We’re putting a few million extra dollars into a spot and from that we learn. And we’ll see whether we do it next year or not, but it would be predicated on having some important launch that we really want to talk about.
Ken Gawrelski: Thanks so much.
Josh Silverman: Yes. Thank you.
Debra Wasser: Thank, Ken. We — I know we went over, but we did have one more in the queue. We want to try to get to. So operator, can you give us the last one.
Operator: Our final question comes from Mark Kelley with Stifel.
Josh Silverman: Hi, Mark.
Mark Kelley: Hey, how everybody? Thank you very much for squeezing me and I appreciate.
Josh Silverman: Thanks for staying late.
Mark Kelley: Of course. My first question, and I know you touched on this a little bit in the prepared remarks, but can you just walk me through the dynamics of Easter a little bit? Was it not as much of a tailwind because it was like the exact last day of the quarter and maybe some shipments occurred in Q2? I guess maybe some of the dynamics there would be helpful just to start.
Josh Silverman: Okay. So the headline is that it was a bigger headwind to March than we expected and an equal better tailwind to April than we expected. We, obviously, knew when Easter was, we obviously forecasted in, and so it was just a bigger impact than we expected. And from what I can see from others, and I can’t validate this, but little data. It looks like it might have been a bigger impact for us than for other people. And the truth is we have several hypotheses as to why, but no really great data. It candidly wouldn’t matter because it was a headwind to March and a tailwind to April of roughly equal and opposite size. The only reason it matters is it fell right smack between Q1 and Q2. If it was all in March, it wouldn’t have mattered. So we haven’t spent an enormous amount of time trying to dissect something that ultimately was just a timing shift. But that’s what it was.
Rachel Glaser: It’s not just Easter. It’s spring break. And what happens when people go on spring break is they’re spending money on travel instead of spending money on Etsy, and they’re also traveling, so they’re not shopping. And that shift, we think, had a pretty big impact more so than we thought it would on March. And so that’s why — in a way, it’s a tailwind to April because last year when they were traveling, and it was we get them back now because they’re not traveling anymore and they’re focusing on summers. Also, Easter is not a huge holiday for Etsy. It’s not like — I’m told, I’m Jewish, but it’s not — sounds that a huge gifting holiday. So people do buy things for their branches and their dinners and things like that, Easter baskets and apparel for the kids and stuff like that. But it’s not like a Mother’s Day or a winter holiday, so to speak. So it was really more of the way we forecasted it in and the time spent shopping disrupted.
Josh Silverman: Yes. And then what Rachel said, the leading hypothesis of ours that travel during that period impacted this holiday more than others. And it again speaks to the macro that people are just having to make tougher choices. And if they’re spending discretionary budget traveling to see the family, it’s just fewer dollars. They’ve got to make tougher choices and they’re going to just have to spend less on discretionary product.
Mark Kelley: Okay. All right, perfect. I appreciate you clarifying that and making you repeat yourself, sorry. Second one is just the commentary around the team building Gift Mode in four months. Can you talk a little bit more about just using large language models to be more efficient on the R&D side. That seems like a pretty short window and maybe what else we could think about there in the future?
Josh Silverman: Yes. Well, large language models were really helpful for Gift Mode. So for example, there’s 200 different persona in Gift Mode. And then within each persona, there are three to five different gift ideas and the ability to ask large language models, what are 200 examples of persona and it wasn’t quite this simple, but it does give you a head start on that. If I’m a foodie who also loves to travel, what are three things I might buy on Etsy, three different ideas for gifts on Etsy, like it does help to come up with a lot of ideas more quickly. The productivity gains, large language models are starting to help us with coding productivity as well. But in addition to that, we talked last year about key paths and democratizing machine learning.
And that work is paying off, not all of this is large language model machine learning, but we use a lot of machine learning. Once we have an idea for our gift, what’s the five very best examples of that idea amongst tens of thousands of opportunities on Etsy. We use machine learning for that. And so the democratization of machine learning has also been really helpful. We do make steady investments in allowing the teams to work more quickly and those pay off. So I am proud of the engineering culture at Etsy that allows our teams to move really quickly. We have a very entrepreneurial and really agile, really fast development culture at Etsy, and this is a great example of that.
Mark Kelley: Right, perfect. Thank you very much. I appreciate you letting me in. Thank you.
Josh Silverman: Thanks.
Debra Wasser: Thanks. Great way to end, I think. Operator, I think that’s it for tonight.
