Why Choosing the Wrong Tech Framework Nearly Cost This Product Manager Everything

I mean, I, uh, let me, let me divide the, divide the tech, the aspects of technology choices into, uh, things that I think like, um, that probably are more in the purview of engineering versus product versus like the things that I think product can be responsible for, which is like, you know, sometimes like technology vendors and, and, you know, uh, partnerships and things like that, right? Um, because I think like, OK, if I were to look at, uh, uh, a decision. Um, that, you know, we made a Dollar Shave Club back in the day to move to like a, a, a, like a single page web app, um, which was a, a big thing for an e-commerce store to be able to, to do something like that. I I realize they're all sort of the rage now, but, um, you know, I think that as a product leader, like, you know, together with design, like you're advocating for a certain type of experience, right? Like, you know, you want, in, in our case, we were, we were advocating for this app to feel like, well, we were advocating for the web app to feel like an app. Like we wanted to work well on mobile, so it was responsive. We wanted like loading times to be minimized. We wanted to be able to have transitions and, and, and, and things that are just sort of commonplace today, um. Ultimately, it was an engineering's um purview to be able to say, look, I think we're gonna use this front-end framework, which was the wrong one we used a framework called Ember, which I don't think anyone really uses anymore. Yeah, man, like it was a fucking mess, but whatever, um. Yes, we, we made, they made a bad technology choice, but, you know, the reasons for making it was like the promise was that that we would be able to do all of these things. That was one thing. The second decision that the second thing that drove that decision was it was a new framework and as an, as an engineering team that took pride in building. All of our own things, working on the hottest technologies or the newest technologies with a way to like get other great engineers into the company and so that's, that's sort of recruiting tactic that is,

so here's the thing, I mean, not me not being an engineer like I, I mean, I, we did for the LA area. I mean, I remember like I was talking to our, our head of engineering at the time. I was like, oh man, why, why did we pick this framework? It's is so problematic. All the things you promised. We're not able to do and he's like, well, listen, man, like, you know, if you look at the people that we're able to recruit, like, like go look at the GitHub submissions for LA for JavaScript developers and we have 1, we have 2, we have 3, we have 6, we have number 7, like I guess there's some kind of ranking system in GitHub when you do your submissions and so we, we were able to like recruit some really great people as a result of like sort of that, you know, build versus buy. Uh, you know, culture that they developed. So, so anyway, my, like bringing this back to your question though, like I think that like, you know, the choice of front-end frameworks is really about like that's in engineering's work, you know, that's their decision, um, but ultimately they need to like sort of respect. The experience that we're trying, trying to, to ultimately build, and so I think like there's a, there's conversations and, you know, that, that ultimately help you arrive at that decision, but they need to make that. In contrast, picking a vendor, like I think that, um, I think there's often, uh, you know, if I look at. One vendor that we picked it, uh, we actually picked it both at Figs and Dollar Shave Club. Um, it was, uh, uh, a, a company that helps with like, uh, shipping and tracking. So, like, and, and you, and you, you see them everywhere, like if you buy anything from Nike or you buy anything from like Sephora or any, any, it's like this one company that handles everyone's tracking page effectively, right? And so, you know, we had a problem with Dollar shade and figs where people would, you know, get sent their boxes and the tracking numbers were all, were all off. They weren't, it's not that they were off, it's the data and the delivery dates were off because ultimately a lot of times when the moment you get an email with a tracking number on it, like it hasn't actually arrived,

Yes, so, so this, so this. Company, um, ultimately like takes like like that tracking info and they, they take like the provider that you're like users are sending the data over to them, um, the, they take the shipping provider and they take like the, the origin zip code and the destination zip code and they have so many partners that they can like actually predict like with with some amount of reasonability like. Exactly when this box is going to show up, you know, down

that's, yeah, exactly, and they know based on this zip code to that zip code like it's gonna, it takes typically 3 days and so and then of course like as UPS actually scans things and like starts to provide more, you know, more accurate data, they can change their estimates to get even more accurate. But, but the point is that they do all of this for, you know, the, the, the. Clients, um, and so I think like for, you know, when it came down to the evaluation of that partner, like there was a couple of things that we were, we, you know, we were trying to solve like one, like let's dramatically reduce the number of customer care inbounds and people being like, look, you sent me this tracking info and it has no, it has the tracking number, it has no info on it, right? Like, and, and so we're getting a lot of inbounds from people, from people and you're talking about like we're sending 80,000 boxes out a day, right? That's a lot of, that's potentially a lot of thousands of people. Like, where's my box? There's no data in here, right? So that, so that was thing number 1. Thing number 2 was just around like, you know, when you think about like the tracking experience, like at the time, like USPS, uh, even FedEx, they weren't like mobile responsive and they said that it didn't look good and so, um, so that was another thing. And then the third thing was was just like from a a marketing upsell perspective, like, you know, if I send you to USPS like there's no chance for me to You know, either give you content that I, I think you might find interesting or even upsell you on other things that I think like you might want to know about, right? So this company, you know, was able to provide all of that through like a fairly lightweight integration. And you know, mind you, it was, it was something cheap. I mean it was like $10,000 a year. I think maybe now it's like a lot more because they've sort of like taken over the world in this space, but like, um, at the time, $10,000 to get rid of thousands of inbounds per month, um, or maybe tens of tens of thousands of inbounds per month seemed like a no-brainer. Plus we got the added benefit of having better experience and getting upsells. It's actually like we're making. Money off of this, this, this email that formerly we didn't make any money off of. So, um, so I think like for that like I mean, you know, being able to evaluate the, the choice across those three dimensions was like, you know, sort of made it a no-brainer and so those are the types of things I think like product is probably more attuned to, to balancing.

Well, I mean, I think like, I think it's like. Like any product you release, I mean you have to, you don't just like put it out into the ether and then be like, OK, it's done, like let me, let me, you know, forget about it, right? Like if you launch a new onboarding experience and you AB tested and it wins and you're seeing like crazy results, you want to know that those results are going to hold out over time, right? And so, So you put together uh some amount of instrumentation whether it's through Google Analytics or whether it's through reports in your own, you know, uh, uh, data analytics interfaces in our case was Looker, like you're constantly looking at those, at those reports. You automate those so that, that they, you don't forget about them and they, they show up in your inbox like once every couple of weeks. So, initially you're looking at it every day, right? And then eventually as you start to settle, settle into the patterns you're like once every, you know, once every week, maybe once every 2 weeks, and then, then at some point when you see that there's a nice established pattern, it's like once, once a month, but you don't forget about it. I think like you, you constantly have to like be looking to make sure that the results actually hold up.