KE Holdings Inc. (BEKE) Earnings Call Transcript
August 21, 2026
Earnings Call Speaker Segments
Good day, everyone, and welcome to the KE Holdings Second Quarter 2026 Earnings Conference Call. Speaking to you is Siting, the company's IR Director. I want to let you know that management's prepared remarks along with the Q&A portion today will be conducted entirely in Chinese. An English simultaneous interpretation will be provided on a separate line throughout the session. If you wish to hear the gold in Chinese, please dial in to the Chinese language line. <<operator_script>> This conference call is being recorded. Earlier today, the press release containing the company's financial and operating results was issued and made available on the company's IR website. Joining us on today's call are Mr. Stanley Peng, our Co-Founder, Chairman and Chief Executive Officer, and Mr. Tao Xu, our Executive Director and CFO. An overview of our business update and financial performance will be given by Mr. Xu. After that, Mr. Pang will discuss in greater detail how our strategic transformation is progressing. Before we go further, please take note of the safe harbor statement included in our earnings press release, which applies to this call, since we will be making forward-looking statements. Be aware that KE's earnings press release and this conference call both cover discussions of unaudited GAAP financial information and unaudited non-GAAP financial measures. For a reconciliation of the unaudited non-GAAP measures to comparable GAAP measures, please see the company's press release. Finally, all figures referenced on this call are denominated in RMB unless stated otherwise. Some statistical and other information concerning the industry in which the company operates that will be mentioned on this call has been gathered from various publicly available official or unofficial sources. Such data has not been independently verified by the company or any of these representatives, and it may involve a number of assumptions and limitations, so you are cautioned against placing undue weight on this information and assets. On today's call, management will conduct the review primarily in Chinese. Please be aware that the English translation is provided solely for convenience. Should any discrepancy arise, the management statements in their original language will take precedence. At this point, I will hand the call over to our CFO, Mr. Tao Xu.
Thank you, Siting. Hi, everyone, and welcome to the earnings call for Q2 2026. I'll start by walking through the main financial points. Our total GDV went back to growing. Even though revenue slipped modestly compared with a year earlier, profits climbed substantially, delivering results well ahead of both GTV and revenue. During Q2, that figure rose 6.3% versus the prior-year period, even as revenue fell 5.7% year over year. The drop in revenue came mainly from changes in our home renovation and furnishing business, together with revenue recognition effects tied to iterative product modeling in home rent services. Non-GAAP Net income climbed 74.9% compared with the same period last year, hitting CNY 3.185 billion. Our non-GAAP net margin landed at 13%. That's up 6 percentage points compared with a year ago, and it's the strongest we've seen in 3 years. The improved profitability stemmed from a cost base that's gotten leaner, disciplined management of finances, and more efficient operations. Every one of our core business lines posted higher contribution margins, both versus last year and versus the prior quarter, which pushed the group's gross margin up 6.7 percentage points from a year earlier to 28.6%. Meanwhile, GAAP operating expenses fell 14.1% year over year. It was the combination of a wider gross margin and lower operating expenses that drove our profit growth. I'll now walk through how each segment performed financially. Starting with existing home transaction services: in Q2, scale came back to growth and profitability improved significantly. On the gross transaction value side, we reached RMB 629.89 billion — that's up 8% from a year ago and 17.9% higher than the prior quarter. Turning to revenue, we posted RMB 7.02 billion, which represents 4.5% year-over-year growth and 14.5% growth quarter over quarter. The reason GTV outpaced revenue on a year-over-year basis mainly comes down to a larger contribution from non-Lianjia GTV, since platform service fees there are recorded on a net basis; in this quarter, revenue from non-Lianjia platform services climbed to 27.8% year over year and 29.8% quarter over quarter. As our network scale remained stable, we kept sharpening our operations to boost per-store output, and that enabled connected stores to outperform the market while also improving efficiency across the platform. For Q2, contribution margin was 46.1%, up 6.1 percentage points from a year earlier, driven by lower fixed labor costs and a greater share of platform service revenue, which carries higher margins. Sequentially, it also improved by 4.8 percentage points, aided by operating leverage, a rebound in net revenue and further gains in business mix. Now turning to the new home business. During Q2, scale remained level compared with the prior-year period while profitability continued to strengthen. GTV totaled RMB 258.39 billion, an increase of 1.2% year over year and 77.1% quarter over quarter. Revenue was RMB 8.95 billion, up 3.8% year over year and 75.9% quarter over quarter. Despite a challenging market, we maintained stable scale by focusing on high-quality projects, improving customer conversion and reducing costs. For Q2, contribution margin reached 28.8%, which is up 4.4 percentage points compared with a year earlier, supported by optimizing the cost structure through refined operations. Those same factors, along with operating leverage tied to revenue growth, also drove a 3.1 percentage point increase versus the prior quarter. Now for the third item: home renovation and furnishing. Revenue in Q2 totaled RMB 3.19 billion, down 30.1% from the same period last year, but up 36.4% from the previous quarter. The decline year over year reflects the changes we intentionally made to underperforming customer acquisition channels, as well as our withdrawal from cities where unit economics were weak. Softness in the new home market additionally dampened renovation demand. The sequential rise in revenue reflects the usual seasonal pickup in business. In Q2, contribution margin reached 39.6%. That's up 7.5 percentage points from a year earlier and 3.4 percentage points higher than in Q1, driven by cheaper materials thanks to centralized purchasing and tighter cost control. Now, fourth, home rental services. Revenue for the second quarter was RMB 4.83 billion. Year over year, that's down 14.8%, and compared with the previous quarter it fell 3.6%. What explains this? We moved Carefree Rent to a lighter, lower-risk product model that recognizes revenue on a net basis. While that shift reduces reported accounting revenue, the count of managed rental units continued to grow rapidly. At the end of Q2, we had crossed 790,000 managed units — up about 34% year over year — and net base products accounted for over half of that. For Q2, contribution margin was 15.3%, an improvement of 6.9 percentage points from the same period last year. That gain reflects a better product mix as our operations improved, thanks to lower labor, installation, and post-lease costs. Sequentially, contribution margin rose 0.5 percentage points, driven by consistent growth in net based products. Fifth, emerging and other businesses: Q2 revenue reached RMB 550 million, which is up 26.4% year over year and 7% quarter over quarter. Moving on to costs and profitability. Costs tied to our stores came to RMB 560 million during Q2, which represents a 25.9% drop compared with the same period last year and was essentially unchanged from Q1. That improvement versus last year stems from rent cost optimization at Lianjia as well as adjustments to our network. Our gap operating expenses for Q2 overall reached RMB 3.989 billion, down 14.1% from the prior-year period, driven by stronger organizational efficiency, more optimized marketing spending, and ongoing financial discipline. Sequentially, operating expenses rose 21.3%, which we attribute to greater selling expenses resulting from the seasonal rebound in home renovation plus bad debt provisions within the new home business. Specifically, G&A expenses came to RMB 2.04 billion, representing a 2.1% year-over-year reduction. A full bad debt provision of roughly RMB 280 million — made after a prudent review of receivables and collateral tied to Seneca — is what drove that 18.9% rise from the prior quarter. Marketing costs landed at RMB 1.4 billion, down 26.1% versus a year earlier, which reflects better sales optimization, personnel savings, and smarter marketing outlays, and yet compared with the preceding quarter they climbed 29.6%, driven by the usual seasonal lift in rental and home renovation selling expenses. R&D spending totaled RMB 550 million, a 13.4% year-over-year decline from lower labor and technical service costs, while quarter-over-quarter it rose 11.4% on higher technical service fees. Turning to the bottom line, GAAP operating profit for Q2 reached RMB 3.026 billion, an increase of 185.6% over the equivalent quarter a year ago. On a non-GAAP basis, operating profit reached RMB 3.59 billion, which represents a 123.6% increase for the year. GAAP operating profit rose 137.8% compared with the prior quarter, with a margin of 12.3% — that's 8.3 percentage points higher than a year earlier and 5.6 percentage points above the previous quarter. Non-GAAP operating profit climbed 115.7% quarter-over-quarter, with a 14.6% margin, which is 8.5 percentage points above the prior year and 5.8 percentage points above the prior quarter. Higher gross margins and reduced operating expenses for issues were the main drivers behind this margin performance on both a year-over-year and quarter-over-quarter basis. In Q2, GAAP net income came in at RMB 2.624 billion, up 100.8% versus a year ago and 109.1% versus the previous quarter. Non-GAAP net income totaled RMB 3.185 billion, marking a 74.9% increase year-over-year and 97.6% quarter per quarter. Now, let me wrap up with the areas of cash flow, our balance sheet, and what we return to shareholders. In Q2, our operations brought in net cash of RMB 6.61 billion. The tenor for net new home accounts receivable came in at roughly 39 days, which is about 12 days lower than the same period a year earlier — a sign that our risk management is working well. If you set aside customer deposits, our broad cash balance at the close of Q2 stayed at approximately RMB 67.3 billion. That level of liquidity gave us more strength in managing risk and being resilient, while also funding business growth and returns to shareholders. In the second quarter, about USD 250 million was used to repurchase shares, and that amount covers the first buyback we carried out in the Hong Kong market. Across the first half, repurchases totaled approximately USD 450 million, which marks a rise of about 14% compared with the same period a year earlier and equals roughly 2.4% of the shares we had outstanding at the close of 2025. From the time this buyback initiative began in September 2022 up through Q2 2026, we have bought back shares worth around USD 2.99 billion, which works out to about 14.8% of the shares that were outstanding before the program was introduced. To sum up, the better profitability seen in Q2 comes from a mix of cost optimization, operational improvements, and a favorable business mix. Going forward, keeping a strong balance sheet along with plentiful liquidity will serve as the foundation for our growth over the long term. For every one of our core new and technical investments, we will require rigorous ROI display and treat customer value, operating efficiency, and sustainable returns as our priority, or our key metrics. In the end, we will strike a balance between developing the business and returning capital to shareholders so that long-term value is created consistently. I will now hand the call over to our Chairman and CEO, Mr. Stanley Peng.
< > Thanks. Good evening, investors and analysts. During the prior quarter, we talked about moving toward a consumer-centric transformation; this quarter, I'll explain how those changes show up in how we run things. In Q2, I saw 2 trends: the operational base became stable, and our organization genuinely got organized. That base makes long-term change possible. I'll take up 5 key questions. First, what shifts as transformation reaches the failure operations stage; second, whether consumer-centric means going around agents; third, whether AI progress will make even become obsolete; fourth, how AI is deployed in our business and what comes of it; fifth, how we'll recognize we're on the right path going forward. On the first question — what has changed as transformation reaches failure operation — this quarter I devoted lots of time at the front line < > stores, properties, construction sites, talking through problems with clients, agents and store owners, and those changes come down to 3 areas. First, refined operation: we're moving away from the 1 side fit oil approach toward district-specific and project-specific strategies. Instead of looking at one city-wide metric, we examine particular districts or projects to < > solutions, and what the solution is for each < >. For instance, in the high-end community where clients feel property across districts, our legacy geographic bound model field, and we regrouped operational units around extra clients viewing < > assigning project efforts, along with professional presentations and client experts, to address particular family needs. With 600 projects driving half the cities volume, turning these professional judgments into a clear division of labor lets us replicate this model, and other cities have become similar operations — exploration. Second, shifted metrics: scale and market share still matter, yet now we care more about consistent agent transactions, writing agent efficiency and income, healthy store profitability, and stable service quality. Leasing LSG plays perfectly. In 2025, we had at most 700 agents for leasing at the peak, and the average agent efficiency fell below 2 transactions. Rather than adding headcount, we divided it into smaller blocks, rematching properties, clients and agents based on familiarity and the capabilities. From April to July, average agent efficiency jumped from 3 to 5.6 transactions, and the zero transaction ratio dropped from nearly 25% to under 10%. So I think what matters is that the effective organization matters more than near headcount. Third, mobilized people: managers have elected meeting rooms for front line. This quarter, managers personally sold < > listings, revisited debt leads and accompanied agents to signing centers; and my only requirement for manager is to presence. You cannot learn to swim without getting into the water. In short, operationalizing transformation wins refined operations, shifted metrics and mobilized people. This stems from a sing approach: solving real consumer and frontline problem first, than reorganizing our people resources in the platform. Alright, so changes are underway, and we're now seeing them show up inside the operational units. Now, the second question: when we say consumer-centric, does that mean agents get cut out of the picture? That framing assumes that if the platform gets nearer to the consumer, it has to come at the agents' expense. In the past, we simply divided one transaction's commission, which amounts to a zero-sum setup. That's exactly what we did before. To break that equation, though, we have to generate more light, generate higher-value tasks, rather than just carving up the same pool of money again. Consumers aren't static. Good used to mean aesthetic property < > nowadays, I'd say good means the right match < > determine goods widened from 1 to 3: the property, the family situation, and the service provider too. That makes the service provider a crucial variable now, not merely a conduit. When the decision gets harder, the tasks have to be broken into segments. There are 3 reasons. First, the knowledge required. It exceeds any one person's capacity; for instance, we had to understand the properties, the clients, their circumstances, mortgages, renovations, and the furnishing side of the business. So it goes beyond what one person can handle. Second, developing expertise demands a mutually exclusive path. You either embed yourself deeply in one project, or you follow a group of clients. Doing both at once isn't possible. That's reason number two. The third: the most valuable action has shifted away from offering options toward confidently removing them. I'd say we aren't just handing consumers more choices; at the start, we had to help them filter. Filtering, though, isn't the same as a transaction. If income depends only on closings, genuine professionalism never takes shape. In my view, professionalism has to be financially viable. So because of that, we are < > income from closed deals, aligning it completely with the buyers or seller. This assisted < > is the client manager. Before, the platform, in effect, halted the moment a lead reached an agent. The client manager guarantees continuity. AI arranges the data, while human services < > stage and needs. The agent gets clients who are fully provided, and since client managers aren't paid per transaction, they stay purely objective; so as I said, managers aren't paid per transaction. From May to July, this will handle over 50,000 < >, reaching a 7.4% related to showing conversion rates, beating the broader market's 5%. So the platform I mentioned is evolving from leading commission toward building a structure in which every specialized scale is independently verified and compensated. ECM is shifting from a single listing workflow into a modular ecosystem, covering consulting, showing, contracting, reporting, marketing materials, renovation, leasing, and so on. Anyone generating incremental value counts as a service provider. That's our definition, and it's expanded. The main goal is enabling professional service providers to win over the long term; so being consumer centric means turning the single agent into a group of independently valuable specialized roles. So at this point, AI is lending us a hand and adding momentum. That brings up a question: as AI keeps getting better, will agents be left behind? That view presumes agents merely pass along fixed details that AI can pull up without effort. But technology rearranges where value sits — certain things lose worth, while other things grow harder to find. So we ought to ask: which things are losing worth, and which are growing harder to find? On the harder-to-find side, what progress can a platform managed service provider actually achieve? And what's losing worth? Fixed details, < > revenue, price and year built, plus the home's floor plan. In my view, that sort of detail can't carry a decision, and it's extremely simple to obtain. So if all we do is pass along — if all we do is move details around, we might run out of chances down the road. What's harder to find is dynamic, deeply inspiring, and in fact can't be manufactured — say, why someone is selling, what renovation could yield, or a read on the local market from experienced managers, along with how things stand in the communities according to those managers, what's closing, and how the deal went previously. That kind of detail sits inside people's heads, and this industry favors a pipeline that captures and reuses it over time. And at bottom, AI isn't the outcome of bad choices. And AI might not bear any responsibility.
So when the expense of getting housing wrong goes up, buyers have to cut back on uncertainty-driven expansion. That leads to three outcomes. First, the sector gains value by lowering uncertainty; carving out segments creates worth that's tough to achieve, since it demands rich data and even deeper survey work. The initial point is that anyone who adapts in this direction gains value, and that covers platforms along with managers. So instead of needing information plus participants, we need experts willing to exercise judgment and shoulder accountability. That ties back to the earlier question regarding the sector and service providers. When we scan the landscape and when we examine ourselves, the question becomes: how does AI get used inside our operations, and what comes of it? In truth, the operation itself acts as a reduction function — what goes in, what comes out — and within that function sit human capital, labor capital, and technology. Given where AI stands today, we ought to understand its position within the sector. It could be a subitem or a direct variable. If it's a supply unit and an efficiency instrument, or if it's a direct variable, then attrition is required. So attitudes had to shift first; we've now also unlocked certain foundational pages, and we're bringing the threshold down. We worry about whether disruption lies ahead, and we're considering how AI might serve as a fresh production factor instead of an adversary that enables innovation. My view is that ultimately consumers < > value. My view is that consumers deserve a superior experience, and we must fix consumers' problems. Second, it alters management. Across the last 200 years, we've seen gains in finance and management, and management requires data we can quantify. I believe all of us gain from this methodology. At KE Holdings and likewise at Lianjia, standards are necessary, and we also require tools to drive improvement. Yet things that can't be quantified can't be measured — that's a major issue too. Still, at times we might fixate only on that figure, and we discover that — at times we discover the figures are overly abstract, and consumers wind up as numbers, and wind up as the #1 in the standard. Yet thanks to AI, AI delivers unstructured data and < > language, and numbers are entirely distinct information and signals, and granularity moves from managing the average toward managing individual properties, clients, and agents. Before, we managed the average, but today we possess the computation power and the knowledge, and we can craft a tailored solution for every individual. And this third party concerns how AI alters the division of labor. We discuss task segmentation, or segmentation within the company, brought by AI. Now we have these scenarios, spanning finance, human resources, product, technology, and also progression stage by stage, plus the computation power. But at present, we are dismantling the threshold, and every one of them falls within AI's computation power. And before that — when the old divisions combine, fresh ones appear. Within our Lianjia new home operation, we reallocated work between people and artificial intelligence. AI assists in comparing proposals through a dynamic, knowledge-based system, which lets agents concentrate on grasping what clients need; that way, agents could refine how well they understand each customer. This yields signed deals as well as valuable organizational skills; those arise only from the frontline; this kind of disruption restructures the company. It touches on four points. The first is cost. AI reduces fixed expenses while raising variable ones, which makes quick iteration possible. So whoever iterates quickly generates greater value; the second is experimentation. In earlier times, this demanded considerable effort. At present, going through the — evaluating, testing, and validating a proposal follows a lengthy path. The larger the company, I mean, the longer that chain becomes. Consequently, plenty of people simply hold back. At present, AI moves innovation away from heavy, sluggish investments toward frequent, inexpensive profitability improvements; this lets us try out and examine several models simultaneously, and our odds of coming out ahead in the game rise. The third is the frontline and the mid-office. Frontline staff < > are able to quickly construct and try out solutions, and the mid-office can then expand them. Finally, and importantly, managers. In earlier times, the larger the company, the weaker its efficiency. At present, I've spoken with numerous managers; we don't sense much of a percentage of value. At present, AI < > the organization is altering the role, taking care of reporting. We're pushing managers to quit acting as a megaphone and begin producing genuine business value; thus they aren't merely showing numbers, they're truly generating authentic value from the frontline. Since they're involved in producing that value. Lastly, in the end, the bottleneck moves toward humans. Consider KE. We have an extensive industrial process. AI can refine many of the workflows. And in those spots where human invention turns into the bottleneck, there's this human and human minor action that AI has to substitute; so whether we can bring people together and give them training, enable them to work effectively alongside AI, one aspect is culture and the other is evolution; so this is fundamentally the shift we're discussing across the entire industry. Now returning to the very first question, is AI a direct variable? Since it alters whom we serve, our judgments, our process, and our organization. So this is a direct variable; that means we aren't merely installing AI into the company, we're regrowing the company with AI. Now, when we think about what comes next, how do we actually tell that we're headed in the right direction on the red track? That really comes down to 2 separate things: where do we have to put big bets, and where are we looking for answers? In my view, there are 3 areas where we're making those big bets — deep service, deep data, and a platform ecosystem. As information becomes democratic, deep data turns scarce, decisions get harder to make, and deeper service grows more valuable. As labor becomes more specialized, you need a platform to orchestrate it. So while we're hunting for answers, what AI ultimately becomes and how management and expertise will finally be structured are still unclear. Matters that are directional call for big bets. So how do we capture users' evolving needs? Management, naturally, is where that value sits. Matters that are more logical, though, call for small investments, fast testing, and cutting losses early. So why must we separate things by certainty? Because, once again, we've already shown that directional matters demand unwavering bets, while morphological matters demand more investment in rapid testing. Looking back at the past 2 quarters, we've acknowledged in some areas that continuing to invest where marginal returns are low makes no sense. A model driven purely by skill is debt. We should halt those pointless investments. Above all, we have to validate 4 things. First, professionals. When faced with AI, can they use it directly or indirectly to create value?
Are there fresh criteria for defining professionalism, and are they truly dedicated to that idea? And next, when it comes to managers, can they get back onto the front line and deliver top-notch judgment that rebuilds that feeling of value? Third, there's processes and judgments. As services go deeper, are they able to gain customer trust, and can they achieve wider recognition — stronger recognition or trust? Fourth, organizational capabilities: can one win be converted into something repeatable? So amid this kind of discontinuous transformation, for me, industries are basically up against the same test. From my perspective, human conviction serves as the leading indicator, while numbers act as the lagging one. A lot of management tends to keep their expertise tucked away inside themselves. So absent open sharing, we can still turn that into a repeatable successful model. Thus our central test is whether we can consistently carry out consumer centricity and let professionalism win; that has to be woven into our culture and workflows. So we'll mainly succeed across 4 pillars: customer, service provider, operations and replicability. So when you examine these 5 things, we need to briefly lay out our playbook. Consumers now face tougher choices; that is fueling deeper specialization. Meanwhile AI is devaluing role into while lifting genuine expertise and reorganizing internal work; so our course is clear — deep service, deep data and a platform ecosystem; so Q2 isn't the finish line, it's only the start. Thank you. I'll now hand the call over to the analysts for Q&A.
Thank you, Stanley. <<operator_script>> The first question comes from Timothy Zhao from Goldman Sachs.
Nice work on the solid second-quarter numbers. I'd like to ask about the property market as a whole. During Q2, we observed volume and price moving in different directions, and there's been some choppiness and momentum carrying into Q3. With so much uncertainty lying ahead, which levers can the company actually control for Q3 and for the entire year?
Thanks, Timothy. During the first half, transactions in the pre-owned housing sector demonstrated a structural rebound, with pricing reaching its floor. That rebound grew clearer in Q2, extending to parity in payment terms across different cities and price tiers. As of < >, deal volumes picked up more quickly in Tier 1 cities, and there the first-half prices also proved more resilient on a sequential basis. During Q2, registered pre-owned home transactions in Tier 1 cities grew year over year at a pace exceeding that of other cities. Data from Baker Research Institute shows that in the first half, pre-owned prices in Tier 1 cities climbed 3.6% cumulatively on a quarter-over-quarter basis, whereas prices nationwide stayed broadly flat compared with a year earlier. Pricing at every tier is still undergoing an adjustment. On our platform, sales of cheaper homes expanded at a quicker pace than those in the middle and upper price brackets. Still, the mix of transactions by unit size has held steady, which tells us housing demand has not shifted broadly toward smaller units. What it actually shows is a migration downward in transaction price bands, driven by price adjustments. At the same time, pricier homes experienced less severe year-over-year price drops, demonstrating durability in core upgrade-focused and premium-quality residences. Turning to newly built homes, volume overall in Q2 stayed under strain. Projects located in core cities that featured compelling product offerings drew better support. From a structural standpoint, pre-owned homes made up more than 80% of total national residential transaction area during the first half, becoming the market's < > for housing demand. Broadly, we are witnessing a structural transaction rebound while prices keep bottoming out; core cities along with high-quality supply are holding up better, yet the market remains pelorized. With a wider array of property options, buyers are making decisions carefully, placing value on professional judgment and certainty of transaction. They require professional decision support, rather than merely transaction matching or facilitation. This underscores the service capabilities our platform has built up. On that basis, we will concentrate on 3 areas. First, seizing structural market opportunities to bolster revenue recovery. We will direct resources according to market performance across cities, customer groups and property tax, strengthening our coverage in higher-tier cities. At the same time, centering on content treatment engagement, precise matching and professional execution will assist customers in making sounder decisions and turning genuine demand into completed transactions. Second, we'll keep reinforcing financial discipline and flexible resource allocation. Our leaner cost structure sharpens our capacity to < > market volatility. Should pressure continue, we will reallocate resources dynamically, putting our core professional service provider network ahead of near-term profit. Even in the event the market picks up, we will not revert to extensive expansion. New investments < > ROI and service validation before scaling, making sure transactions convert efficiently into profits and cash flows. Third, we'll likewise put cash flow and a solid balance sheet first; we'll manage receivables and collections strictly, control risk exposure and cap nonessential investments so as to preserve flexibility. So because of that, how we run things in the back half of the year won't hinge on what the market does. When it comes to revenue, stronger decision support is going to let us bring in additional customers. And on the financial front, since our cost structure is in better shape, it'll shield cash flow along with our core capability when markets are soft, while unlocking more operating leverage as the market gets better. Thanks.
Our next question comes from John Lam from UBS.
Thanks for taking my question, Tao. I'd like to ask about Q2, when the pace of revenue growth had a significant impact on profitability. Could you walk us through how much of that came from business performance, from operating efficiency, and from the expense baseline? Did any one-time items play a role? And looking ahead, how sustainable do you expect those gains to be over the longer term?
Appreciate the question. In Q2, the profit improvement was driven primarily by stronger contribution margins across the core business, along with lower operating expenses. On the core business side, contribution margins increased on both a year-over-year and a quarter-over-quarter basis, lifting the group's gross margin to 28.6%, which is up 6.7 percentage points from a year earlier. At the same time, GAAP operating expenses were down 14.1% compared with the prior-year period. Three factors are behind this. To start, the reduced cost and expense base we had in the previous year resulted from streamlining Lianjia < > along with the agent framework—specifically, by broadening management expense, merging resources, and cutting back investment in low-productivity areas. That lowered our fixed labor cost and our breakeven point, giving us a lasting baseline as well. Next, improved operating efficiency in housing transactions for new homes—through securing coverage or high-quality projects and boosting customer conversion—made our transaction business more resilient. On top of that, we maintained steady monetization, and greater channel efficiency fueled profit growth. When it comes to the homes that are already part of our portfolio, we zeroed in on what mattered most — getting more listings and sharpening the operational backing we give to connected stores. That combination delivered a meaningful lift to both revenue and profit contribution from those connected stores, which in turn strengthened the unit economics and shifted the mix within our new business. On top of that, we moved procurement toward a centralized model and tightened up how we manage costs, which brought down the ratio of material costs within home renovation. Over in rental services, contribution margin came out better, driven by a mix shift toward a net basis revenue product, plus the gains we saw in January around labor, installation and post lease costs. When we think about the upcoming two quarters, and assuming the market behaves in a mutually consistent way, that reduced cost base is going to help us fund the incentives tied to our profit driver marketing channel. At the same time, some of what we spend on frontline sales could move up or down from one quarter to the next, depending on how much revenue we bring in, what mix we sell, and normal seasonal patterns. Rather than taking one quarter's profit and projecting it forward in a straight line, our attention stays on growing revenue and profit in a balanced way. Should conditions in the market get better, the extra revenue we gain would pair with stronger operating leverage coming off that lower baseline, which means more profit on the upside. If instead the pressure sticks around, having a healthier structure makes our profit less sensitive to swings in the market — to put it plainly, the setup we have right now gives us more room to gain when things go well and more cushion when they don't. Over the longer term, though, this optimization creates a sound operating foundation. This is the first of one phase in our strategic overhaul. We're reworking how resources get distributed given where the market stands right now, and that's the way we can handle the ambiguity. Phase two involves steering our constrained resources into efforts that generate customer funding, instead of merely reducing expenses. In the end, by addressing workflows, evaluations, inventive, and the platform tools, we'll drive efficient resource distribution throughout our dealer organizational capacities, which supports sustainable growth.
Thank you, Mr. Tao. The next question comes from Xiaodan Zhang from CICC.
Nice work on the strong Q2 results. My question concerns existing homes. During Q2, existing GTD was up 8% from a year earlier, and contribution margin gained 6.1 percentage points. How much of that comes from the market bouncing back versus what the company itself is doing, and which metrics show this operating < >?
Thanks, Xiaodan. Great to hear from you. Briefly, although the market rebound gave us a base for transaction volume, our existing home operating offer didn't come from growing our network or rising prices; it came mainly from greater unit productivity inside a stable network and from converting platform service value into revenue more effectively, which the simultaneous margin gain proves we didn't trade profitability away for growth. To be specific, in Q2, existing home transaction volume across our key cities recovered modestly, with prices stabilizing sequentially, which gave some external support, and we help the help that external support whoever the year-on-year average transaction price stayed in adjustment, offering low-price twin. Against that backdrop, our Q2 existing home GDV rose 85% from a year earlier, transaction volume climbed almost 25% year over year, well ahead of the market. The more direct offer source was higher unit productivity in our connected store network. In Q2, connected store transaction volume increased almost 30% year over year, network scale did not grow, active stores and agents stayed broadly flat versus a year ago, yet average transaction per active connected store climbed 26%. That demonstrates our network is moving from expansion toward high-quality operation. As earlier connected stores mature and platform collaboration deepens, that network value turns directly into greater per store output and high efficiency. The second offer was better conversion, or platform service value, into revenue. Q2 < > platform service revenue rose 27.8% from a year earlier < >; in a buyer's market, professional marketing, property presentation, and transaction facilitation delivered clear value and are chosen more and more by homeowners. Meanwhile, the existing home contribution margin increased 6.1 percentage points year over year to 46.1%, confirming growth wasn't post expense or profitability. Looking ahead, we will track whether connected store output and the conversion of platform service revenue stay stable across different markets. And going forward, we will focus more on the output of the connected store and also whether the conversion remains stable across different markets to validate the sustainability of this offer.
Thank you, Mr. Tao. Our next question comes from Alvin from CMS.
The performance in the new home segment is quite strong too. What's behind the Q2 new home alpha, given that the shift in operations—moving from traditional channel collaboration toward integrated marketing plus project services capabilities—keeps generating value over time; and along the way, how are you managing the trade-offs among growth margins, connection prediction margin collection cycles, and developed credit risk?
Thanks, < >. Good evening. Through the first six months of this year, conditions in the new-build market have stayed under pressure. Even so, during the second quarter, things got better: the pace at which year-over-year sales fell for the leading 100 developers eased, with demand shrinkage reaching 9.3%, while newly added supply became more heavily focused on prime urban locations, premium developments, and offerings aimed at buyers looking to trade up. Given those conditions, our gross transaction value for new homes during Q2 rose 1.2% compared with the same period a year earlier, largely thanks to broader access to top-tier projects and better rates of converting interest into sales. To start, we got in early with top-tier developers and worked alongside them on their freshly introduced initiatives, which let us broaden our reach and boost results across the market's leading projects. Next, we sharpened how we spot needs, qualify opportunities, and run the pipeline; by looking at conversion rates, we directed our resources toward the opportunities with the strongest potential. Looking ahead to the back half of the year, our expectation is that conditions stay in a correction phase, with buyers holding back, so we're concentrating on fine-tuning the mix of projects and lifting conversion rates to strengthen operating efficiency that we can control over time. Our new home segment isn't simply about tacking on another service—it's meant to help customers work through their housing choices. In a buyer's market, people grapple with complicated options and many alternatives, so what they require goes well beyond simply gaining access to the project. They have to grasp how scalable this project is, what the product is worth, and how it stacks up against what's available nearby — price payouts, amenities, and whether it can actually satisfy their requirements. Beyond that, we're shifting away from being just a transaction channel toward a customer-centric, full-cycle project services model. Our goal is to put consumers at the center, deliver services across the entire cycle, and weave consumer insight into project research, positioning, sales, and the decisions that support their buyers. Upgrading in this way is driven by consumer value, while developer value comes from serving those consumers more effectively. Along these lines, we're developing 3 capabilities. To start, we draw on prior consumer research and matching efforts. By leveraging information from home transactions that already exist—things like and veins—we gain insight into demand project positioning and marketing, which cuts down on the gap between what developers build and what buyers actually want. Next, we convert what a product offers into decision metrics that can be compared. Complicated elements such as location, layout, natural light, and amenities get turned into content that's easy to grasp; on top of that, we provide explanations along with additional services that support decision making. Take the deliver make levels in Guangzhou—there we offer 3D community presentations and layout analysis, letting consumers understand the product in an intuitive way and boosting on-site conversion. Finally, we bring end-to-end project operating capabilities grounded in customer feedback. At this point, we connect the analysis of customers, the content, and the channel sales for a given project, and we also factor in timing adjustments along with how resources get allocated. Take a project in < >, for instance: the developer assisted in acquiring knowledge of the local market, and we went back and reexamined the target consumers. We fine-tuned based on what the market told us, we tweaked the sales approach, and we tied channel acquisition to on-site conversion, which lifted sales efficiency. < > capacities are still at the stage of early validation. We'll adapt them on a project-by-project basis, confirming consumer value, operating results, and economics prior to scaling across every one of those projects; and my view is that validation needs to be sustainable, which lets us replicate things better. As our services grow and the scope of what we offer gets deeper, we'll handle payment terms and development credit risk with even greater prudence, steering clear of irrational risks, purely for the sake of growing GTV. Thus over the long run, expansion will rest on a deeper grasp of consumers and aggressive matching, which in the end turns into revenue of high quality, healthy profitability, and robust cash collection, and that gives us high-quality growth. Thank you.
Thank you. Mr. Xu. The next question comes from Griffin from Citi.
I'd like to ask about home renovation and the care fee. Our Q home renovation revenue fell more quickly compared with a year earlier, yet contribution margins got much better. What caused that drop, and are the changes we made earlier basically finished? When should revenue come back? Also, how do you weigh scale, contribution margin, and delivery quality? Does Carefree Rent's profitability, or margin, rise considerably? And how do we keep that going?
Thanks, Griffin, for asking. A deep supply-demand restructuring is what this industry is going through. Because property adjustments flow into renovation, deliveries of new homes have fallen. Businesses that previously focused on newly built residences are now aggressively entering the resale housing sector, which shapes how competitive the landscape has become. Under circumstances such as these, navigating through the cycle depends on operational excellence, how competitive products are, and delivering quality at scale. In the second quarter, the decline stems from two things. First, over the past year we made a proactive choice to exit cities, locations, and acquisition avenues that were not efficient. Second, overall demand remains squeezed since fewer newly built residences are being handed over, and that weighed directly on the renovation of homes business, while competitors are using price vehicles and hefty channel incentives to fight for our resale home clients. This intentional realignment is essentially now complete. We anticipate no additional broad-based contractions in the current year. Even with revenue under pressure, contribution margins rose considerably. Centralized procurement along with supply chain optimization brought material costs down meaningfully. < > provider productivity per store likewise rose versus a year earlier, and store costs were optimized too, pointing to healthier retained capacity and cost structure. As for revenue recovery, so contract value acts as a leading indicator. Although reported revenue trails because of construction cycles, positively, end metrics such as July showroom visits rose quarter-over-quarter thanks to restored internal cooperation and incentives, though translating that into revenue will take time. Looking ahead, we won't sacrifice profitability for scale. Over the long haul, our growth hinges on how well we deliver, achieved through consistent actions, plus enhancements to the user experience and the competitiveness of our products. Those improvements will arrive through customized renovation offerings and integrated showroom spaces located at transaction centers — we're aiming for high-quality products and a healthy business, built on three pillars: a growth strategy set to fuel substantial expansion in both revenue and profit. Turning to Carefree Rent, the number of units we manage has been climbing steadily, reaching just under RMB 790,000, which is a 34% increase compared with a year earlier. Revenue totaled approximately RMB 4.83 billion, with a contribution margin of 15.3%, representing a 6.9 percentage point gain year-over-year. The drop in revenue compared with last year mirrors the ongoing evolution of payer free strength toward a product with lighter, net-based revenue. Improved margins came from our structural repositioning along with broad operational efficiencies across labor, installation, and poster leasing expenses. As for whether that level of profitability can be maintained, my view is that simply adding more units won't be enough — what's needed is stewardship of an asset base that sees reduced churn-driven release activity and more renewals. Accordingly, during this week, expenses for labor and channel will increase at a slower pace than revenue itself. Going forward, we'll concentrate on three main priorities. The first is stabilizing the portfolio of units under our management, which should bring down the costs associated with the re-leasing channel. As our managed unit count expands, an increasing portion of those units are entering the renewal stage. Homes already in our portfolio are being listed for release. By managing leases in a proactive way and providing high-quality service, we will lift both renewal and retention. During Q2, the renewal rate for owners reached 74%, a gain of 4 percentage points, while the tenants renewal rate came to 56%, which is 1 percentage point higher than the same period a year earlier. Second, we want to raise efficiency so the delivery cost for each unit falls. In Q2, management units per asset manager climbed 40% compared with a year prior, reaching roughly 170. As we go forward, we'll try separating work tied to transactions—things like sourcing and leaving—from the management tests, which include renewal and post lease. That should sharpen specialization and boost output per person. AI also has a part to play. We can bring in AI planning to deal with scale complexity by refining service areas and matching, scheduling of tasks, plus many other operational improvements. Third, raising the quality of incremental scale will require adding more asset-light offerings that can withstand swings in rental rates. On top of that, in each different city, we plan to use differentiated product solutions, which will generate more robust unit economics. Most importantly, service quality is the bedrock beneath all of these gains. So whenever a tenant or an owner makes the choice to renew, I mean, < > depends on reputation, repurchase, and the channel cost as well; therefore we will devote particular care to reputation and to bringing channel costs down. Thus we hold that profitability can only last when service experience, renewal, and efficiency together create a positive cycle. That is why we are strengthening this base so that growth in our scale turns straight into growth in profit. Thank you.
Appreciate that, Mr. Xu. That wraps up the question-and-answer portion. Thanks again for being with us today. Should anything else come to mind, you're welcome to reach out to Baker's IR team using the details listed on our website. That brings today's call to a close, and we hope to talk with you on the next one. Thanks, and farewell.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete KE Holdings Inc. transcript - plus 131,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.
Get an API key View API docs →For developers and AI pipelines
Programmatic access to KE Holdings Inc. earnings transcripts and 131,000+ others is available through the
EarningsAPI REST API and the hosted MCP server.
Quarterly plans from $162 - full transcripts, speaker segments, full-text search,
and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.