BEKE (Trans): Net margin at 3-yr high, revenue still contracting

Dolphin Research Trans of$KE(BEKE.US) FY26Q2 earnings call

I. Key takeaways

1. Shareholder returns: stepped-up buybacks, first repurchases in HK

a. Q2 buybacks totaled approx. $250 mn. This included the company's first purchases in the HK market.

b. 1H buybacks reached approx. $450 mn, up ~14% YoY. That equals about 2.4% of shares outstanding at end-2025.

c. Since the program launched in Sep-2022 through Q2 FY26, cumulative buybacks were approx. $2.99 bn. This equals ~14.8% of pre-program shares outstanding.

2. This quarter's KPIs: profit growth far outpaced GTV and revenue

a. Aggregates: total GTV grew 6.3% YoY, while revenue fell 5.7% YoY due to intentional trimming in home renovation and recognition shifts from rental product model upgrades. Group GPM expanded 670 bps YoY to 28.6%, and GAAP Opex declined 14.1% YoY.

b. Profit: GAAP OP was 3.026 bn, +185.6% YoY and +137.8% QoQ, with OPM of 12.3% (+8.3 pp YoY, +5.6 pp QoQ). Non-GAAP OP was 3.59 bn, +123.6% YoY and +115.7% QoQ, with OPM of 14.6% (+8.5 pp YoY, +5.8 pp QoQ).

c. Net profit: GAAP NP was 2.624 bn, +100.8% YoY and +109.1% QoQ. Non-GAAP NP was 3.185 bn, +74.9% YoY and +97.6% QoQ, with a non-GAAP NPM of 13% (+600 bps YoY), a three-year high.

d. Segments: contribution margins rose YoY and QoQ across all five lines — existing homes 46.1% (+6.1 pp YoY, +4.8 pp QoQ), new homes 28.8% (+4.4 pp YoY, +3.1 pp QoQ), renovation & furnishing 39.6% (+7.5 pp YoY, +3.4 pp QoQ), rentals 15.3% (+6.9 pp YoY, +0.5 pp QoQ).

e. Expenses and one-offs: GAAP Opex totaled 3.989 bn, down 14.1% YoY but up 21.3% QoQ on seasonal ad spend for renovation and new-home bad-debt provisioning. Store-related costs were 560 mn, down 25.9% YoY and flat QoQ; G&A was 2.04 bn, down 2.1% YoY but up 18.9% QoQ, mainly due to an approx. 280 mn full bad-debt provision after prudent assessment of a developer receivable and collateral; S&M was 1.4 bn, down 26.1% YoY but up 29.6% QoQ on the renovation peak season; R&D was 550 mn, down 13.4% YoY and up 11.4% QoQ on higher tech service fees.

3. Cash flow and B/S: receivable days shortened meaningfully

a. Q2 operating cash inflow was 6.61 bn. This reflects solid collections and disciplined working capital.

b. New-home net A/R days were ~39 days, about 12 days shorter YoY. This shows improved risk control.

c. Excluding customer deposits, broad cash was ~67.3 bn at end-Q2. Liquidity remains ample.

4. 2H outlook and capital allocation principles

a. Do not extrapolate a single quarter: under a neutral market, a lower cost baseline should continue to support profits. However, channel incentives and some frontline sales costs will fluctuate QoQ with revenue scale, mix and seasonality.

b. Aim for balanced growth in revenue and profit. When the market improves, a lower base unlocks stronger operating leverage and greater profit elasticity; if pressure persists, a healthier mix reduces profit sensitivity to volatility.

c. Enforce strict ROI discipline for all core new biz and tech investments, focusing on customer value, operating efficiency and sustainable returns. Even in a recovery, there will be no return to blunt expansion.

II. Detail from the call

2.1 Executive remarks

1. Existing homes: scale returned to growth; alpha from per-store productivity, not footprint expansion

a. Q2 GTV reached 629.89 bn, +8% YoY and +17.9% QoQ. Revenue was 7.02 bn, +4.5% YoY and +14.5% QoQ.

b. GTV grew faster than revenue as non-Lianjia GTV with net revenue recognition mix increased. Non-Lianjia platform service revenue grew 27.8% YoY and 29.8% QoQ this quarter.

c. The store network held steady, with per-store output lifted by refined ops. Connected stores outperformed the market.

d. CM improved YoY on lower fixed people costs and a structural shift toward higher-margin platform service revenue. QoQ gains reflected operating leverage, recovery in net-method revenue and further mix optimization.

2. New homes: stable scale, driven by quality project coverage and conversion

a. Q2 GTV reached 258.39 bn, +1.2% YoY and +77.1% QoQ. Revenue was 8.95 bn, +3.8% YoY and +75.9% QoQ.

b. Under market pressure, scale held steady by partnering with quality projects, improving client conversion and optimizing costs. Execution quality was the key differentiator.

c. CM rose on cost structure optimization from refined ops, with QoQ upside also from operating leverage on revenue growth.

3. Renovation & furnishing: proactive downsizing largely done

a. Q2 revenue was 3.19 bn, -30.1% YoY and +36.4% QoQ. Seasonality aided the QoQ rebound.

b. The YoY decline reflects exit of inefficient channels and cities with poor unit economics, while a weak new-home market curbed renovation demand. The QoQ recovery was seasonal.

c. CM gains mainly came from centralized procurement lowering material costs and ongoing cost discipline. Efficiency improved across stores.

4. Rentals (Hosted Rent): revenue recognition down, managed units growing fast

a. Q2 revenue was 4.83 bn, -14.8% YoY and -3.6% QoQ, due to a shift to lighter, lower-risk models with net revenue recognition. Financial revenue fell, but managed inventory kept rising.

b. Managed units exceeded 790,000 by end-Q2, up ~34% YoY, with net-method products above 50% of the mix. Scale is compounding.

c. CM improved on product mix upgrades and better ops in labor, fit-out and post-rent costs. QoQ, higher net-method mix also helped.

5. Emerging & other: Q2 revenue was 550 mn, +26.4% YoY and +7% QoQ. Growth remained broad-based.

6. Three operating changes from the transformation

a. Refined operations: move from one-size-fits-all to zone- and project-level strategies. Instead of city-level metrics, drill to micro-markets to find solutions; for example, when premium buyers view cross-district, the old geo-boundary model breaks, so units are reorganized along actual viewing paths with specialist explainers and client experts.

b. In one city, 600 projects drove half of transactions, and standardizing this professional judgment into clear division of work enables replication to other cities. The playbook becomes portable.

c. Metrics shift: scale and share still matter, but we now put more weight on agent consistency, per-capita efficiency and income, store-level profitability, and stable service quality. Health over sheer size.

d. Rentals exemplifies this: in 2025 peak, rental agents hit 700 and per-capita productivity fell below two deals. Rather than add headcount, the org was split into smaller units and rematched listings, clients and agents by familiarity and capability; from Apr to Jul, per-capita deals rose from 3.0 to 5.6 and the zero-deal ratio fell from ~25% to below 10%, proving organization beats headcount.

e. Mobilize managers: managers left the meeting room for the front line, shadowing listings, reviving cold leads, and accompanying agents to signing centers this quarter. The only ask of managers is to be on site — you cannot learn to swim without getting into the water.

7. Consumer-centric does not mean bypassing agents: from splitting commission to creating new roles

a. Slicing a single deal commission is zero-sum. To break that, we must create more high-value tasks, not just re-cut the same pie.

b. Consumer needs are changing: 'good' used to mean the property itself, now it means the right match. The deciding variables expanded from one to three — property, family situation, and service provider; the service provider is now a key variable, not just a conduit.

c. Harder decisions force task specialization for three reasons: required knowledge already exceeds one person's capacity, professional depth requires mutually exclusive paths, and the most valuable work shifts from offering options to confidently ruling out options.

d. But screening is not closing. As long as income relies entirely on closing, true professionalism will not scale, so we delink this role's income from transaction completion and align it fully to buyer- or seller-side needs — this is the client manager role.

e. Previously, platform service stopped when the lead reached the agent. Client managers maintain continuity: AI organizes data, humans handle stages and needs, and agents receive well-served clients; because they are not paid per deal, client managers stay objective.

f. From May to Jul, client managers handled over 50,000 leads, achieving a lead-to-viewing conversion of 7.4%. This is above the market's ~5%.

g. The platform is evolving from commission-splitting to a structure where each professional capability is independently validated and compensated. Biz is moving from single-listing flows to a modular ecosystem — consulting, viewings, signing, reporting, marketing materials, renovation, rentals — where anyone creating incremental value is a service provider.

8. Will AI replace agents: what depreciates, what becomes scarce

a. The premise is that agents only sell static info, which AI can easily fetch — layout, size, price, vintage, floorplan — and that is depreciating fast and insufficient for decisions. Pure info hauling will have no future.

b. Scarcity lies in dynamic, deep, non-fabricable insight: seller motives, renovation potential, senior managers' local assessments, community realities, how the last deal was negotiated. These live in people's heads and lack industry pipes to be codified and reused.

c. Fundamentally, AI bears no liability for wrong decisions. As the cost of mistakes rises, consumers value uncertainty reduction more.

d. Three things will happen: value shifts to digesting uncertainty, value creation gets harder and needs deeper data and research, and those who pivot in this direction — platforms and managers — become more valuable. The industry needs professionals who judge and own outcomes, not info couriers.

9. AI in-house: treat AI as a direct variable, not just an efficiency tool

a. The biz is a production function with inputs of human capital, labor and tech. The core question is whether AI is a support tool or a direct variable that requires org redesign.

b. First, change the mindset: we opened basic capabilities and lowered usage thresholds. The aim is to make AI a new factor of production, not an opponent.

c. Change management: for ~200 years, management and finance advanced on quant data, while the unquantifiable was ignored. AI turns unstructured data and language into usable signals, shifting granularity from managing averages to managing each home, each client and each agent.

d. Change division of work: as compute and knowledge barriers fall, old divisions fade and new ones appear. Lianjia's new-home biz is re-split between humans and machines — AI generates and compares plans from a dynamic knowledge base, while agents focus on understanding clients and fine-tuning judgment, producing both transactions and organizational learning.

e. Org impact has four aspects and ultimately comes back to people: on cost, AI lowers fixed costs and raises variable costs, rewarding fast iterators; on trial-and-error, AI makes innovation high-frequency and low-cost, enabling parallel tests and higher win rates; on frontline vs. middle office, the frontline can rapidly build and test solutions while the middle office scales them; on managers, AI takes over report-chasing, forcing managers to stop relaying and start creating real business value.

f. The bottleneck returns to people: BEKE's value chain is long, and while AI can optimize many processes, human-to-human interaction remains the bottleneck and cannot be replaced. The key is whether we can rally people, train them and enable efficient human–AI collaboration — a culture and evolution challenge; AI is therefore a direct variable, reshaping clients served, judgment, processes and the org. We are not fitting AI into the old system; we are regrowing the company with AI.

10. Next-phase big bets and validation criteria

a. Split issues into two types: directional topics that deserve big bets, and form-factor topics where we must search for answers.

b. Three big bets: deep service, deep data, and platform ecosystem. The more information is democratized, the scarcer deep data becomes; the harder the decision, the more valuable deep service is; the finer the division of labor, the more orchestration a platform must do.

c. Still searching: AI's end-state form and the ultimate line between management and professional specialization remain uncertain. For form-factor questions, invest small, test fast, and cut losses early.

d. We have validated over the past two quarters that continued spending in low marginal-return areas makes no sense. Pure scale-driven models no longer work, and such spending should stop.

e. Four things to validate: whether professionals can use AI to create value directly or indirectly and whether we redefine professionalism and commit to it; whether managers can return to the front line, make high-quality judgments and rebuild their value; whether deeper service can win client trust and broader recognition for processes and judgments; and whether the org can turn point successes into replicable capabilities.

f. In a discontinuous transformation, conviction is the leading indicator and numbers lag. Many managers' professional strengths remain personal and cannot scale without openness and sharing; the core test is whether we can persistently execute consumer-centricity and let professionalism win, embedding it into culture and process, measured along four dimensions: client, service provider, operations and replicability — Q2 is a start, not a conclusion.

2.2 Q&A

Q: With volume–price divergence in Q2 property markets and choppy momentum in Q3, what levers can you control in Q3 and for the full year?

A: In 1H, existing-home markets showed a structural recovery in volumes and a bottoming in prices, with divergence across cities and price bands more evident in Q2. Tier-1 volumes recovered faster, and Tier-1 prices were more resilient QoQ; Q2 Tier-1 existing-home online signings outgrew other cities, and per BEKE Research Institute, Tier-1 existing-home prices rose a cumulative 3.6% QoQ in 1H while nationwide prices were roughly flat YoY.

Platform data show low-price segments grew faster than mid-to-high, but the mix by floor area stayed stable. This indicates demand did not broadly downgrade to small units, and price adjustments simply shifted the transacted price band lower; meanwhile, higher-priced homes saw smaller YoY price declines, showing resilience in core upgrade and quality segments.

New-home sales remained under pressure in Q2 overall, with strong, well-positioned projects in core cities holding up better. Structurally, existing homes accounted for over 80% of national residential transaction area in 1H, now the primary bearer of housing demand.

Our view: volumes are recovering structurally while prices continue to base, with core cities and quality supply more resilient but with ongoing divergence. With more choices, clients decide more cautiously and prize professional judgment and certainty — they need decision support, not just brokering, which maps to our accumulated service capabilities.

Given this, we will focus on three things in 2H. First, capture structural opportunities to drive revenue recovery by allocating resources by city, client segment and property type, strengthening Tier-1 coverage, and converting real demand through content, client engagement, precise matching and professional execution. Second, reinforce financial discipline and elastic resource allocation with a leaner cost base to weather volatility; if pressure persists, dynamically redeploy resources and prioritize the core professional network over near-term profit, and even if the market improves, require ROI and service validation before scaling to ensure efficient conversion of revenue to profit and cash flow. Third, prioritize cash flow and the balance sheet through strict A/R and collections management, risk control and limiting non-essential spend.

We will not rely on the macro tape in 2H. On revenue, better decision support should win more clients; financially, a healthier cost structure protects cash and core capability in a weak market and releases more operating leverage if conditions improve.

Q: Q2 profit growth far exceeded revenue growth. Can you break down contributions from segment performance, operating efficiency and the expense baseline? Any one-offs, and how sustainable are these improvements?

A: Q2 profit improvement came mainly from higher CM in core businesses and lower Opex. Core CMs rose YoY and QoQ, lifting group GPM by 670 bps YoY to 28.6%, while GAAP Opex fell 14.1% YoY.

There were three drivers. First, the expense baseline moved lower: over the past year we optimized Lianjia's org and agent structure, trimmed G&A, consolidated resources and cut low-output spend, reducing fixed people costs and the breakeven point — this baseline effect is durable. Second, operating efficiency improved: in new homes, better coverage of quality projects and higher client conversion bolstered resilience, with stable monetization plus channel efficiency driving profit; in existing homes, focus on quality listings and refined support for connected stores markedly lifted their revenue and profit contribution, improving unit economics. Third, mix improved: in renovation, centralized procurement and cost control reduced material cost ratio; in rentals, CM improved as products shifted to net-method models, alongside better labor, fit-out and post-rent cost control.

For the next two quarters under a neutral market, the lower cost base should keep supporting profits, while channel incentives and some frontline sales costs may fluctuate QoQ with revenue scale, mix and seasonality. We will not simply extrapolate a single-quarter profit, but pursue balanced growth in revenue and profit — upside is larger in a recovery as incremental revenue layers on a lower base to amplify leverage, while in a tough market, a healthier structure reduces profit sensitivity; today we have both more upside and better downside protection.

Longer term, this optimization lays the foundation for healthy ops and is step one of a strategic transition — optimizing resource allocation and digesting uncertainty for the current market. Step two is to deploy scarce resources to what truly creates client value, not just cost cuts, and ultimately codify efficient allocation via processes, evaluation, incentives and platform tools to support sustainable growth.

Q: Existing-home GTV rose 8% YoY in Q2 and CM improved 6.1 pp — how much from market recovery vs. company execution, and what metrics evidence operating improvement?

A: In short, the market recovery provided volume support, but existing-home alpha did not come from adding outlets or raising prices. It came mainly from sustained per-store productivity within the network and better revenue capture of platform service value — the concurrent margin expansion shows we did not trade profit for growth.

Specifically, in Q2, existing-home volumes in key cities recovered modestly and prices stabilized QoQ, offering some external support. But average prices still adjusted YoY and low-price mix rose; in this context, Q2 existing-home GTV grew 8% YoY (the transcript misprinted 85%, inconsistent with the disclosed 8%), and transaction count rose nearly 25% YoY, materially outperforming the market.

The more direct alpha came from higher per-store productivity in the connected network. Q2 connected-store transactions rose nearly 30% YoY without network expansion — active stores and agents were roughly flat YoY, but average deals per active connected store rose 26%, showing a pivot from expansion to high-quality operations.

The second source was better conversion of platform service value into revenue, with non-Lianjia platform service revenue up 27.8% YoY. In a buyer's market, professional marketing, listing presentation and deal facilitation are delivering clear value and are increasingly chosen by sellers.

At the same time, existing-home CM rose 6.1 pp YoY to 46.1%, confirming growth did not come at the expense of cost or profitability. We will track per-store output in connected stores and the stability of platform service revenue conversion across markets to validate the sustainability of this alpha.

Q: Where did Q2 new-home alpha come from, and can upgrading from traditional channel distribution to integrated marketing and project services create sustainable value? How do you balance GPM/CM, cash collection and developer credit risk?

A: The new-home market stayed under pressure in 1H, though Q2 saw some improvement as top-100 developers' sales decline narrowed to 9.3% YoY, with demand and new supply further concentrating in core cities, quality projects and upgrade products. Against this backdrop, Q2 new-home GTV grew 1.2% YoY, driven by better coverage of quality projects and conversion — earlier identification and cooperation with quality and new-to-market projects lifted coverage and performance on leading projects, and resources were allocated by conversion prospects to high-potential projects.

Assuming continued adjustment and cautious clients in 2H, we will focus on optimizing project mix and conversion to improve controllable operating efficiency. Longer term, the new-home goal is to solve the customer's housing decision, not just extend the service chain; in a buyer's market with complex choices, consumers need to understand expandability and product value, compare pricing, payment terms and amenities versus alternatives, and assess whether their needs are met.

We are shifting from a transaction channel to client-centric, full-cycle project service that embeds consumer insight into research, positioning, sales and decision support. Consumer value drives this upgrade, and developer value comes from better consumer service.

We are building three capabilities. First, earlier consumer insight and matching using existing-home transaction data to inform positioning and marketing, reducing mismatch between products and real demand. Second, translating product value into comparable decision metrics and intuitive content — e.g., 3D community renderings and unit analysis in Guangzhou helped consumers grasp products and improved on-site conversion. Third, end-to-end project ops tied to client feedback by linking client analytics, content and channel sales, and adjusting pacing and resource allocation — e.g., for a project where the developer lacked local knowledge, we re-segmented target clients, tuned sales strategy to market feedback, and integrated channel acquisition with on-site conversion to boost efficiency.

These capabilities are still in early validation and will be customized by project, scaling only after proving consumer value, operating results and economics. As scope expands, we will manage payment terms and developer credit risk prudently to avoid taking unreasonable risk just to grow GTV; long-term growth will rest on deeper consumer understanding and precise matching, translating into quality revenue, healthy profitability and strong collections.

Q: Renovation revenue declined faster YoY in Q2 while CM improved sharply. What drove the decline, is the adjustment largely done, and when will revenue recover? How do you balance scale, CM and delivery quality?

A: The industry is undergoing a deep supply–demand reset: property downturns flow through to renovation, new-home deliveries fell, and players focused on new-home renovation flooded into existing-home renovation, intensifying competition. Under such conditions, cycle-resilience depends on operating quality, product competitiveness and scalable delivery quality.

Q2 decline reflected two factors: over the past year we exited inefficient cities, stores and channels, and broader demand stayed weak as fewer new-home deliveries directly suppressed renovation. Competitors also engaged in price wars and high channel incentives to fight for existing-home clients; the proactive adjustment is now largely done, and we do not expect large-scale further contraction this year.

Despite revenue pressure, CM improved materially as centralized procurement and supply-chain optimization lowered material costs, store productivity per service provider rose YoY, and store costs were optimized. This shows the remaining capacity and cost structure are healthier.

On recovery timing, contracted value is the lead indicator — P&L revenue lags due to construction cycles. Early signals are positive: store foot traffic improved MoM in Jul on restored internal collaboration and incentives, but it will take time to flow into revenue; we will not trade profit for scale, and long-term growth will rely on delivery quality from high-frequency actions, a better user experience and product competitiveness via customized packages and integrated showrooms in transaction centers, aiming for quality products and healthy revenue and profit growth.

Q: Hosted Rent profitability and margins improved significantly. How will you sustain this?

A: Managed units grew steadily to nearly 790,000, +34% YoY; revenue was ~4.83 bn with CM of 15.3%, up 6.9 pp YoY. Revenue declined YoY due to iteration toward lighter net-method products; profitability improved from this mix shift plus operational optimization in labor, fit-out and post-rent costs.

Sustaining profitability is less about taking more units and more about managing a low-churn, low-vacancy, high-renewal asset pool so that growth in labor and channel costs trails actual revenue growth. This is the operating core.

We will focus on three areas. First, stabilize the managed-asset mix to lower relet channel costs; as scale expands, more units enter renewal cycles and more existing units face relet, so we will proactively manage leases and service quality to lift renewal and retention — in Q2, landlord renewal was 74% (+4 pp YoY) and tenant renewal 56% (+1 pp YoY). Second, improve efficiency to cut per-unit delivery costs: units per asset manager rose ~40% YoY to ~170 in Q2, and we will pilot splitting transaction tasks (sourcing, letting) from management tasks (renewal, post-rent) to raise specialization and per-capita efficiency, with AI assisting via service-radius optimization, matching and task dispatch to manage complexity at scale. Third, raise the quality of incremental growth by increasing light-asset product mix to buffer rent volatility and tailoring products by city to achieve healthier unit economics.

Service quality underpins all of the above — landlord and tenant renewal ultimately drive word of mouth, repeat and channel costs. We will focus on reputation and suppress channel costs, and only when experience, renewal and efficiency form a flywheel will profitability be sustainable; we are laying this foundation to convert scale growth directly into profit growth.

<End of text>

Risk disclosure and statement:Dolphin Research disclaimer and general disclosure