Marvin Fong: Yes. I mean you guys kind of basically touched on my next question which was sort of you’re reiterating reaching EBITDA breakeven in the second half of the year, you guys feel good in the sense that the consumer remains weak or even gets a little bit weaker, that you’ve stress tested it and you still feel good that you get that even if the consumer is softer.
Yaniv Sarig: Yes. I mean I think — yes, I think you’re right. First of all, the category — the consumer softness across the board, right? But I think also our experience with some of the categories and the positioning that we have on those in Q2, Q3 is giving us a lot of comfort around that. But again, the most important point here is really the cost base of the product, right? The contribution margin expansion that we were expecting is due to the fact that the efforts that we put in Q4 to make room for that inventory at a lower cost basis is going to create that money expansion. And our logistics team has done a great job of securing a very significant amount of the inventory we needed for the second half at a lower cost which also gives us a lot of comfort around that, right?
I guess there was a lot of challenges Again, as I mentioned in the last 1.5 years, right, not just about the price of shipping but also the ability to even get your goods on a ship. And so now we feel quite confident in our projections given the fact that we were able to secure a position on the ship at a lower cost and then we made room for that inventory to replace the older inventory, right? All these things combined give us that comfort with our prediction.
Marvin Fong: Got you. And I guess my last question, maybe a more fun topic is you mentioned leveraging ChatGPT and OpenAI and that sort of thing. And also curious, I think you guys had always employed some form of AI looking at reviews and helping guide your — our business decisions and product dividends. Could you just kind of expand on what capabilities you’re achieving now or expect to achieve in the near future with these new large language models that maybe you weren’t able to efficiently in the past.
Yaniv Sarig: Yes, absolutely. So we use machine learning and automation across many different aspects of our business. I think if you really want to bucket in the 2 kind of big areas is 1 is understanding the consumer, their sentiment, what they think about other products, they think about our products. And the other side of it is just managing the complex quantitative they do the effort of managing the products, right which includes, for example, forecasting which is where we have developed our own machine learning base forecast and things like media buying and pricing, they use automation as well, right? So when it comes to what you mentioned with ChatGPT and large language model, I mean, I think the excitement obviously for us is huge.
And as I mentioned in my comments before, we already looking at — I mean looking — we are already using ChatGPT to, in a way, actually augment some of the efforts that we’ve had with our own code around sentiment analysis on reviews. But Marc, the most important thing that I think a lot of people that are not necessarily in the weeds on AI don’t realize is that — the hardest thing about AI is actually having good data. And when I say that, what I mean is a lot of organization, except — especially in the consumer product industry that is not necessarily a fact driven, right? Not necessarily designed or have the DNA or have the systems and infrastructure and access — and their own data in a way set up for AI, right? And that’s something that really gives us some advantage, right?
Because we — we’re a consumer company, right but we’re a consumer company that uses all our technology and things like a technology company, we have put use in a position through the years of effort that we put that we are set up to use across the board in many different ways. And again, with the large language models, I think that we’ll see in the next few years, a lot of disruption across many different functions of business but I think only companies that actually have put the effort to prepare themselves to put that data to normalize it to make it available to these models are going to be the ones benefiting from it, right? And I think that’s where again, both with our proprietary software and all the efforts that we’ve made to kind of make the data available to our own algorithms, we have the ability now to relate these advances of technology and get even more efficiency across many different functions, right?
So, we’re very excited about everything that’s happening out there and we’re really happy that we again set ourselves for success by always see in that future and preparing for it and not waiting for it to happen actually putting it to work with our own hands, right? So really excited about all that stuff and what it could do for us in the future.
