, it was already in prod and we could go to the individual engineering teams. So my team as a whole just continued trucking on with their existing priorities once again. on the other hand, continued with my existing priorities, but in addition, went engineering team by engineering team, embedded with them, helped them analyze their PII, because I had already done some amount of analysis myself, and helped them integrate this tool and get hooked in, get their workflows up and jamming, and then Do a QA test team by team by team. So the total time commitment per team for this was I think it was about a week, two weeks total. very low lift because I really invested a lot of time in enablement there at the start. We had a year-long deadline for this, and we shipped this in about three months. the stakeholders were worried that this would take over a year to ship, and we shipped it in three months. That meant that we had the additional revenue from those large contracts nine months ahead of schedule. It meant that we significantly our risk of churn. There were a number of customers that had it turns out, GDPR CCPA and compliance written into their contracts. And we were able to avoid significant churn. Actually, it turns out that the avoided churn was larger than the dollar amount of the signed contracts. So increased revenue, didn't lose revenue. we were to something like two billion dollars worth of potential fines. The you know, the the percentage chance of that actually happening is so infinitesimal that it doesn't really matter. Aside from saying the the amount of potential fines was astronomical. additionally, the tech that we built, essentially module that I had built allowed us to hook into more than just PII. It was configured just for PII, but with additional in-code documentation that allowed us to do things like data as we transferred it over to test environments, which meant that we could spin up new testing systems much rapidly. meant that we could use sanitized data in our test suite. which meant that essentially our risk of bugs and our risk of defects went down significantly. So overall that resulted in additional revenue, ⁓ success for success for marketing, because we had a great splash with landing this particular customer, and additional velocity additional for engineering. not every project works exactly this way. This was an incredibly difficult one. ⁓ due to mean leading hundreds of engineers over a period of time, going off into the wilderness and building something up front is always a bit risky. But also it's frequently to address the risk early in the project than to allow the risk to fester ⁓ and then out that a lot more risk than you realized when you're moving towards shipping. I realize I'm a a bit general and not naming company names, but I'm not entirely sure if they'd be pleased about me talking too much about this. But it was a really cool project. It was one the hardest things that I've ever done. Really continuing ship of your deliverables while taking on a large, risky project with lots of uncertainty and then delivering it significantly ahead of schedule. But once again, this has been photo breakthrough. I'm Hunter Harris. Thank you very much for listening. I hope this is helpful. Please ⁓ share, like, subscribe. We're on the major podcasting platforms. We're on you maybe you you're watching this YouTube, maybe you're listening on Spotify. But if you know somebody that's interested in ⁓ starting companies and what it really to run them, especially B2B SaaS companies, please share this with them. we got a lot of exciting guests coming up, so I hope you're looking forward to for future episodes. and until the one, I hope you have a fantastic day. Peace.