YouSaid · the spoken record

Peter Reinhardt

lines on the record
66
first
2021-05-20
most recent
2021-05-20
sittings or episodes
1
sources
podcast

Every line below is reproduced as it was said and linked to the record it came from. Nothing here is summarised or generated. Directory · Search · Corrections

  1. One excited, but two like self confidence or like self worth or like judgment was just like, I've been wrong three times now. You know, I had two ideas. They were wrong that I thought were good and the world told me they were wrong. And then this one that I thought was right, the world told me the opposite. I was like so baffled. It took me like a month and a half to recover from that as we started getting more traction with the library. But yeah, the real learning for me was like the world could not give a shit how you think it should operate. It really has problems that it wants solved if you go and humbly solve those problems it rewards you. And if you try to tell it how it should be done, like just get your butt kicked.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  2. Just to take data from our website and send it through to these analytics tools and so on downstream. It's called Analytics.js. He was like, you know, I think that little analytics data abstraction could be a really big business. And I was like, that is the worst idea I've ever heard. 300 or 500 lines of JavaScript. It's already open source. I don't understand.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  3. Trying to build an analytics tool to compete with Mixed Panel or Google Analytics or any other sort of players in that space. That was also a disaster because again, we had this mentality that like we knew it was right and we knew how it should work and we're going to go build that maybe later in life. You've only seen enough that you can have an opinion like that, but certainly for some college kids, like not a good strategy. And I think about a year and a half later we were basically realizing that our analytics tool was also failing. So we're a year and a half in, basically failed at everything. We've got 100K less than the bank of the 600K that we've raised. We realized we have about like four to six months of runway left. And that was pretty intense. We realized we got like one more shot on goal. And so we had a big discussion over a couple days and it ended with a huge fight where my co-founder Ian had this idea that we had built this open source library a bit over a year before that for our own use, which was

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  4. Yeah, we were very excited about the classroom electric tool because we thought that it was the way that the world should work. You know, we were like, hey, you should be able to give feedback on a lecture. And a professor should go and improve their lecture. And people should care about how good these things are. And this should be a great way to go get all of this learning and maybe extract it and build it into like some kind of massive online research or learning opportunity. Yeah, the world just doesn't really care how you think it should work. It just doesn't. The world has its own problems. And we actually made this mistake again after class metric failed, which is we were like, hey, we should have been able to look in our analytics and see that the students were distracted. And so we then spent a year.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  5. Of them said, well, all of them, except a couple, said we invested for the team. So go find something else. And then there were two that we paid back.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  6. Except focusing on the lecture. By having a thing where the professors went and said, like, everyone, please get out of your laptops and open up, it's called class mentrip, open up class metric, and we'll use it as a way to engage better during the lecture, total disaster. I remember standing in the back of this one classroom at Boston University. I think it was an anthropology class, and we started counting screens, just standing in the back. Like, what are people doing? How many of them are actually using our tool? I think at the beginning of class, 60% of students were not using Class Metric at any given time. And by the end, it was about 80% were not using class metric. So basically, yeah, the most horrible thing you could ever put into the classroom pretty quickly realized that that was not going to work. But we'd raised 600K at a demo day on the back of kind of professors trying it out in the fall. And so we called back all the investors that we had raised money from a few weeks prior and said, hey, we just tried this in the classroom. It was a disaster. We should do something else. What do you want us to do with the money?

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  7. In 2011, I was studying aerospace engineering at MIT. My two co founders, Ilya and Calvin, were studying computer science at MIT and our fourth co-founder is studying design at Rhode Island School Design. And naturally, we were really excited about things to help students. The idea that we had and that we applied to Y Combinator with and that we got into the program with in 2011 was a classroom lecture tool where students could push a button to say I'm confused. And the professor would see this graph over time with how confused their students were. And we had talked to some professors who seemed to be excited about it, including Professor Morris at MIT, who's also on the YC committee. And we built a little product where people could do this thing. It turns out that this is an incredibly distracting thing to put into the classroom. Basically, as soon as you put this in front of a student, they go straight to Facebook, they go straight to Twitter, they go like everywhere.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  8. Of just like, yeah, sure, you can do it, it'll be okay, it won't ever be done high quality unless you're sort of amortizing all of this R&D over thousands of customers. And it's a huge maintenance burden. So all of these integrations change their APIs. They change how they want to accept. They change what features they have. They watch new things. They need new types of data. So that is just a huge workload to keep up with as well.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  9. Depth within the quality that one of these integrations needs to have. And then we've got hundreds of them. So if you think about as a company like, oh, I'll just go build this myself, that is a giant tarpet of time that you can waste going and rebuilding these integrations that are already super high quality. So that's one. The other is, turns out data deliverability is no joke. So, okay, great. You got the data point in. Sure, you send it off to the downstream provider. But what if that provider has downtime? This happens all the time. It's not like a what-if scenario. It's like we've got 300 partners. I guarantee at least one of them is down right now or having significant deliverability issues. So how then do you queue that data and then redeliver it at a time when the API is up? We have a cool status page at status.guent.com where you can actually see the deliverability to all of those partners in real time. And there's like substantial losses along the way that we patch over by redelivering the data when they're up or when they're ready to accept it. So there's probably 100 things in this realm.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  10. It turns out that there is devils underneath in the details, specifically in a few different places. One is how do you translate those two API calls, simple abstraction? How do you translate that then to how the data should show up in Google Analytics? We have been building our Google Analytics integration and refining it for almost a decade now. And it shocks me that we still find aspects of that integration that still need to be better. I guarantee you that when you have 20,000 customers using an integration, they're still finding things that can be improved about it

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  11. So on. But in reality, you actually only need basically two API methods. One is the identify call, which is who is the customer, and the other is the track call, which is an observation that user X did Y at time z. That's it. That is entire API. We've since added a little bit to that, but that simplification down to this abstraction of these two API methods, I think has allowed it to be a single pipeline and what allowed a lot of engineers to look at it and say, like, well, that is an elegant solution to this massive spaghetti code that I've been writing.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  12. I think it was getting the abstraction layer of the API and credit for this goes to my co-founder, Ian. There's a really actually awesome paper by the author of the R programming language. It's called Tidy Data is the name of the paper. And he basically explains what it is that makes data in this concept tidy. Basically what it boils down to is a long table of observations of a thing at a time. And that has a bunch of nice properties to it. And you can aggregate all kinds of things on top of that, but you really want to start with data that is things like user A did X at time Y. And if you start with that, you can reassemble the current state of the user. You can reassemble all these other things. We discovered that same thing by accident, I think, with our API, which is very, very simple. So the API for collecting all this data, you can say like, well, there's so many different tools. You're going to need such a complex API to support all this different explosion or advertising tools and email marketing tools and analytics and data warehouses and so on.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  13. And build beautiful visualizations of how people are interacting with my site, what products they're using, all these things. And oops, now I need that same data in a data warehouse. Over time, the rat's nest of all these data pipelines multiplies as companies become more complex and want to make use of data in more and more places.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  14. Small company that's just getting started, often the first thing they'll do when they launch their website is look at Google Analytics on the page. And so that's copy pasting a snippet, adding little pieces of tracking code around the website, and they start getting a sense of like how many people are coming and viewing the page, how many people are clicking through my e-commerce flow. And then they start to notice that, hey, not everyone is actually going all the way through my checkout flow. So I'd love to be able to send them an email. Okay, well, now I need to get that same data about people moving around the web page into my email marketing tool so that I can send emails to the people who started but didn't complete the checkout. So I actually also want to send blast emails once a month to your customers announcing my new product launch. So, okay, now I need a different email marketing tool. It'll allow me to send these batches. Oops, I need to get that same email data and tracking data into the other email marketing tool. And then as the company gets a little bigger, they're like, now I actually, I really want to be an analyst that can sit down and run arbitrary SQL queries.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  15. Before us, and frankly, still most of the market today goes and builds this in house. So they've got engineers who get assigned to requests from the marketing team to go integrate the newest ad campaign pixel into the website, like bespoke, use the API directly alongside the other 20 or 30 marketing pixels that they've got going on, pull data out of this Stripe API or out of this Salesforce API and load it into our data warehouse. lots and lots of bespoke engineering going on to build each one of these pipes independently And what we stumbled on is that there was a relatively clean API to be built that would act as one giant pipe for all that stuff.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source

  16. Yeah, so we help companies collect and manage all of their first party customer data. So all the interactions that they have with their customers on their mobile applications, their websites, their help desks, their payment systems, all those things, all that data they float into segment where we assemble this holistic record of their customer interactions. And then we're like the pipes. We sort of move that data out to all the different marketing sales support tools where they actually then make use of that data to interact better with their customers.

    2021-05-20 · Invest Like the Best · Peter Reinhardt - Learning How to Sell – [Founder’s Field Guide, EP. 34] · IDENTIFIED FROM THE TRANSCRIPT · source