Jeff Weinstein: recommendation

Transcript context

…So there's some really cool lessons here of just how to pick a good metric. Just to kind of maybe summarize what I'm hearing is, you kind of worked backwards from essentially NPS, like, "Why aren't people recommending Atlas?" You found, "Okay, well people that are complaining and having issues with support and running into problems, most likely are going to be detractors, not going to want to recommend Atlas, so let's have those really ambitious goals. Eventually nobody has a support ticket/let's just track how many people have zero issues." And then you identify, "Okay, what's driving a lot of these support tickets? It looks like this risk timeline till they get to this certain milestone, let's make that our goal for the next quarter or whatever, and let's focus there and then make an impact." And then I imagine move on to other levers within this bucket of zero context. You have to pick the right metric for your audience in... I wouldn't wouldn't fully export that metric to everyone. That was not a terrible one to export, but in the founders choosing where to get started mindset, again, this isn't just deeply spending time listening to your customers. They all ask their friends, "Hey, how did you start your company?" They want to talk to an older sibling of sorts about how it went. And so, we decided that our go-to-market strategy would be to delight our current customers such that they would tell their friends. And other businesses, that's always somewhat useful, but you can also reach people with billboards and Google ads and other types of upsells. That's very difficult in the moment that someone is starting a company that's sort of a National Geographic, can you get the picture of the bird at exactly the right time? How are you even supposed to go find people who are about to start companies? Thankfully, they have this practice of just asking their friends. And so, if their friends loved it, they're going to just recommend it. And that has the metric and the tactic and the go-to-market all lined up, rather than sometimes in cases you might hear someone say, "Well, let's make the product quality better." Well, we all make the product quality better, but why? Why is this actually going to move what our customers want, and the business board? And when you can line them all up, it can be quite beautiful, especially when you can see it month-over-month for a long period of time. One other metric I think that I would export to anybody, is if you are unsure what to measure, we have this, I don't know if we stole it from somebody else or if we came up with internally, whatever, is just users having a bad day, where we will just emit a log line anytime we think that a user bumped into a problem. So maybe they hit a 404, or maybe their payout was one day after the ETA, or they had more than 10 payment declines, you can kind of brainstorm again or just listen to customers, what would cause you to personally have a bad day, and then just emit an event when that occurs. And then you could just make a stacked bar chart of all the bad day reasons and the frequency by which they're happening, and it's eye-opening to see those frequencies. And it's kind of a metric I hadn't thought about until Stripe's scale, in which you just don't know what's happening until you emit the log line and count it. The frequencies could be... ... emit the log line and count it. The frequencies could be kind of mind-blowing. And I think for almost any scale business, if you are bored one day and you're not sure what to measure, just make a user having a bad day chart, emit a log line and count it as a bar chart and then anybody else can add their own bar chart on top of it. bored one day and you're not sure what to measure, just make a user having a bad day chart, emit a log line and count it as a bar chart and then anybody else can add their own bar chart on top of it. And so, it's become a way for teams to scale their understanding of users through metrics by just saying, "Hey, look, anytime anybody has an idea about why a customer is having a bad day, just emit a log line, put it on this chart, and then we can choose over time which bad day reasons we want to burn down and hopefully just eradicate them, not just minimize them. But it gave us kind of a background noise counting system for where there are problems and anytime there's an incident or some customer issue, the first thing I think is, "Ooh, I wonder if we have a bad day reason for this." And if we do, I actually feel okay. I'm like, "Oh yeah, it is a bad day for this customer." I wish I didn't have it, but at least we're aware and we can evaluate it against other bad days that we want to burn down. What does sometimes a little bit grind my gears or gives us an opportunity is like when we didn't know about that bad day. And it's a surprise to us too, that is for me is immediate action. It's like, okay, cool. We have to figure out a way to count this bad day. We got to get it on the chart. That way we can make an informed decision about when to invest improving it.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.