--- title: "Reddit (Trans): Google traffic dries up faster; betting on 'login-only'" type: "Topics" locale: "en" url: "https://longbridge.com/en/dolphin/post/44645289.md" description: "Q3 revenue grew 53%, with TTM revenue exceeding $3.0bn. For Q4, guidance is $1.00–$1.03bn (+38%–42% YoY).However, search referral traffic is deteriorating faster, and Q4 DAU (QoQ) could decline by low-teens from the 130 mn base. The company will also test a 'login-only' mode outside the U.S. in Q4." datetime: "2026-10-29T23:31:04.000Z" locales: - [en](https://longbridge.com/en/dolphin/post/44645289.md) - [zh-CN](https://longbridge.com/zh-CN/dolphin/post/44645289.md) - [zh-HK](https://longbridge.com/zh-HK/dolphin/post/44645289.md) author: "[Dolphin Research](https://longbridge.com/en/dolphin.md)" generator: "portal-rs" --- # Reddit (Trans): Google traffic dries up faster; betting on 'login-only' **I. Earnings Recap: Key Takeaways** 1\. Shareholder returns and dilution a. Buybacks: Repurchased approx. $100 mn of stock in Q3 at an Avg. price just under $150 per share. Remaining authorization stands at roughly $660 mn. b. Dilution: Fully diluted shares fell QoQ to 206.3 mn. YTD total dilution of 1.2% is within the long-term 1%–3% annual guide, and after buybacks, YTD fully diluted share count is up just 0.1%. c. SBC and related taxes were $130 mn, or 15% of revenue. The ratio fell 80 bps YoY. 2\. Q4 outlook a. Revenue of $1.00 bn–$1.03 bn (+38%–42% YoY). Midpoint implies ~40%. b. Adj. EBITDA of $480 mn–$510 mn (+47%–56% YoY). Midpoint implies ~49% Adj. EBITDA margin. c. The next internal 'North Star' is $5 bn in annual revenue. Management reiterated this long-term target. 3\. Margins and cost structure a. Cash and profitability: FCF was $363 mn (+98% YoY) with a 41% margin. Cash and investments reached $3.03 bn, topping $3 bn for the first time. YTD Adj. EBITDA exceeded $1 bn, nearly double YoY. b. Margins: Net margin was 32% and Adj. EBITDA margin 45%. Incremental Adj. EBITDA margin was 54%; GPM was 91.5% (+50 bps YoY), marking the ninth straight quarter above 90%. c. COGS was ~ $76 mn (+44% YoY), invested in ML, site performance, measurement, and safety. Gains were more than offset by incremental revenue and infra efficiency programs. d. Opex: Adj. total costs and expenses were $493 mn (+41% YoY), while revenue grew ~1.3x faster. Adj. operating expenses were $417 mn, at 47% of revenue (vs. 51% a year ago). e. Headcount and marketing: Headcount rose ~3% QoQ and 12% YoY. G&A headcount was flat vs. Q3 2024; user marketing spend remained a mid-single-digit % of revenue, flat vs. Q2. 4\. Tax and disclosure a. Q3 tax expense was approx. $5.8 mn, for an effective tax rate of ~2%. The company still maintains a full valuation allowance against U.S. federal and state deferred tax assets and is updating internal forecasts to assess whether it remains appropriate. b. Capex was $1.2 mn. Capital intensity stayed very low. c. User metrics: Will continue to disclose logged-in/logged-out DAUq by region. The team will keep evaluating whether user KPIs match how the biz is managed. **II. Earnings Call Details** **2.1 Executive commentary: key points** 1\. Users and traffic: search referrals under pressure a. Search referral traffic fell notably in Q3 and accelerated lower into quarter-end, continuing into early Q4. The impact was greatest on logged-out and web traffic, with some spillover to logged-in and App users, as search is a common re-engagement source. b. In Oct., search referrals declined by over 1 mn on Avg. per week. If that pace holds, total DAUq in Q4 could fall by a low double-digit % vs. the current 130 mn base. c. Logged-in and App DAUq grew QoQ. These cohorts drive the vast majority of engagement and over 80% of monetization, and remain the top priority. 2\. Product and retention a. ML Feed and App performance upgrades are working: first-week return rate for new installs rose double digits QoQ and over 50% YoY. Video creation and consumption are up; next, the team will test a more consistent, predictable core Feed, plus new experiences in games and visual podcasts. b. Community: 1 bn new posts and comments were added in Q3. The company will make joining subreddits more ceremonial and provide mods with AI tools; median enforcement time by the in-house trust & safety team fell from 58 hours to 5 seconds this year. c. New user channels: word-of-mouth installs, reinstalls, and mobile web-to-App conversion. These are elaborated in the Q&A. 3\. Moving to 'logged-in only': broad tests start in select ex-U.S. markets in Q4. Results will inform the global rollout. Rationale and test focus are detailed in the Q&A. 4\. Ads biz a. Volume and pricing: Ad revenue grew 54% YoY, with pricing outpacing impression growth. Clicks rose over 50% YoY and conversions more than tripled. b. Channels and regions: Revenue from scaled channels grew over 80% YoY, and active advertisers in the mid-market rose over 50%. U.S. ad revenue grew 51% while Intl rose 68%; finance, pharma, and travel were strong; shopping and App install objectives each grew over 95%. c. Automation: Reddit Max entered open beta in Q3, with advertiser adoption more than tripling QoQ and a new asset-group feature added. Model training and real-time sequence/session/graph signals were upgraded, lifting full-funnel performance. d. Full-funnel: Video views are the fastest-growing upper-funnel objective, with revenue more than doubling YoY, and a 15s interactive video view objective in open beta. Shopping is the fastest-growing lower-funnel objective, with catalog support expanded to 20 mn SKUs. e. B2B: Over 20 mn business decision-makers use Reddit globally, and 9 in 10 use it to research business products. B2B advertisers leveraging in-market audiences saw cost-per-action improve by ~7% vs. other first-party targeting. f. Ecosystem: Ad API ecosystem revenue rose over 45% QoQ. New partners include Innovid and Skai; beta integration began with HubSpot, launching the first third-party Max management integration. g. Shopify integration has scaled since the May GA launch. A luggage brand case showed 4x+ ROAS and a 40% improvement in CPA efficiency. h. Measurement: Third-party studies show Reddit users have 17%, 19%, and 11% higher LTV in personal care/beauty, QSR/beverages/snacks, and retail beauty, respectively. Meta-analysis of 25 TV lift studies showed an Avg. 18% incremental lift, with 15% longer watch time per household than peers’ benchmarks. 5\. User marketing a. The 'People are the Best' brand campaign is gaining traction, lifting relevance and ad recall. The company held its first community pop-up in NYC, covering 75 communities. b. U.S.: Some budget shifted from web acquisition to App installs and brand. U.S. web-acquired users retain poorly with weaker ROI, while App-installs retain on par with organic users. c. Intl: With lower brand awareness, web acquisition spend was increased, partly offsetting the decline in Intl search referrals. Over the next few quarters, the team will evaluate performance and explore reactivation targeting logged-in and App users. **2.2 Q&A** **Q: Why move to 'logged-in only'? What are you testing this quarter, and how will it affect top-of-funnel and new user acquisition?** A: Other platforms have already made the shift, and Reddit’s timing is right; tests focus on where to place the login wall and when to trigger it. Reddit operated successfully with logged-out access for years, but at today’s scale, the product works best when logged in. There are four reasons: only when logged in can we personalize communities and content; ad targeting signals are stronger; most data scraping occurs among logged-out users, so gating protects the corpus; and it is easier to adapt to new regulations. The trade-off likely favors engagement over raw reach, but 80%+ of monetization already comes from the App and logged-in users. Tests will preserve the 'taste test' on Reddit.com and enable link/post sharing. The team will study conversion rates across surfaces and post-conversion engagement changes and will start quickly in several ex-U.S. markets. **Q: Revenue growth keeps outpacing user growth. By 2027, how much headroom remains vs. other social platforms?** A: ARPU still has substantial room to catch up to Meta, but management did not quantify it. Over 80% of monetization comes from logged-in and App users whose engagement and frequency are an order of magnitude higher than logged-out search-referral users. More than 60% of revenue is performance-based; clicks rose over 50% and conversions more than tripled, the inventory performance advertisers want, driven by ML and signal improvements. This quarter’s growth skewed more to CPM than impressions. In an auction, lower-funnel result CPMs can run into the triple digits, as advertisers pay for intent and conversion. **Q: Within Google traffic changes, which one or two factors stand out the most?** A: Two working hypotheses: aggressive rollout of AI modes that do not drive out-clicks, and possible downranking of machine-translated content. The most significant factor is AI Overviews and AI Mode scaling rapidly in and outside the U.S., shifting from a feature to a full search experience that pushes other results below the fold and does not outlink, fundamentally changing Google’s relationship with the rest of the internet. Downranking of machine-translated content mainly affects ex-U.S. markets, but visibility is limited. Management does not view this as a typical, transient algorithm tweak; Reddit remains a top Google search term, indicating stable user demand, consistent with the strategy to build a direct destination. **Q: Can you scale paid user acquisition more in 2025?** A: App-install campaigns are early but show retention on par with organic users; no forward spend magnitude was given. Spend is primarily in the U.S., with scale still small, though management sees room to ramp. Core product upgrades across Feed, notifications, and personalization should lift retention broadly, improving install ROI. Outside the U.S., brand campaigns improved App install efficiency. The company will also begin evaluating reactivation campaigns for logged-in and App users. **Q: After going logged-in only, will you keep a lightweight logged-out version? Is this just a test or a committed direction?** A: The shift to logged-in only is a firm direction. Tests aim to define the boundary for limited logged-out access before requiring login, balancing value for users arriving via shared links, search, or first-time trials with nudges to sign in. There is ample room to iterate on details. The rationale remains engagement, ad targeting, corpus protection, and regulatory adaptation; management notes Reddit is currently the only major platform that has not yet done this. **Q: Beyond user numbers, any positive signals on engagement?** A: Feed improvements are lifting engagement for both new and existing users, though total sessions are down due to weaker search referrals. Feed has been a top priority this year; new-user retention has improved meaningfully YTD, and posting and commenting grew QoQ. The corpus now totals ~27 bn posts and comments, with 1 bn added last quarter. On-site time is solid, with the divergence mainly in total sessions; the most important logged-in and App engagement is growing. **Q: After moving to logged-in only, how do you view data licensing to OpenAI and Google?** A: The stance is unchanged by the shift; no updates, though negotiations are active, and deals may or may not be reached. Any agreement must advance long-term strategy and represent fair value exchange. Deals are not binary and could include training access, in-product display and usage, and more specific content use cases. The company will only pursue arrangements that are strategically and economically optimal. **Q: Does this period resemble any prior phase, and how do you balance near-term and long-term goals?** A: It is reminiscent of the period ~11 years ago when leadership returned; near-to-mid term, the priority is product work that makes the core App experience more consistent. Then and now, the company is re-examining what Reddit is and its place in the ecosystem; the mission remains to create community and belonging for everyone. There is overlap with social media and AI search in areas like search and shopping, but Reddit’s essence is community and conversation. The core App will help new and existing users reliably find Popular, News, frequent subreddits, games, and video, solving the long-standing issue that the first session struggles to explain the product. **Q: U.S. ARPU has been growing 50%+. How do you view that trajectory?** A: No outlook on the growth rate, only that Reddit’s ARPU potential matches any platform. Drivers include outcome-based ads, a broader affluent audience many of whom are not on other platforms, and communities organized by topics like skiing and travel that provide inherent commercial context. **Q: On product, how do you convert low-frequency users into DAUs?** A: Impose structure on the Feed, classifying posts by use case, content type, and mood to make the experience predictable. Today’s core experience mixes links, text, and video with varying tone; after the redesign, users can quickly find what they want, such as news, last night’s show or game threads, lean-back videos, or games. Knowing what to expect on open fosters daily habits and converts low-frequency users into DAUs. Labels like Popular, News, Games, and Memes also help new users understand the product; this will require ongoing builds and tests with extensive iteration and no timeline given. **Q: What kind of traffic is being lost from Google, and what sticky use cases can offset it?** A: No breakdown of traffic mix; management attributes declines to AI Mode not driving out-clicks rather than demand falling. Reddit remains one of the most searched terms on Google, and crawling intensity has not diminished, but AI Mode does not outlink for any site, fundamentally changing how Google operates. Sticky use cases include staying current on events, discussing last night’s shows and games, and unique content spanning hobbies, interests, and life stages. The company aims to make these experiences more consistent. On the Feed, input signal volume and model scale have grown by an order of magnitude, with sharply shorter training and deployment times; last quarter saw notable gains, with roughly a year of runway left, already boosting engagement for both new and existing users. **Q: In 2027 renewals, do brand clients want GEO exposure via Reddit in frontier models?** A: This is not core to advertiser partnerships, and management is skeptical of GEO. Advertisers buy Reddit to reach highly engaged users and drive scalable, repeatable, measurable outcomes, which is why wallet share is rising. Reddit is inherently hard to game; while the company understands why advertisers consider GEO amid search changes, it is not a primary driver. **Q: Will you acquire to expand the user base?** A: Not considering M&A to buy users; priority is to mine three in-house channels. Historically, deals focused on small teams and tech, never on large user-acquisition plays; modest improvements in new-user retention are hard to match via M&A. The three channels are: a large volume of unattributed word-of-mouth downloads with new-user retention up 50%+ YoY; a large number of reinstalls, likely for specific content with uninstall afterward, which an incognito App mode may mitigate; and converting search-referred mobile web users to the App, with conversion prompts reaching only ~13% of eligible users this quarter. **Q: Post shift, how should we model the share of DAU that is logged-in vs. ARPU of logged-in users?** A: No modeling framework was given; tests are designed to answer these questions, and conversion rates are not yet known. Over 80% of revenue comes from logged-in users, so the shift will necessarily expand that cohort, but the time path of conversion is uncertain. Peers have taken varied but successful approaches; the company will find the one that fits best. **Q: Will logged-in only prevent scraping, and what is the impact?** A: Scraping will be harder but not impossible, which is one reason for the shift. Peers have long required login for this reason; Reddit’s corpus is highly valuable, and the goal is to protect and govern access. Ecosystem impact will be evaluated through testing. **Q: Will logged-out users trend to zero long-term?** A: They will not go to zero, but over time App, logged-in, and total DAU should converge with only a small % gap. The company will always preserve logged-out entry points such as search referrals, shared links (which it aims to expand), and the homepage; the core product is in App and logged-in, and users will be steered there. Convergence could arrive earlier than the long term. **Q: With Google ads under pressure, could the pendulum swing back to outlinking?** A: No assumptions about Google’s path; focus is on direct traffic the company can control. More changes from Google are expected; AI Overviews/AI Mode do not outlink, and the company is not counting on them. Work to build a direct, sticky daily destination is showing results. **Q: Why was the upside vs. expectations smaller this quarter?** A: Demand was strong but concentrated in Sept., and some bookings on a few peak days could not be delivered. Management noted that 53% growth is rare. About 50% of the quarter was pre-booked at the start, with the rest estimated; certain Sept. days were especially strong, and a few days’ bookings could not be fulfilled, but demand was robust throughout the quarter. \ **Risk Disclosure and Disclaimer:**[**Dolphin Research Disclaimer and General Disclosure**](https://support.longbridge.global/topics/misc/dolphin-disclaimer) ### Related Stocks - [RDDT.US](https://longbridge.com/en/quote/RDDT.US.md) - [MAX.US](https://longbridge.com/en/quote/MAX.US.md) - [GOOG.US](https://longbridge.com/en/quote/GOOG.US.md) - [GOOGL.US](https://longbridge.com/en/quote/GOOGL.US.md) - [INHD.US](https://longbridge.com/en/quote/INHD.US.md) - [GOOGN.US](https://longbridge.com/en/quote/GOOGN.US.md) - [GGLL.US](https://longbridge.com/en/quote/GGLL.US.md) - [GGLS.US](https://longbridge.com/en/quote/GGLS.US.md) - [GOOY.US](https://longbridge.com/en/quote/GOOY.US.md) - [GOOX.US](https://longbridge.com/en/quote/GOOX.US.md) ## Comments (1) - **Benzard_epnOP2 · 2026-10-31T05:35:48.000Z**: can bounce back? --- > **Disclaimer: This article is for reference only and does not constitute any investment advice.**