It is time. We would like to begin ZOZO's FY 2026 second quarter earnings briefing. This session is being live-streamed only, and this session will end at 5:30. Afterwards, from 5:40, on a different Zoom channel, we will have a Q&A session for institutional investors. Let us introduce the presenters today. President and CEO, Kotaro Sawada. Thank you. Executive Vice President and CFO, Koji Yanagisawa. Thank you very much. We have two presenters today. Without further ado, we would like to have the CFO, Yanagisawa, take you through the results.
I will take you through the FY 2026 second quarter earnings results. This presentation material has been already uploaded on the IR site, so please take a look. Let me take you through our results for this quarter. The GMV for FY 2026 second quarter declined by 1% Y/Y, landing at JPY 309.4 billion.
GMV excluding other GMV increased by 5.7% to JPY 309.2 billion. Adjusted EBITA increased by 5.1% Y/Y to JPY 34.6 billion, and the adjusted EBITA margin was 11.2%, flat from last year. We achieved 45.6% of our plan with respect to GMV, excluding other GMV and 44.4% of our adjusted EBITA plans. GMV decreased Y/Y, but this is because as of the end of September 2025, we ceased recognizing GMV generated by stores operating under ZOZO Option agreements on Yahoo! Shopping. GMV excluding this fell slightly below our plan, but LY Corporation Commerce over-achieved our plans. In total, our performance was broadly in line with our expectations. On the other hand, regarding the adjusted EBITA, we improved the efficiency of our shipping logistics center operations, resulting in packing and shipping costs, as well as logistic cost savings.
Moreover, some actual promotion expenses planned for Q1 remain unused as we are tracking slightly ahead of plans. We have achieved record high GMV, excluding other GMV and adjusted EBITA for the second quarter. Moving on to page nine, here are the quarterly trends and the consolidated results. GMV excluding other GMV in the second quarter increased by 6.3% Y/Y. We talked about this a little bit already, but the ZOZOTOWN business was impacted by lower summer temperatures, which resulted in a lower demand for summer products. Thus, as I mentioned, its performance fell short of plans. On the other hand, the GMV excluding other GMV for LY Corporation Commerce was largely in line with expectations due to increased investment in promotional activities. The adjusted EBITA increased 7% year-on-year, and the adjusted EBITA margin was 10.4%.
This surpassed our expectations due to packing and shipping and logistics cost savings. Moving on to page five of the document. Next, I would like to tell you about the initiatives we undertook in the second quarter towards achieving our mid-term management plan. First, starting with the key area, More Fashion. We took pre-orders of the T-shirt collection commemorating 30 years of Pokemon. The initiative worked extremely well and contributed to new acquisitions as well. Moreover, with WEAR, we utilized our proprietary data and AI and introduced a new feature, Niau Room by ZOZO, which helps users find styles that suit them. The feature comes with Fashion Core Assessment, which finds the most suitable style from out of 28 styles, and users can also virtually try on clothes using Gen AI. We hope that you will give this a try.
Next, with respect to Near Fashion, the one in the middle. In the second quarter, we accelerated our efforts to direct ZOZOTOWN users to HIGH LINK's key service, Coloria. We also opened Coloria's shop on ZOZOTOWN, where they sell their own private label products and selected perfumes. Lastly, regarding our global initiative, the key initiative for this year was to add the checkout function on LYST, and we are continuing to onboard new partners. We started offering the B2B service for ZOZOFIT Pro in the U.S., which enables trainers, clinics, and other businesses to manage and use their customers' body measurement data. So we have started to offer the ZOZOFIT Pro service in the U.S. We will continue to make steady progress in initiatives across all business areas to achieve our mid-term management plan targets.
Now I would like to take you through page 10, but take you through the business results. So here is the year-over-year comparison of the increase/decrease analysis of the adjusted EBITA. Compared to the same period last fiscal year, the adjusted EBITA increased by JPY 1.69 billion, from JPY 32.92 billion to JPY 34.61 billion. Four factors contributed to the increase in adjusted EBITA. First, the ZOZOTOWN business and the LY Corporation Commerce GMV grew, resulting in an increase of gross profits by JPY 3.75 billion. The advertising business sales contributed JPY 0.31 billion. Third, longer LYST consolidation period and HIGH LINK consolidation resulted in an increase in gross profit of other businesses by JPY 0.97 billion.
Fourth, we have also successfully reined in packing and shipping costs due to improved economic terms agreed with our logistic partner since the third quarter of last year, and we also improved the efficiency of our logistics center operations, thereby driving down logistics-related payroll, reducing the variable costs by JPY 0.53 billion. Three factors drove the adjusted EBITA down, on the other hand. First, depreciation of material handling equipment began in the existing logistics centers, and HIGH LINK consolidation added to the number of consolidated employees, resulting in an increase in fixed costs by JPY 1.41 billion. Second, acquisition and promotional costs increased, and we also saw an increase in the expenses related to the LYST and HIGH LINK consolidation, resulting in an increase in actual promotional expenses by JPY 1.58 billion.
Third, higher consumable expenses driven by rising packaging material costs, along with increased IT and communication expenses due to greater use of cloud services, led to an increase in other costs by JPY 0.88 billion. Next, page 11. This is the balance sheet. Compared with the end of the previous fiscal year, cash and deposits under current assets decreased due to share repurchases, transfer of funds to a securities account in preparation for further share repurchases, and payments of dividends and corporate income taxes, among other factors. In addition, treasury stock increased with shareholders' equity because of the share repurchases. Regarding share repurchase, we are making steady progress, and as of the end of September 2026, we have completed JPY 19.1 billion of the maximum JPY 30 billion. Next, page 12 of the materials. Here are the changes in cash flow.
First, cash flow for business activities increased by JPY 4.2 billion year-over-year, primarily due to an increase in profit before income taxes for the first half. Second, cash flow from investing activities decreased by JPY 6.3 billion year-over-year, since expenses related to the acquisition of HIGH LINK was relatively lower than that spent for LYST. Lastly, cash flow from financing activities. This increased by JPY 19.1 billion as we repurchased our shares. Next, page 23 of the materials. I would like to take you through the changes in SG&A expenses. SG&A expenses as a percentage of GMV was 22.6%, decreasing by 0.4 points year-over-year. There are two key reasons why.
First, we improved our shipping efficiency, as we were able to agree on more favorable economic terms with the logistic partner in October of last year, and we saw the ratio of smaller packages, NEKOPOS, increase as well as AOV, resulting in 0.5 point decrease. Second, we also optimized the inventory levels in the logistics centers and improved operational efficiency, reducing logistics-related payroll by 0.2 points. There are also two reasons why the SG&A expenses increased. First, in addition to the increase of digital ad expenses for ZOZOTOWN, the longer LYST consolidation period and HIGH LINK consolidation led to an increase of 0.2 points. Second, goodwill amortization expenses increased by 0.2 points due to the longer LYST consolidation period and HIGH LINK consolidation. Next, page 26 of the materials. Here are the trends in the actual promotional expenses.
Actual promotion spending in the second quarter, which includes advertising and promotional expenses, as well as point-related costs and other incentives recorded as deductions from revenue, was 5% as a percentage of GMV. Year-on-year, this percentage rose by 0.3 points because of these three main factors. One, free shipping campaigns and strengthening our personalized discounts initiatives on ZOZOTOWN. Second, increase in advertising for LYST. And thirdly, recognition of HIGH LINK advertising expenses. These are the three key factors. The spending for the second quarter was largely in line with our plans, but we still have unspent budget from the first quarter. The spending in the first half fell slightly short of plans. We plan to spend these funds in the third quarter onwards, so the spending for the full fiscal year will be similar to that of last year or 4.8% as a percentage of GMV.
These plans have not changed. From here on, I would like to take you through ZOZOTOWN's KPIs. Page 27. The number of total buyers. From here onwards, the results of LY Corporation Commerce and LYST or our B2B business are not included. The annual buyers increased by 130,000 to 13.55 million, comprised of active members, which increased by 120,000 to 12.84 million, and guest buyers increased by 10,000 to 710,000. Our digital ads and various initiatives within ZOZOTOWN continued to steady acquisitions in the second quarter as well. Especially in September, ZOZO Week, which we did not organize the same period last year, contributed to new acquisitions. On the other hand, because we strengthened digital ad efforts and introduced referral campaigns last fiscal year, when compared to the recent couple quarters, the increase in the number of active members was rather moderate.
There was an offsetting of positive and negative impact. Next is the average retail price trends. This is page 32 of the earnings material. The average retail price was JPY 3,640 , increasing by 1.6% year-on-year. Although the prices of new fall/winter items increased slightly, overall price levels of new products remained largely unchanged year-on-year. We organized the ZOZO campaign in September. Thus, the discounted sales ratio increased marginally, but as the average discount rate declined, the average retail price increased slightly. Moving on to page 33, this is average order value trends. The average order value was JPY 8,388 , up by 2.5% year-on-year. We conducted more frequent free shipping campaigns. Thus, the number of items purchased per order increased.
This, combined with the increase in average retail price, led to a higher increase in the average order value than the average retail price. Lastly, page 36. Here are the consolidated earnings forecast and dividend forecast for this fiscal year. The forecast remains unchanged. We would like to continue to do our best to achieve our targets. That concludes my portion of the presentation. I would like to give the floor to our CEO, Sawada.
I would like to tell you about some of our strategies that we are considering going forward. What I will be talking about is AI or AI agents. I am sure you hear AI agents quite often nowadays. You probably hear those words every day. I am asked this question quite often, so I would like to talk about our thoughts regarding AI agents.
As you can see here, this is a viable model. Internet services, this can be applied to most Internet services, and of course, this applies to us as well. What is written here is that supply and demand. The internet services provide value by linking supply and demand. There are many businesses that are helping connect supply and demand. Step 1, this is from before, but differentiation through volume, many companies start there. We started there by welcoming a lot of brands. First step, we evolved through adding supply, increasing supply, and then step 2, differentiation through search convenience, and recently, differentiation through personalized convenience as well. We have been providing services that enable consumers to find what they want, find things that match their preferences. I am going to talk about AI agents now.
Right now, AI agents are embedded in PCs and smartphones of our consumers. AI agents, in terms of scenario 1, they can actually directly contact supply to find items that would match the preferences of consumers. Scenario 2 is something that we are also thinking about. Of course, AI agents will continue to recommend items that the consumers would want, match their preferences, but supplying usual agents will be a little bit more generic. The recommendations will be more generic. If we have the supply agent and the demand agents converse with one another, we would be able to provide even more optimal products to consumers. That is scenario 2. It is not either/or. It depends on the product, generic AI agents and recommendations, as you can see below.
Scenario one, of course, where information on merchandise and services is clear and widely available as public information. Those kind of products, scenario one would apply. On the other hand, if the service and product information is a little bit vague, ambiguous, and public information is not available, supply side has proprietary data around such products. We are developing agents that can reference our proprietary data and make even more sophisticated recommendations. Scenario two would apply to fashion. Of course, there is publicly available data, but depending on who wears it, the product will look different, fit different. There is a lot of that kind of data. We have that data on purchase history and various preference data as well. We have a lot of different types of data. Right now, ZOZO is trying to head towards scenario two.
Moving on to the next page. This is a more detailed diagram of what we just talked about. Through ZOZOTOWN, ZOZO has done basically the blue arrow portion. Through touchpoints, we promote various products, and we have the users discover what they want, and we invite them to ZOZOTOWN, and they make purchases through ZOZOTOWN. Once they buy something, then we will continue to compile information or data about what they purchase, what they like. By running this cycle, of course, we continue to improve the customer's LTV. That is something that we continue to do. Our positioning of the AI agent, the blue loop. The AI agents would actually, we were trying to create the orange portion. This is a portion that the AI agents will play in. It is a loop that sits within this blue loop.
As I mentioned before, the way products look or the way consumers think about products, way it fits differs depending on the person. What looks best on them? It is a lot of complicated information, and we are creating algorithms and agents that will be able to tell you what would look best for them. It is not just about buying products. People will be able to know what looks good on them, what is optimal for them, and they can make a purchase. They can also learn about how better to wear the products that they have as well, so they can ask the algorithm or AI agents. We want to create another loop around styling. What is included within this orange arrow is that on the right, this is what we have done in the past couple of years.
The most primitive way is to introduce how to wear the clothes that you have on YouTube. We do not only introduce products, we of course, wanted to show users how they can wear them. We have done this over the last three years, and we have a lot of viewers now. There is a lot of demand for how to wear the clothes, online as well. On our LINE account, we have an AI conversational styling proposal. We call it AI Lab-kun, but we release this AI recommendation feature so consumers can speak to this AI, and we are fine-tuning the styling recommendations. WEAR, as Yanagisawa mentioned, we have the Fashion Core Assessment, so you can find something that you like very easily, and you can also see the style that would look best on you, suit you best.
We would be able to do an assessment and propose the best styling for the user. In recent years, GenAI or virtual fitting is becoming quite standard, but we are also offering a virtual try-on service as well. Also on ZOZOTOWN, we are still building this, but we are building an AI where people can talk to that while they are selecting products. We are trying to create these kind of agents right now, and we also believe that touch points outside of internet is necessary, and we are also trying to create something that we have temporarily coined Neo Mirror. Those kind of devices we are developing right now. Going back to the proprietary data, we do have unique data, and we want to utilize that data to sophisticate our niau-styling algorithm so that we can continue to recommend the best styles for our consumers, our users.
This is a summary of what I just explained, but as you can see here, the objectives are slightly different. However, we want to basically enable, empower people to find their style. We want to create a system or engine service so that people can find their style. We will continue to fine-tune our services so that we can realize the orange loop that I touched upon before. That concludes my explanation. Thank you very much.
That concludes ZOZO's FY 2026 second quarter earnings briefing. Thank you very much for joining us.