MediaAlpha (MAX) 2026年第3四半期決算説明会：売上高23%増、第4四半期ガイダンス

要点

• MediaAlphaが発表した2026年第3四半期決算は過去最高となり、売上高は前年同期比23%増の3億7600万ドル、コントリビューション（貢献利益）は26%増の5700万ドル、調整後EBITDAは36%増の3900万ドルに拡大しました。
• 自動車保険収入は17%増加したほか、住宅・借家人保険収入は市場環境の改善とシェア拡大に支えられ、116%増の6600万ドルと過去最高を記録しました。
• 代理店向け売上高は約46%増の2100万ドルとなりました。MediaAlphaの活発な利用代理店数は前年同期の約2,000社から3,000社以上に拡大しました。
• 経営陣は第4四半期の売上高を3億5500万ドル〜3億7500万ドル、調整後EBITDAを3600万ドル〜3900万ドルと予想しています。2026年通期のフリーキャッシュフローの見通しは9000万ドル〜1億ドルに据え置かれました。
• 同社は保険会社の支出拡大、高利益率のオープンマーケットプレイスの拡大、損害保険業界におけるデジタル広告のさらなる浸透を理由に、2026年以降も主要財務指標で2桁成長が続くと見込んでいます。
• MediaAlphaは2025年9月以降、約1億1500万ドルの自社株買いを実施し、発行済株式数を約1120万株（16%）削減しました。

主要財務データ

第3四半期の売上高の10%以上を占めた保険会社は3社でした。第4位の保険会社は9.6%を占め、主要顧客への集中度は維持しつつも顧客ミックスの多様化が進んでいることを示しています。

事業および業績の動向

自動車保険は、MediaAlphaの損害保険部門において最大の成長ドライバーであり続けました。経営陣によると、保険各社は契約獲得に向けて競争を活発化させており、広告予算を増やして同社のマーケットプレイス経由の支出を拡大させています。

住宅・借家人保険は最も高い成長率を示しました。フロリダ、カリフォルニア、テキサスなどの州で保険会社が受入容量と広告を拡大したため、売上高は116%増の6600万ドルに達しました。経営陣は、この改善の主な要因を好調な市場環境とシェア拡大によるものとしています。

代理店チャネルも拡大を続けました。MediaAlphaは第3四半期に代理店経由で2100万ドルの売上高を計上し、3,000社以上の活発な代理店と直接連携しました。保険契約の半数近くが依然として代理店経由で結ばれているため、経営陣はこのチャネルを長期的な大きな機会と捉えています。

同プラットフォームは、当四半期中に年換算で25億ドル以上の損害保険広告支出をサポートしました。MediaAlpha独自の最適化モデルを使用したある保険会社のテストでは、コンバージョン率が約80%向上しました。マッチング精度とキャンペーン成果の向上により、オープンマーケットプレイスに更なる広告主予算とパブリッシャーの供給を惹きつけることができると経営陣は述べています。

経営陣の見通し

第4四半期の売上高のうち、医療保険の割合は約4%となる見込みです。調整後EBITDAの見通しには、65歳未満向け医療保険からのコントリビューションが前年同期比で約90万ドル減少することが織り込まれています。

65歳未満向け医療保険を除いた場合、中間値ベースで第4四半期のコントリビューションは21%増、調整後EBITDAは26%増を見込んでいます。MediaAlphaは2026年通期のフリーキャッシュフロー見通しを9000万ドル〜1億ドルに維持しました。

2027年以降を見据え、経営陣は売上高、コントリビューション、調整後EBITDAのすべてで2桁成長を見込んでいます。この見解は、上位2社のパートナー保険会社による継続的な支出、3位から5位の保険会社の支出成長の加速、高利益率なオープンマーケットプレイスの比率上昇に基づいています。

資本配分

MediaAlphaは第3四半期中、自社株買いおよび租税受取合意（TRA）債務の削減に2100万ドルを投入しました。平均価格11.25ドルで約80万株を900万ドルで取得しました。

また、同社は2300万ドルのTRA負債を47%割引の1200万ドルで買い戻し、1100万ドルの利益を計上しました。経営陣は、本来であればTRA保持者に支払われていたはずの追加的なキャッシュフローを、年間約500万ドル手元に残せると見込んでいます。

四半期末から10月26日までに、MediaAlphaはさらに180万株を平均価格10.38ドルの計1800万ドルで買い戻しました。1億ドルの自社株買い枠のうち8200万ドルを使用しており、自社株買いのペースを維持するため四半期末後にリボルビング信用枠から1000万ドルを借り入れました。

リスクおよび注視事項

• マーケットプレイスへの参加は広がっているものの、売上高は引き続き大手保険数社に集中しています。
• コントリビューション利益率は、導入方式の構成比、プロダクトミックス、パートナーシップの動向、およびオープン／プライベートマーケットプレイス間の移行によって変動する可能性があります。
• 経営陣は、パーソナルAIエージェントが保険の比較検討（特に顧客獲得ファネルの上流部分）をどのように再構築するかについての不確実性を認めています。
• 第4四半期には通常の季節性があり、損害保険の供給がやや減少し、医療保険の動きが活発化します。
• 保険料の急騰に伴う消費者の保険比較検討の急増は落ち着いていますが、経営陣は今後数年間、活発な状態が続くと予想しています。

アナリストQ&Aの要点

経営陣によると、損害保険業界は成長指向サイクルの初期段階にあります。保険会社の強い収益性がシェア獲得競争、価格引き下げ、広告支出の拡大を支えており、これがMediaAlphaのマーケットプレイスにとって追い風となっています。

AIについて経営陣は、消費者に代わって検索・比較を行うパーソナルAIエージェントの影響はまだ見られないと述べています。同社はAIが情報収集や調査段階に影響を与えると予想していますが、ブランディング、規制、開示事項、顧客体験の観点から、見積もりや契約締結の管理権限は保険会社が維持すると考えています。

また、経営陣は、住宅保険や中小企業向け保険といったより複雑なカテゴリーにおいて、AIが消費者直接（D2C）販売の追い風となる可能性があると見ています。より精通した消費者は、購入を完了するまでに代理店やコールセンターからの支援をそれほど必要としなくなる可能性があります。

利益率に関して、経営陣はコントリビューション利益率が比較的安定して推移すると述べており、第4四半期の見通しはモデリングの合理的な基礎を提供しています。ただし、パートナーシップの調整やマーケットプレイスの構成比の変化により、四半期ごとに変動が生じる可能性は依然としてあります。

業績説明会（コール）トランスクリプト全文

決算説明会の完全なトランスクリプト

経営陣による説明

Operator

Hello, and thank you for standing by. My name is Lacey, and I will be your conference operator today. At this time, I would like to welcome everyone to the MediaAlpha, Inc. Third Quarter 2026 Earnings Call. [Operator Instructions]

Thank you. I would now like to turn the call over to Alex. Please go ahead.

Alex Liloia

Thanks, Lacey. Good afternoon, and thank you for joining us. With me are Co-Founder and CEO, Steve Yi; and CFO, Tigran Sinanyan.

On today's call, we'll make forward-looking statements relating to our business and outlook for future financial results, including our financial guidance for the fourth quarter and full year of 2026. These forward-looking statements are subject to risks and uncertainties that could cause actual results to differ materially. Please refer to our SEC filings, including our annual report on Form 10-K and quarterly reports on Form 10-Q, for a fuller explanation of those risks and uncertainties and the limits applicable to forward-looking statements. All the forward-looking statements we make on this call reflect our assumptions and beliefs as of today, and we disclaim any obligation to update such statements, except as required by law.

Today's discussion will include non-GAAP financial measures, which are not a substitute for GAAP results. Reconciliations of these non-GAAP financial measures to the corresponding GAAP measures can be found in our press release and investor supplement issued today, which are available on the Investor Relations section of our website.

I'll now turn the call over to Steve.

Steven Yi

Thanks, Alex. Hi, everyone. Thank you for joining us. The third quarter was the best quarter in MediaAlpha's history across all of our key metrics. During the quarter, our revenue base continued to broaden as multiple carriers stepped up their budgets meaningfully, driven by intensifying competition for policy growth. This resulted in strong volume for our supply partners, who delivered more high-quality Consumer Referrals into our marketplace. We continue to execute very well in these favorable market conditions, which we expect to continue for the foreseeable future.

Auto insurance revenue grew 17% year-over-year and remains the largest driver of growth in our property and casualty insurance vertical. We're also seeing more carriers put real ad dollars behind home insurance, where we've historically had a much smaller footprint. In home and renters, revenue hit a record $66 million in the third quarter, up 116% year-over-year. That growth was driven by strengthening market conditions and wallet share gains, as both traditional and natively online direct-to-consumer digital carriers allocated more of their budgets to us in pursuit of efficient growth.

As carriers evolve towards more multichannel distribution, we're uniquely positioned to help them acquire customers effectively across both their direct-to-consumer and agent-based channels. As a reminder, we started our agent business from scratch in Q1 of 2021 to help captive and independent insurance agents acquire customers efficiently and directly on our platform using the same programmatic capabilities we offer to our carrier partners. We invested in automation to serve a growing number of agents with a much smaller, leaner team, which enables us to scale efficiently. Today, we work directly with over 3,000 active agents, up from approximately 2,000 a year ago. In the third quarter alone, agent revenue was $21 million, up approximately 46% from $14 million a year ago. Agents continue to represent a significant share of insurance distribution and an attractive growth driver for MediaAlpha, complementing our direct-to-consumer capabilities.

Our platform powered over $2.5 billion of annualized P&C advertising spend in the third quarter. We believe this scale, which is the largest in our industry, makes our predictive algorithms more effective at delivering unique insights and superior value to carriers. For example, in a recent test, a carrier that opted into our proprietary optimization models saw conversion rates improve by roughly 80%, which drove far greater sales at a lower cost per policy. We love these kinds of tests because they demonstrate to carriers the value of our data scale advantage through real measurable efficiency gains. As we keep delivering such results, we expect carriers to shift more budget into our higher-margin open marketplace. This drives the flywheel of our 2-sided marketplace by further increasing our scale advantage, which allows us to deliver even more differentiated insights and greater efficiencies over time.

As we look ahead, we expect consumers to increasingly leverage AI to help them shop for insurance. While no one can predict exactly how this will unfold, we remain confident that major carriers will retain a high level of control over how their policies are advertised, distributed and sold. We're also confident that advertising will emerge as the dominant monetization paradigm for consumer AI companies, creating a new acquisition channel for our supply partners. We've seen ChatGPT aggressively roll out advertising to free tier subscribers. In September, according to Similarweb, 43% of free tier ChatGPT users in the United States were shown ads. While the impact of agentic commerce, as enabled by personal AI agents like Muse, will vary widely across consumer products and services. Large brands and merchants are already signaling that they intend to retain control over how AI agents access their systems and interact with customers on their behalf.

Earlier this month, Meta and Sierra with partners, including Shopify, Stripe and Walmart, announced a Personal Agent Protocol, an emerging standard designed to let businesses verify who an AI agent represents and sets rules for what an agent is allowed and not allowed to do. That reflects a simple reality. Companies want to control how third parties access and interact with their websites, systems and customer-facing experiences. Insurance carriers are no different, and we expect quoting and binding to largely stay in carrier-controlled systems, whether the shopper is a person or an AI agent.

As we look ahead to 2027, we're confident in the momentum we're seeing as broader carrier mix reflects the wider demand in the market. We're continuing to operate in a growth-oriented auto insurance market cycle, and our marketplace delivers the scale and efficiency that carriers need to compete. That competition is driving higher carrier budgets, and more of those dollars are flowing through our marketplace. Longer term, the shift towards direct distribution will continue with carriers expected to shift a growing portion of their $35 billion of spend in insurance agent commissions into digital marketing over time. As we wrap up 2026, we believe our business has years of durable double-digit growth to come, and that will translate into real value for our shareholders.

With that, I'm excited to welcome Tigran back as our CFO. Tigran, I'll hand it over to you to take us through our numbers.

Tigran Sinanyan

Thanks, Steve. It's great to be back, and I'm proud to share these results with you and the team. Our record results exceeded the high end of our guidance ranges for revenue, contribution and adjusted EBITDA. I'll cover what drove our results, then cover capital allocation and our fourth quarter outlook.

Revenue for the third quarter was $376 million, up 23% year-over-year, reflecting broader carrier participation and increased spend in our marketplace. That broadening also showed up in our customer mix, and we're pleased with our progress in diversifying demand. Three carriers each represented more than 10% of third quarter revenue, and our fourth largest carrier was just under that mark at 9.6%. Contribution was $57 million, up 26% year-over-year and outpacing revenue. That reflects a mix shift toward our open marketplace, led by carriers that rely on us as a trusted partner for optimization and account management. Adjusted EBITDA was $39 million, up 36% year-over-year. Excluding under-65 Health, revenue and adjusted EBITDA grew 26% and 42% year-over-year, respectively, reflecting very strong performance in our core business.

Growth accelerated as we move down the P&L. Revenue grew 23%, contribution grew 26% and adjusted EBITDA grew 36% year-over-year, reflecting the operating leverage in our model. As contribution grew, a higher share of each incremental dollar flowed through to adjusted EBITDA. Contribution margin, by contrast, can fluctuate from quarter-to-quarter with deployment mix, product mix and partnership dynamics. This quarter, we converted 69% of contribution into adjusted EBITDA, our highest since 2024. That operating leverage drives strong cash generation and enables us to deploy capital aggressively to create shareholder value.

Cash generation continues to be strong. In the third quarter, we deployed $21 million of capital across share repurchases and reductions to our obligations under the tax receivable agreement, or TRA. We repurchased approximately 800,000 shares for $9 million, representing an average price per share of $11.25. We also continued to reduce our long-term obligations under the TRA. In September, we repurchased $23 million of our total TRA liability for $12 million, representing a 47% discount, which generated an $11 million gain that we recorded in the third quarter. We funded this transaction with cash on hand. Since quarter-end and as of October 26, we have repurchased an additional 1.8 million shares for $18 million, an average price of $10.38 per share, utilizing $82 million of our $100 million authorization. This brings our total repurchases since September of 2025 to approximately $115 million, a reduction of approximately 11.2 million shares or 16% of our shares outstanding. We ended the quarter with $18.7 million in cash and $30 million undrawn on the revolver. To support the pace of repurchases early in the fourth quarter, we drew $10 million on our revolving credit facility after quarter-end, which we view as an attractive use of capital.

Turning to guidance. Our philosophy and approach remains unchanged. For the fourth quarter, we expect revenue of $355 million to $375 million, up approximately 25% year-over-year at the midpoint, contribution of $54 million to $57 million, up approximately 18% year-over-year at the midpoint, and adjusted EBITDA of $36 million to $39 million, up approximately 22% year-over-year at the midpoint. For Q4, we expect the health vertical to be approximately 4% of total revenue. Our adjusted EBITDA guidance includes a decline in contribution from under-65 Health of approximately $900,000 year-over-year. Excluding under-65 Health, we expect contribution to increase 21% and adjusted EBITDA to increase by 26% year-over-year at the midpoint. For the full year 2026, we continue to expect to generate $90 million to $100 million in free cash flow.

We're proud of the progress the team has made this year, and we're excited about what's ahead. Overall, we remain confident in the strength of our position and the opportunity ahead of us.

With that, operator, we are ready to take the first question.

Operator

[Operator Instructions] Your first question comes from the line of Maria Ripps with Canaccord.

質疑応答

Maria Ripps

Congrats on the strong results. And Tigran, welcome back as CFO. It's great to hear your voice on the call again. First, I just wanted to ask about your expectations for double-digit growth in 2027 and beyond. Just can you maybe talk about whether that's for revenue or contribution? And then, can you maybe help us bridge that between industry growth and share gains from carriers beyond the top 2? And does that assume that your top 2 carriers hold their spend -- their 2026 spend levels?

Tigran Sinanyan

It's good to be back. I think when it comes to bridging the double-digit growth, I'll say that, first, our top 2 carrier partners, the leaders in direct-to-consumer customer acquisition, continue to invest in the platform and continue to invest in our marketplace. And so, what we're seeing is a continuation of what we saw in Q3, which is an acceleration in the velocity of spend from carriers 3 to 5. I'd say that we expect that double-digit growth to carry across all of the metrics, particularly because a lot of the growth is coming from our open marketplace where we have higher contribution margin.

So just kind of stepping through the industry dynamics as a starting point, premium CAGR in P&C over a long period of time has been roughly 6%. Digital advertising remains underpenetrated in P&C. Our latest stats show about 40% online digital ad penetration relative to other industries, which are at 80%. We're starting to see that catch up. And of course, there's a secular shift to direct-to-consumer. And that, along with that broadening that we've seen in Q3 and what we're seeing into Q4, it's not hard to bridge from that to a double-digit CAGR and expectation for us going forward.

Maria Ripps

Got it. That's very helpful. And then, just in terms of AI agents and sort of what's going on in the industry, so it feels like AI agents like Muse can now fill out forms and sort of complete purchases on behalf of users. Are you seeing sort of any sort of agent-driven shopping sort of already happening in the marketplace? And then, if shopper -- sort of your shopper becomes an agent rather than a person, how does your sort of pay per click model change as a result?

Steven Yi

Yes. Maria, I can handle that question. No, we haven't seen any impact of that of like personal AI agent shopping on behalf of consumers. Certainly, we've seen some of the demos that people have been talking about. I think that this is I think something that's going to foreseeably shape the early part of the consumer shopping experience. But I think ultimately, when it comes to quoting and binding, that the carriers are going to retain control of those aspects of the consumer experience.

First of all, consumers have invested tens, if not hundreds, of billions of dollars in building brand awareness and getting consumers to interact with them directly, be it one of their agents or through their website. The carriers aren't going to relinquish that control easily or lightly by allowing a personal AI agent to dictate and control how a consumer is going to access their site and systems. And so -- and that's not a position that's inconsistent with a lot of other brands in the space. I mean, you saw companies like Amazon and Delta Air Lines, Charles Schwab, JPMorgan, all block attempts by personal AI agents like Muse to actually crawl their sites, and I think for a good reason. A lot of these companies are big incumbents who have a lot more to lose than to gain by allowing for personal AI agents to access their systems without their permission.

In addition to that, these are highly regulated industries. So they want to control exactly what the user sees and what's disclosed to the user before a purchase is made. And certainly, that applies to insurance carriers. And so, historically, there have been other attempts to try to use an agent or a crawler to go and shop on behalf of a consumer and aggregate a bunch of rates. Those types of business models have been attempted for the past 10 to 15 years. I will say that they were very unsuccessful in this regard. And certainly, AI technology makes this a little easier, but it doesn't really change the reasons why it didn't work the first time, which is really largely dependent -- largely because the carriers have a lot vested in controlling the consumer experience. And so, what we expect to happen is like protocols like the Personal Agent Protocol, right, to really allow brands and companies to be able to control exactly how personal AI agents can interact with their websites and systems. And once they do that, we think that the dynamics will remain largely similar where the large branded carriers, again, will really want to have the quoting and the binding happen on their sites and not through a third-party personal AI agent or through an LLM.

Operator

Your next question comes from the line of Andrew Kligerman with TD Cowen.

Andrew Kligerman

So I was interested in the part about the $21 million in agent revenue. And perhaps you could share with us what you see as your share of that market right now and how big that market can be for you? Yes, just what are the KPIs around that? And how do you see that 5 years down the road?

Steven Yi

Sure. I'll start with that answer, which is -- I mean, the agent business, keep in mind, close to 50% of all policies are still written through agents. And so, when you break down the market opportunity, although we expect to see a continuing secular shift to direct-to-consumer sales versus agent-based sales, agent-based sales are still going to remain a huge part of the overall distribution channel for this industry. And these agents, I think, still need to be connected to the direct online shopping ecosystem. And so, what we see is the long-term opportunity here being comparable to what the market share is for agents in general. And so, it's an enormous market opportunity.

In terms of where our market share is compared to others, we started this business, the agent business from scratch in 2021, and we're growing this business. And so, our footprint here is smaller than that of some of our other competitors who've been in this business for 10-plus years. And so, what we expect is that we can gain market share because one of the things that we've been able to do is bring a lot of programmatic efficiencies to how agents buy leads. We're seeing that in just better ROI that they're getting from the leading call purchases that they're making through our ecosystem and using our platform. We're seeing that lead to these partnerships like the Farmers Lead Marketplace, where we're powering lead purchasing -- programmatic lead purchasing for all Farmers agents. And so, the ability to really support that of behalf of carriers like Farmers is really leading to us having a lot of momentum behind our agent business. And so, we expect this business to be a very large one for us going forward and one where we have a lot of market share gains that we can expect in the upcoming years.

Andrew Kligerman

That's awesome. And just as we think about shopping and the pricing cycle, Steven, where are we in terms of that -- the level of shopping right now? And how do you see that playing out over the next year or 2? Have we kind of peaked out in terms of this activity and what MediaAlpha will see? Or do we still have a few more years of it?

Steven Yi

Yes. I mean, I think in terms of -- so I think you're asking about just maybe where we are in the underwriting cycle and what that -- how that may impact shopping. And then, they're slightly different. And so, let me just try to answer. One is -- so I think shopping remains at elevated levels. Now, shopping spikes when carriers raise rates, and they raise rates during these hard market cycles, when we emerged from back in 2024, 2025. And so, I think what you're seeing is that the rates of shopping has come down since that period when people's prices were getting increased by 30%, 40%, and that led to increased shopping. But I think because insurance has seen overall inflation of 40%, 50% over the last few years, I do think that you're going to see heightened levels of shopping persist over the next several years.

And so, now, we've always said that ambient shopping behavior does have an impact on the volume in our marketplace. But what has a far greater impact is carrier demand and carrier growth appetite. And really, where we are in the insurance underwriting cycle is still in the very early stages of what is a growth-oriented soft market cycle. The second quarter of this year was the most profitable period by some measures within the P&C industry in 25 years. And what that's doing is, it's really fueling a market share growth competition between all the major carriers.

You see a lot of the carriers talking about how competitive the market is. And during these competitive market cycles, what happens is that carriers tend to lower rates to have their policies be more competitive, and they spend a lot more in advertising. This has the effect of actually lowering their overall profitability from peak levels, right? But obviously, all of this competition and increasing ad spend has a strongly beneficial effect for our marketplace. And you're really seeing that play out with the broadening of demand in our marketplace. And I think that you can continue to expect these tailwinds to persist for the foreseeable future just because the industry remains at a very, very strong profitability position.

Operator

Your next question comes from the line of Bryan Smilek with JPMorgan.

Bryan Smilek

Look, Steve, great to see the execution. I believe the 4Q guide does imply solid P&C acceleration. Can you just help me parse out the drivers there and just like more specifically by vertical as well? I thought home and renters is growing quite nicely in triple digits. Can you just elaborate a bit on what's going on there? Is it really more macro-driven or share of wallet-driven?

Steven Yi

Yes. So, on the home insurance side, I think it is a lot more macro-driven. I think what you're seeing is a recovery of the homeowner market, which is a lot more state-driven than auto tends to be. And so, what you're seeing is, in part because we just haven't had a lot of catastrophic events happen like in Florida or in California over the past 1.5 years to 2 years. I think what that's leading to is a lot of carriers going back into those large states that had in the past been problematic states, and so states like Florida, California and Texas. And so, as you see these carriers going back into those states and increasing capacity there, you're obviously seeing them now starting to advertise in those states. And so, as those budgets go up, the advertising budgets go up, you're seeing our supply partners really reacting to that by expanding the amount of volume that they can make available to these carriers. And so, I think that a lot of the growth there is based on just increasing capacity for homeowners insurance and the growth appetite that you're starting to see among homeowner insurers and just seeing that demand reflected in the overall volume in our marketplace.

I think overall, in Q4, you're going to continue to see growth being driven by just the increased growth appetite of our auto insurance carriers. Again, that marketplace is in a very, very good position profitability-wise. That's really resulting in carriers investing a lot into direct distribution and direct advertising. Again, you're seeing that being reflected with how broad -- how much broader the demand is within our marketplace versus 2 to 3 years ago. I think you're just going to continue to see those trends in Q4.

Operator

Your next question comes from the line of Tommy McJoynt with KBW.

Thomas Mcjoynt-Griffith

Going back to the subject of personal AI agents, is that a risk? Steve, when you talked a little bit about -- your response to one of the earlier questions, is there a risk that MediaAlpha worries about that the largest direct-to-consumer carriers do establish formal Personal Agent Protocols with popular personal AI agents and that cuts out MediaAlpha? Or do you view your marketplace as still having a solution to provide in those types of arrangements to the extent they do become somewhat commonplace?

Steven Yi

No, it's a good question. I think that both -- I mean, first of all, I think that -- I mean, I guess, my position is that it's -- that what you have to expect is that then what the carriers really don't want, right? They don't want an environment where their rates are displayed outside of their own websites or through someone other than their own agents. And that's not true for independent agent carriers who don't have control of the consumer experience as it is. They sell their products primarily through agents and brokers. But that's certainly the case for direct-to-consumer carriers and the captive agent carriers. And it's really those carriers who represent the vast majority of the overall industry that we expect to really be resistant to efforts by personal AI agents to pull in their rates into an environment so that it can be easily matched against rates from other smaller carriers.

That level of price transparency isn't something that the big carriers want for a variety of reasons. And so, the -- I think the premise that you think that a carrier would do a deal directly with the personal AI agent to show their rates in those, I mean that's the premise I would question. Certainly, I think that you can -- it's foreseeable that a user will start their shopping experience on a personal AI agent through a personal AI agent or through an LLM. But ultimately, they're going to have to go over to the carrier site to complete that process. And that type of a direct referral is something that you see through Google, et cetera. But keep in mind, just the ecosystem that's going to arise around these personal AI agents and LLMs, they're going to be advertising. And so, just like Google, you can go on Google search for insurance go directly to GEICO to buy a policy, right? Or you can or you can go directly to a price comparison site or a lead generation site to shop for insurance. And then, that's where we have to make that referral, right, that connection between one of our publishers and the carriers.

And I think that, number one, our supply base is going to really tap into the advertising ecosystem that a ChatGPT is creating and that you're likely to see with something like a Muse. And then, in addition to that, even if people do end up searching for insurance and starting their experience at LLM or through a personal AI agent, we fully expect our publisher base to remain resilient, tap into other media channels to really generate a lot of strong shopping volume for the foreseeable future that we can then help to direct to the right carriers.

Thomas Mcjoynt-Griffith

Got it. And then, switching over, it's good to see you on track to hitting your buyback goals for this year. And as you think about heading into next year, talking about double-digit increases across the various financial metrics, can you talk about your expectations for free cash flow conversion into next year and specifically how the recent tax receivable transactions will impact that going forward? And then, with that strong free cash flow generation, what are your plans in terms of buybacks and other forms of capital deployment next year?

Tigran Sinanyan

Thanks, Tommy. I'll handle that. With respect to the TRA acquisitions, the effect of that on cash flow is that we'll be retaining around $5 million a year of incremental cash flow that would have gone out to the TRA holders. And now that we've retired that liability, that money stays in the company. And we continue to benefit from the cash tax shield, right? In 2026, cash taxes will be approximately 1%. Going forward in '27, you can expect that to be in the mid-single digits and then in the mid-teens thereafter. And so, there's a lot of cash savings and return that we're generating from those TRA acquisitions. I'd say roughly mid-teens in terms of an IRR. And the cash flow conversion for us will remain high.

We remain a low-CapEx business. And so, as we continue to generate contribution growth, which flows down to adjusted EBITDA at the midpoint of our guidance at 68%, really the major kind of outflows on cash are net working capital and a little bit on debt service, right? So we'll have plenty of opportunity to deploy capital efficiently. And I'll say that the bar remains high for us. We're trading at effectively mid-teens free cash flow yield. And so, that's kind of where we're going to measure ourselves in terms of the returns we can generate. At current prices, I think we've demonstrated over the last 5 quarters that we are bullish on the stock and believe that it's an attractive investment. And we've deployed capital fairly consistently quarter-to-quarter. Between TRA acquisition and share repurchases, last quarter, we deployed $21 million. The quarter prior was $20 million on share repurchases. Again, the quarter prior was $20 million. And so, you can expect us to continue to look at the opportunities ahead of us in terms of generating return and making the most prudent decision.

Operator

Your next question comes from the line of Eric Sheridan with Goldman Sachs.

Eric Sheridan

During the prepared remarks, you talked about one partner running an optimization test that saw much higher conversion. I'm curious if you could maybe pull on that thread a little bit. When you think about your platform and the potential scope for evolution of it, how are you thinking about conversion improvements, greater scale of optimization and how that may result in either a higher return profile from the spend that's already on the platform or the means to possibly attract more net advertising dollars to the platform as you prove out a [ scaling ] ROI on the back of some of those initiatives?

Steven Yi

Yes. Eric, it's a good question. I mean I think -- yes, I mean, I think you summarized kind of how we think about it, which is that as we have the scale advantage -- we're the largest marketplace within the industry, in some ways, several times that of some of our competitors -- that data scale advantage is really translating itself into the ability to leverage predictive AI to deliver better outcomes, better matching for our advertisers and then higher yield for our publishers. And so, ultimately, because we can leverage the data and apply AI against the data to provide better outcomes for both the participants in that marketplace, as you can imagine, what that results in is just the attraction of more budget and more advertisers and then more supply because publishers understand that the yield that they can get and the money that they can make by exposing their advertising inventory through our marketplace is better than what they can get anywhere else.

And so, it's a virtuous cycle. It is a flywheel effect that you really see taking effect here. And it really has to do with our scale and the data advantage that that's giving us and our AI capabilities and the data science capabilities that we have internally and then doing just an outstanding job of optimizing campaigns on behalf of our advertisers and doing that in a way that's better than any of our competitors can do and better than any carriers can do directly for themselves.

Operator

Your final question comes from the line of Mike Zaremski with BMO.

Michael Zaremski

Maybe I'll start with just a bigger picture question, sticking with the conversation about the evolution of AI and shopping experiences. And obviously, a lot of still directions of things can settle out. But I guess, is it reasonable -- is there a reasonable scenario, I guess, in which shopping habits push the industry to -- either your industry and/or the carriers to consolidate in a much bigger way to the extent shopping habits do meaningfully change towards increased [ shopping on ] Internet?

Steven Yi

Let's see. Yes. I'm not sure how shopping habits -- and feel free to elaborate, Mike, because I might be missing the question. But I'm not sure how shopping habits and increased shopping habits by insurance consumers, particularly -- and the evolution of shopping habits to an AI platform or an ecosystem could lead to consolidation within the industry. I would say that certainly, I think -- as I mentioned, I think you're going to continue to see elevated shopping patterns among consumers just because the price of auto insurance is so much higher than it was before. And so, there's just a built-in incentive by consumers to really try to get the best price.

In terms of the AI's impact on the insurance shopping experience, what we talked about is really that having an impact and potentially disruptive impact on the upper funnel of that shopping experience, where instead of going to Google and searching for auto insurance, they start off in ChatGPT or with a personal AI agent and try to get some information through one of those sources, leveraging AI to become better informed about shopping. But that ultimately, the shopping experience is going to have to be the quoting process at least, and then certainly, the binding process will have to be hosted by the carriers themselves. And so, I'm not sure whether or not those trends will lead to consolidation within the space. If there's some aspect of that question I was missing, feel free to clarify that.

Michael Zaremski

No, you got it. That's helpful. I guess, the premise was just there's a number of carriers and you know the universe well that are just kind of watching and don't really play much in direct-to-consumer and are low-cost providers, et cetera. So this could just accelerate the longer-term shift towards D2C as a premise.

Steven Yi

Well that absolutely. And so, what you've seen is, 15 or 20 years after auto insurance was shoppable online, you still have roughly half of consumers buying auto insurance through an agent. And that's just because insurance is a more complicated product, right? Things like homeowners, small business, commercial insurance, slightly more complicated like life insurance, those are products that are smaller within our ecosystem and within the online advertising ecosystem because they're more complicated products that really require human intervention and human explanation.

And now, one exciting thing that we foresee is that when people start off their shopping experience in an AI environment or they leverage AI capabilities that are offered by a carrier themselves in an online interface that they're going to become much better informed about these complicated purchasing decisions and will have less of a need to actually have a human intervene in the shopping process for them to -- before they're ready to make a purchasing decision. And so, it certainly, I think, will have an impact in the auto space. I think it's going to have a pretty big impact in the homeowner and small business commercial space.

And what I'm trying to say is that AI does have the potential to actually have a far beneficial impact to unassisted direct-to-consumer enrollment channels where a consumer can just go and start the process and finish the process really without having to resort to turning to an agent for assistance, again, with the more complicated insurance products. And so, we do see AI as a potential strong tailwind for these non-auto insurance verticals, which, again, right now, primarily are closed by offline agents and call centers.

Michael Zaremski

Got it. That's a good perspective. And I guess, just lastly, in terms of the contribution margin trend line, we can kind of see the guidance. I think historically, there's a little bit of 4Q seasonality from health, but health is just much smaller, right? So just it looks like we should continue to kind of think about the contribution margin trending down a bit unless -- you did make some remarks earlier that might contradict that in the future.

Tigran Sinanyan

Thanks, Mike. No, I think with respect to contribution margin percentage, you can expect that to be relatively stable. And what we've guided to is a good indication of where that is. Now, that could move about, depending on partnership dynamics, mix. But you're right to point out that in Q4, there's a little bit of seasonality with P&C seeing a little bit less supply in a normal kind of seasonal period and health seeing an uptick. Overall, I'd say our open marketplace take rates across those 2 verticals are steady and seeing some growth and stable. And so, something that could impact the contribution margin percentage would be movements in private marketplace, which would lift contribution margin percentage. But overall, I think what you're seeing in the guide is indicative of what you should be modeling.

Operator

There are no further questions at this time. Ladies and gentlemen, this concludes today's conference call. You may now disconnect.