Lendingtree Inc (TREE) 2018 Q3 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the LendingTree Inc. Third Quarter 2018 Earnings Conference Call. (Operator Instructions) As a reminder, today's conference is being recorded for replay purposes.

It is now my pleasure to turn the conference over to your host, Mr. Doug Lebda, Chief Executive Officer. Please go ahead.

Thank you, operator, and good morning to everyone joining the call today, I want to use my time with you to offer my thoughts on the business, run through the progress we're making on key initiatives and provide some context on what we're seeing in the broader market. J.D. will then cover the quarter's financials and our updated guidance. Before we jump in, let me provide the usual disclaimer.

During today's call, we may discuss LendingTree's plans, expectations, outlooks or forecasts for future performance. Forward-looking statements are typically preceded by words such as we expect, we believe, we anticipate or other similar statements. These forward-looking statements are subject to risks and uncertainties, and LendingTree's actual results could differ materially from the views expressed today. Many but not all of the risks we face are described in LendingTree's periodic reports filed with the SEC.

On this call, we will discuss a number of non-GAAP measures, and I refer you to today's press release available on our website at investors.lendingtree.com for the comparable GAAP measure, definitions and full reconciliations of GAAP measures -- of non-GAAP measures to GAAP.

With that, let's get into it. Overall, I'm pleased to report that LendingTree once again delivered a record quarter in terms of revenue, variable marketing margin and adjusted EBITDA, and even more pleased with the strategic and operational successes we've had during this quarter. There are a few key areas I'd like to focus on today: One, the success of our diversification strategy; two, what we're seeing in the mortgage market; three, our track record in M&A; and four, our progress on My LendingTree. First, let's talk about the diversification of our product portfolio.

Five years ago, we consciously set out to expand into minimum loan categories. Although we always had a variety of loan sites through our network, we first put real focused effort on growing our personal loans business. Next, with small business loans, then student loans and credit cards, both organically and through acquisitions. Then followed by deposits, credit services and most recently, insurance.

These new product offerings have truly transformed the entire business, and on top of that, they continue to experience solid growth. 5 years ago, if someone told me that the mix of our revenues would flip flop from roughly 80% mortgage to roughly 80% nonmortgage and then be 5x our size, I would have had a hard time believing it myself.

Our diversified product mix enabled us to weather the storm of various market shifts and credit cycles in individual products. And with each new product offering, we're able to deliver increasingly more value to customers, engage with consumers more frequently and in new and different ways. Now, let's talk about what we've diversified away from, mortgage.

Clearly, mortgage will always be an incredibly important and meaningful part of our business. As I'm sure you're aware, the overall industry is struggling with higher interest rates, rising home prices, low housing inventory and declining volume. But we are working closely with our lenders to ensure that they can navigate this market profitably.

Those longstanding relationships are one reason why our mortgage business continues to do so well considering the industry headwinds. Even the mortgage revenues are down sequentially, I'd like to go through some of the reasons why I remain very optimistic on our mortgage business over the intermediate and long term. Obviously, the pool of borrowers that can benefit from refinancing changes with interest rates. And because consumers enter the market at different times, that pool of borrowers that can benefit also fluctuates.

According to industry estimates, the pool of homeowners who would qualify for and benefit from a refinance is at its lowest point since 2008 with only an estimated 1.5 million household that fall into that category. However, as we work towards fully understanding the customer journey, we're finding our presence in mortgages actually driving traffic and revenue to our other loan products.

Given rising interest rates, many consumers who initially come to LendingTree for a mortgage aren't seeing the financial benefit of a refinance and thus, are increasingly fighting their way to another LendingTree product.

In fact, since 2016, the likelihood of reengagement has, by most measures, internally doubled. Just this year, the percentage of consumers who initially shop with LendingTree for a mortgage and then reengage with LendingTree on a nonmortgage product in the same quarter is up 53%. We also track the unique behavior of a mortgage customer who filled out a form and found multiple matches versus the very different behavior of the borrower who is not able to find a match given their weak credit. And the good news is that both of these cohorts are finding their way to other LendingTree products.

Additionally, we're making great strides on our new mortgage experience. When we first began testing on the new platform, we focused only on refinance. In the third quarter, we launched purchase on the new mortgage experience. We're releasing new features every single day aimed at helping consumers and simplifying the process. Additionally, we have a robust pipeline of more than 20 lenders in the queue, and what was especially encouraging is that we're now able to track new types of lenders who historically have not been able to effectively operate on comparison-shopping platforms, including big banks and new mortgage companies. And the rise of the fully digital mortgage companies are also seeing great success on our new experience as well.

We've increased mortgage traffic to the new experience now at approximately 6%, which cost roughly $1 million of adjusted EBITDA per month because of the difference in monetization. And as we continue to optimize and enhance the experience, we'll ramp up traffic as we improve monetization. Overall, I'm very excited for this game-changing experience and believe this will transform the mortgage experience for both consumers and lenders on LendingTree.

Finally, despite the third quarter challenges we faced in mortgage and the seasonality we expect to come into play in Q4, we're also seeing some signs of life in October that are very encouraging. Considering the consumer engagement and the traction with the new mortgage experience, I'm looking forward to our opportunities in mortgage over the next few months and we will go into greater details during our Investor Day in December.

Moving on to M&A. Since 2016, we've completed 8 transactions for a total consideration value of just over $680 million, including potential earnouts. Five of these acquisitions have been for less than $40 million. Prior to QuoteWizard, which just closed yesterday, our largest acquisition was CompareCards in November of 2016. This transaction was critical to our diversification strategy. In many ways, QuoteWizard is very similar. It gives us a strong presence in an important and large category, and it was evaluated using the same approach we applied to each opportunity. We take a meticulous strategic and disciplined approach to transactions, and we are always looking for revenue synergies and ways to strength in both the platform and the consumer offering. I'm incredibly proud of the team and what we've been able to accomplish in this area.

Next, I'd like to touch on the progress we're seeing with My LendingTree. We now have over 9 million users and the contribution from this product continues to climb, growing 68% year-over-year. We continue to improve our alert functionality and our feedback from consumers continues to improve. We are enrolling new customers from opt-ins across the LendingTree platform, have ramped up app installs to over 8,000 a week and have a robust pipeline of syndication deals very similar to our deal of H&R Block. Overall, I'm thrilled with our progress on My LendingTree.

And now I'd like to turn the call over to J.D. for more details on our financial progress.

Thanks, Doug, and thanks to everyone for joining this morning. With Doug having given his thoughts, I'd like to provide further color on our financial results and some additional context on our guidance for the remainder of the year.

As Doug said, our third quarter results demonstrate many of the same themes we discussed in the second quarter. The macro pressure facing mortgage and more specifically, the margin pressure felt at our lender partners, remains persistent. But despite the sustained pressure in mortgage, the overall business continues to perform incredibly well as our nonmortgage category scale.

Margins expanded significantly, and we once again grew variable marketing margin and adjusted EBITDA by more than 30%. Total revenue for the quarter of $197.1 million was up 15% year-over-year. While a 25% decline in mortgage revenue weighed on overall revenue growth, our collective nonmortgage revenue grew 45% to $141.8 million and now accounts for 72% and importantly, more than 80% of total variable marketing dollars.

Several nonmortgage verticals produced standout performance in the third quarter. First, our personal loans business generated $38.6 million of revenue, up 52% year-over-year. While this is a category that is certainly benefiting from growth in the end market, the fundamentals of the business continue to improve as we see increasing demand among both the newer entrant nonbank lenders and traditional banks.

Although we have seen reports of certain lenders citing credit concerns and moderating their growth expectations, the aggregate demand among our lender network is as strong as ever.

Second, you may recall that after a challenging second quarter, we indicated that we saw some signs that our credit card business was stabilizing. But we're happy to report that revenue and contribution from cards rebounded nicely, growing 8% year-on-year and an impressive 10% sequentially to $42.7 million. Our efforts to diversify our issuer base and align with those partners during the first half of the year are providing for a more stable and predictable revenue stream, and we're starting to introduce more innovative ad units beyond traditional cost per approval arrangements.

Third, our other category continues to grow in both revenue and contribution. In fact, Q3 was the first quarter in which other, the aggregate of those businesses aside from mortgage, card and personal loans, was larger in both revenue and contribution than any of those large businesses individually. Other in total grew 84% year-on-year.

As Doug pointed out, our diversification has been facilitated through both organic efforts and acquisitions. Most recently, we're happy to report the acquisition of Student Loan Hero, and we're pleased to report that the early results from our now scaled student business were very strong in Q3.

The traditional in-school student lending business is very seasonal and Q3 is critical. We are confident that the acquisition of Student Loan Hero helped our already strong SimpleTuition business execute in Q3. And all indications are that Student Loan Hero should benefit materially from being part of the LendingTree platform.

Small business deposits and credit services continue to be standouts among our nonmortgage category. And while these are areas where we had made acquisitions, we were in 2 of these 3 categories prior. Most of our acquisitions have been small, as Doug pointed out earlier, but they've helped us to scale and in turn, become more critical to our partners. From there, we execute a playbook. We go deeper with existing lenders and partners, expand the network and unlock incremental traffic sources.

Finally, let's discuss mortgage. Revenue of $55.3 million was down 25% compared to an exceptional third quarter last year. It should not be a surprise that the decrease was entirely driven by softness in refinance activity, where industry originations continue their decline. In this difficult environment, we're focused on maintaining healthy relationships with our lenders, many of whom are struggling. We're focused on lender economics, and we're consciously optimizing our marketing efforts to deliver high-quality traffic for our lenders. At times, to the detriment of increasing volume. While the current environment is certainly a challenging one, we are encouraged that the strength in our other products are enabling us to whether this period while staying focused on improving our mortgage offering and continuing to deliver results for shareholders. Now let's move on to margins, which are the story once again this quarter.

As we've been saying consistently, we run the business to optimize for variable marketing margin dollars and grow adjusted EBITDA. In the third quarter, we delivered $76.8 million of VMD, up 30% year-over-year. Even including the expensing of a series of off-line advertising test run in the quarter, our variable marketing margin, as a result -- or pardon me, as a percent of revenue improved to 39%, the highest such measure since the first quarter of 2015.

While we aren't managing to the percentage, you can appreciate that our efforts to drive more traffic from organic or near organic sources are beginning to really materialize, and we're clearly benefiting from the continued expansion of our product offerings. Most importantly for shareholders, adjusted EBITDA grew 31% to $45.3 million. After a few quarters of accelerated headcount growth to scale the business, we are returning to demonstrating operating leverage in the portion of the cost structure beneath variable marketing expense. From a GAAP perspective, net income from continuing operations came in at $28.4 million or $2.05 per diluted share, and adjusted net income per share, which excludes certain items expensed under GAAP, was $1.92, up 64% year-over-year. With that context in hand, let me provide some color around our revised guidance for the remainder of the year.

With QuoteWizard just closed yesterday, we are layering some upside on to our preexisting outlook to account for the 2 months of impact the deal will have on our reported financials. With that, we are increasing our full year revenue guidance to $765 million to $775 million. This reflects softness in the mortgage business coupled with seasonality, offset by an estimated contribution from the new insurance vertical.

VMD is now expected in the range of $283 million to $288 million, up from $275 million to $285 million. While mortgage continues to present challenges in the top line revenue, we remain confident in our ability to generate VMD at levels consistent with what we promised all year. And adjusted EBITDA is now expected to be $152 million to $155 million for the year, an increase from $148 million to $152 million and now representing year-over-year growth of 32% to 35%.

Having just closed the acquisition yesterday, we're not in a position to provide a great deal of context on insurance today, but we look forward to doing that and updating you on -- updating you all on our outlook for 2019 at our Investor Day in New York on December 4. With that, I'll hand it back to Doug.

And with that, operator, let's open it up to questions.

(Operator Instructions) Our first question comes from Mark Mahaney of RBC.

Two questions, please. One, can you talk about the credit card segment. It looks like that recovered a little bit in Q2, but -- I'm sorry, Q3, but talk about that going forward with the sustainability of that recovery that you saw. And then in terms of the Q4 guidance. Did you just say how much of that guidance increase is simply due to the acquisition of QuoteWizard? Or is it -- is the organic -- is there an organic reduction in revenue kind of offset by that increase? Just quantify the QuoteWizard contribution.

Sure, Mark, it's J.D. I'll start with credit card. As you recall last quarter, we talked about some signs of stabilization. And we in fact broke it down by month, and we indicated that May was really the difficult month and that we saw some signs stabilization in June and as we began the third quarter. Simply put, that played out. One of the things we talked about was we were getting closer with our issuer partners. We're happy to report that only have we gotten closer with many of them and seen expanding wallet share there, but we're actually -- we've actually grown the issuer network in the third quarter as well. And then the economics have just improved. We've done a better job managing the marketing mix. Keep in mind, there are 2 ongoing transformations since the CompareCards acquisition. One is the diversification of issuers and the other is layering on different marketing channels into our card mix. And so we're seeing real benefit from both. It's driven by both. We talk about the sequential growth, the contribution in the quarter, just a sequential contribution improvement was in excess of 30%. So card really did deliver for the bottom line in the third quarter, and we're excited about that business go forward. Now, that's against the backdrop, by the way, where when we talk about this whole mix between reward cards and balance transfer, it is not -- in the broader environment for card, we've not seen in the channel the balance transfer cards become prominent again. So we're executing in a somewhat challenging environment for card and the growth that we're delivering, the sequential growth that we're delivering is a function of execution, not the external environment. We're not yet seen the issuers come back with balance transfer cards. So that's the card business. And we're encouraged that if we can execute in that environment, at some point, those balance transfer cards are going to come back. We get paid more for those. They're more valuable for the issuers. And so we're just going to continue the playbook of expanding of the network, improving the marketing mix and being there when that market recovers. But if we can deliver growth in this environment, it's a pretty good indication for 2019. So that's card.

Your second question was with regard to how much of our contribution. We talked last -- when we announced the acquisition, we got asked about seasonality in insurance. It does not have unique seasonality. But like our business, November and December are always months that were somewhat conservative with projections. So we're getting 2 months of QuoteWizard. We've layered on, I think, appropriate upside effectively to our full year plan. We modestly adjusted for mortgage downward on revenue only, not on VMD and EBITDA, just to be clear. Just like last quarter, we can deliver the bottom line, but we did adjust the revenue guide for mortgage modestly. And the reason for that as you look at the aggregate year, look at the fourth quarter, it is the most significant decline in refi. It's expected to be down 38%. So in that environment, we thought it was appropriate to take the revenue for mortgage only down just modestly.

Our next question comes from Nat Schindler of Bank of America Merrill Lynch.

Can you help me out on the mortgage business? I mean, the gap between you and the industry looks like it's narrowed in this quarter. Was that conscious on your time and that you decided that it wasn't worth as much to fight in refi if it wasn't going to do as well, so you shifted marketing dollars mid quarter? Or is there something more fundamental going on?

Yes, so it's a great question and everything we do is conscious. And we've talked before about the flywheel, and this ultimately boils down to what we call, CPL and RPL, or revenue per lead and cost per lead. And we just -- in the mortgage environment that we have, where lenders are -- where there's not enough refinance volume as they're switching over to purchase, we adjust our marketing spend to be in tune with whatever those revenue per leads are. In a purchase -- it purchase customer, if you remember, monetize about half of the refinance customer. So you drive more purchase volume from organic sources, SEO, and just people knowing about LendingTree TV, et cetera. And so that's really the switchover that we're seeing. So I wouldn't take any alarm with it because we're basically just maximizing our VMD every single day. And as I said in October, we're seeing unit revenue improve, which is giving us a lot more confidence going into Q4. We're also seeing, importantly, the cost per lead is coming down as -- we always say there's 2 sides of this equation; as mortgage companies increasingly focus on their most profitable channels, they're going to be doing more business with LendingTree and less direct marketing on search and other things, so that helps out our marketing expenses.

Great. And also just a quick follow-up. Can you break out the various growth rates between purchase and refi as you've done in the past? Is that possible?

It's possible, but I don't know that we've done that for J.D. Do we...

No, we're not -- yes, we're not breaking that down. We've talked about our revenue relative to the industry, but the actual refi activity...

But the decline in the mortgage business -- the revenue -- the top line decline in mortgage is 100% refi driven and offset by some growth in purchase.

I mean, as we've -- and that's, I guess, one of the things we talked about already. The purchase business is just -- it's a harder business. we get paid less for it because of conversion rates, as we've talked about in the past. Right now, we're in an environment where that refi activity is de minimis. And so we're having to execute in a lower margin product effectively. Now, just like we talked about at the end of Q2 in card, we're internally seeing some signs, as Doug points out, particularly in the cost side in mortgage. They are beginning to be encouraging. But that's against a backdrop where we're expecting a 38% decline in broader refi activity in Q4, coupled with Q1 being a tough comparison for mortgage. As you remember, we were able to drive RPLs double digits in Q1, that was the last quarter where we had RPL expansion. So we're seeing good signs internally in mortgage. But only internally, the cost equation is getting better and that's great, but we're going into a tough -- we're in a tough fourth quarter for refi, and we do have one difficult comparison ahead of us in Q1 of next year. But importantly, those initial signs that the mix between RPL and CPL is improving are there.

And the only other thing I'd add, increasingly over time, we're going to be looking at the total platform revenue, the individual products are important. But as I said earlier, we're seeing a lot of crossover from mortgage where people traditionally may click on a mortgage ad, then they come in and they see there's no benefit. They pop over to a personal loan, et cetera. And then the other thing I would add is that the new mortgage experience completely changes the game on conversion rates as we improve that monetization. And just one encouraging sign, we're seeing lenders locking loans at about a 4x higher clip on the new mortgage experience than the current mortgage experience. And the Net Promoter Scores are very, very high on that. So as we work to sort of automate the mind of the loan officer on the site and we get the monetization equal, then that's going to really the change the game in mortgage business and hopefully give us a new leg of growth.

Our next question comes from John Campbell of Stephens.

Doug, you mentioned a flipping of mortgage to about 20% of rev. Clearly, you guys have the addition of QuoteWizard that's going to, I guess push the mortgage mix shift a little bit lower. But just looking at the forecast, if we stuck with that, if we just kind of went with that 20% mix, I'm thinking you might have to see mortgage down again next year. I know you guys mentioned a tough comp in 1Q of '19, and I'm sure you guys will talk about this more at Analyst Day. But am I thinking about the phasing of that mortgage rev right? And may -- is it pretty difficult to grow mortgage rev next year?

Yes, we are definitely planning on growing mortgage next year. And the percentage mix just changes based on the individual growth rates. Some of the other businesses are growing faster, but mortgage, we absolutely expect to grow significantly next year.

Okay, that's great to hear. And then on the broadcast spend, can you talk just kind of broadly how that looked year-over-year? And if you guys maybe intend on stepping on it, accelerating a little bit more as you get into next year?

Yes, I'll hit it at a high level and then let J.D. comment as well. Our broadcast spend, we've been testing. We have been running a decent amount of TV. We're also running ads now increasingly on our nonmortgage products, particularly credit card, including some ads on the CompareCard brand, which are either running or they're in process. And so year-over-year, marketing spend off-line is down because of the mortgage environment, you just don't want to market into lower revenue per lead if you can't do it profitably. But yes, we have a -- we'll talk about in December, we have a significant off-line spend anticipated for next year, and we expect we'll be able to do that very profitably.

Yes, John, all I'll add is interestingly, I mean, clearly, off-line spend is down in 2018. It's down in part because they're intentionally, as Doug points out, in mortgage specifically, the RPLs didn't necessarily justify that TV spend. But it's also down because we've been going through an evaluation of how we should be spending those dollars. We've got a new logo that you might have noticed. We, as Doug points out, are going to advertise not just broad LendingTree, but specific products. And in Q3, it was actually up meaningfully relative to Q2 because we were testing. And so that's what we were referencing. We're testing different ad units in regional markets. So in Q3, it was actually up, but in a testing format. We will roll it out more specifically at investor Day what our intention is, but it should be up meaningfully.

And the only other thing I'd add as well on the marketing flywheel effect, we have built over the last couple of years a significant SEO business that's now tracking roughly 20% of our revenue, and that's from almost zero a few years ago.

Our next question comes from Jed Kelly of Oppenheimer.

Yes, just following up on the advertising discussion. Online advertising, I guess, it continues to actually grow faster than your revenue this year. Does that become harder to leverage in 2019 as you stop -- as you start to comp some of the lower, I guess, the lower expenses or the lower spends you made in broadcast TV? I guess, how should we think how you're optimizing for your online advertising spend?

So the online advertising spend is optimized in minute, if you will. So it's in real time, we're constantly looking at the supply and demand equation and marketing up to the last profitable dollar, anywhere we can find it. So you're looking the demand from your lenders, looking at the availability of inventory, making sure you can do it cost effectively, and you optimize the online stuff in real time.

The off-line spend this year, we've done a lot of analytics and data tracking. It does make money over the period of, let's say, six months, but you sort of go negative and then you go positive from any amount of ad spend, and we can draw those curves up fairly precisely and it gives us a lot of confidence to go to market into next year, particularly as you're getting increasing monetization from My LendingTree and people are buying multiple products, that does nothing but improve your lifetime value and gives even more juice to go market against.

And then on the digital mortgage, can you give us an update just how consumer throughput is doing? And as we move through a purchase environment, how much outreach is going to be -- have to be done on your part to educate lenders on how to manage leads and better drive leads on -- I guess, on a product that's harder to convert or/and has a longer sales cycle?

Yes. So on mortgage, could you repeat your first question, Jed?

How's the consumer throughput on digital mortgage system?...

Okay, yes. Got it. Yes, so the digital mortgage experience. So first off, LendingTree in the new model takes a consumer from an inquiry, which is filling out the form, thinking about like a search query, shows you the results and then LendingTree helps you make a selection. Once you select a lender, you either deal with that lender, "manually" even though manual today is very, very automated, or there are some lenders, Better Mortgage is a good example, who have a nonhuman touch, fully digitized experience once you fill out the actual application and then actually lock online. So the advent of the digital mortgage experience, whether it's the Rocket Mortgage of Quicken Loans, the things that loanDepot and many others are doing and that technology is becoming more and more available. But it's really sort of a click over to a fully digital application. Think of it like the personal loans product. It reduces a lot of friction. And we're helping lenders get increasingly automated.

To your second point, the notion of touch points, particularly on purchase, think of a -- and you've heard me talk about the mind of the loan officer. The log and experience of LendingTree needs to over time emulate what a loan officer would talk to you about. What are your goals? Let's see what products we have? How long are you going to stay in that house, et cetera, et cetera. With purchase, you layer on the fact you need to keep a realtor in the loop and the time lag, between the time the customer comes in and the time they ultimately close could sometimes be as long as six months. During that period, you need to incubate them. And we do that mostly through technology. We were very -- we had to do this when we own LendingTree loans. So we're bringing those -- that muscle tone back so that we know how to interact with customers. But most of it we are not doing over the phone, most of it is coming through text, e-mail and then the online experience, plus alerts telling you that rates are changing, et cetera, et cetera.

So as that incubation process gets better and more automated, we think -- we actually think that purchase will be easier in the new mortgage experience than refi because that -- because, basically, we'll be running the same incubation process across all of our lenders, and we'll be doing that once as opposed to each of our lenders doing it 5x.

Our next question comes from Mike Grondahl of Northland Securities.

Two quick questions. One is how is the home equity business doing? And secondly, outside of mortgage, could you rank your sort of products from best visibility to maybe least visibility?

So let me comment on home equity. The home equity business is doing fine, but I would say it has not reached the running speed of several years ago. And I'll talk about why that is. But that's starting to change. It -- before 2008, you had lenders, mostly banks, doing home equity and loans and keeping them on their balance sheets, and they did them in a highly automated way with drive-by appraisals and most of it done online. Very, very high conversion rates, just like we see in personal loans. The banks have not yet brought that process back and there is not really a liquid secondary market for home equity like there used to be pre-crisis. So with that, the home equity business is growing more slowly. But as technology automation happens and we've got some exciting things on the docket for next year, then we get the RPL up and then we can market into that.

The other thing I would say about home equity, you also get a lot of business where people come in for a home equity loan, then they can get a full refinance on their home. So you also pick up some refinance business through home equity. But overall, I would say we're waiting for and helping the automation to home equity to happen so then we can then market into it.

Yes, Mike, it's J.D, all I would say is we've, traditionally, called out home equity, but it wasn't about calling out home equity, it was taking the nonmortgage category and away from those individual businesses, like card "PL" ( Personal Loan) that are big enough that we need to identify the absolute dollar number or dollar amount. We obviously, for competitive reasons, we don't want to give individual business scale, but we do want to give you and investors a sense for what's driving the broad -- increasingly broad nonmortgage category. And so we always highlight those that have contributed most, and that's why we've, in this quarter, home equity just wasn't among the top 3 that we pointed out through in excess of 100%. The business is fine. But the percentage growth rates are not as overwhelming partially because a year ago, that business was probably inflated by lenders who were buying home equity leads to convert them into a refi product.

And the only other thing I'd add on product visibility, it's kind of interesting, as we got ready -- as we're getting ready for a new TV ad campaign, obviously, we do a lot of research. Consumers still think of LendingTree as primarily a mortgage business, as a mortgage comparison shopping service. And we are going to change that next year with our new ad campaign, and I think it's all upside. If people are thinking about us for mortgage and it's not a significant part of our business, once they realize, "Oh, wow! LendingTree all does all of that, too?" We get much more enthusiasm about the brand and we can tell that story very easily through advertising.

Our next question comes from Michael Tarkan of Compass Point.

Just a technical one here. I saw on the VMM calc on the back page, you had a $3.6 million cost of advertising resold to third parties. Can you just provide a little color on that and if that flown through to revenue in some form or another?

Yes. No, absolutely. So we've talked in the past, over the last year, about business development partnerships. And so on the media side, we own a fair -- a good amount of inventory with CNN and MSN that we use for our own properties. And periodically, we decided to resell that inventory. And so rather than -- so we classify that differently from accounting perspective as cost of revenue rather than traditional classification for our own business. So we're just making -- drawing the distinction. So that's when we sell it to a third-party and it's not a LendingTree sale, but sold to another publisher.

And the way -- conceptually, if you think about it, basically, we've done a partnership with a partner. And instead of running LendingTree a parent, LendingTree ads or units or rate tables constantly, you reach a point of we're out with the consumer and then we ship that inventory off to other ad buyers through all the ad networks in a highly automated fashion. So it's -- and over time, as our other products, as that inventory can be better spent on LendingTree products, we will simply do that.

Is that activity that we should expect to continue? Or is this sort of a one-time situation or just sort of temporary? How do we think about that sustainability of that?

I would think about it in terms of just revenue. Because if we weren't reselling that inventory, we'd be running a LendingTree ad and it would be a LendingTree revenue. So it will fluctuate up and down depending on those deals. But I would just -- it's effectively taking revenue from one LendingTree product and sticking it in another one. So I wouldn't think of this like a huge growth engine of the business. It's more of just a supplement or a substitution for other LendingTree product revenue.

Okay, great. And then a question on sort of pricing versus volume. In the past, you've talked, I think, directionally about how mortgage you had pricing power and then you're sort of flatten out. Is that eroding a little bit on the pricing side? And then same question on personal loans. Are you still able to take pricing up on your lender base?

Yes, let me take mortgage. I'll give J.D. personal loan. So pricing power, if you think about it, really comes from lenders' conversion rate. So they put in effectively a bid or what they're willing to pay for given a customer introduction or a lead. And we've -- and that's where the pricing power comes from. And in October, we've actually seen pricing move up with some of our significant lenders because conversion rates are better. They've worked through some of the -- the switchover cost from refinance to purchase. And they're able to convert better. And because of that, they upped their bids. Because of that, we can then market into it. So we're actually seeing pricing strengthen in Q4 over that.

And by the way, the other nice thing about that, that tends to stick as you go through another cycle. So if you think -- put yourself in the mind of a lender, you're cutting every unprofitable marketing channel back. And your LendingTree channel, which is generally sustainable and very profitable for lenders, and they can toggle the volume up and down, that tends to be the last channel they turn off. And then not only that does help our revenue per lead, but on the cost per lead side, as I mentioned, as lenders pull out a direct advertising, that improves our economics online as well.

Our next question comes from Youssef Squali of SunTrust.

Maybe you can just speak to the -- your guidance or particularly, what the implied VMM growth relative to revenue growth. Looks like maybe at the midpoint, it's about $79 million less of VMM for Q4. Maybe you can just help us think through that and also, the contribution of QuoteWizard to VMM, if you can help share that.

Sure. We've seen -- we obviously have reached a recent peak year in the third quarter with respect to VMM percentage at 39%, as we pointed out. QuoteWizard operates at a very similar margin profile to LendingTree businesses. It's one of the things that attracts us to it. We'll get 2 months of it in the quarter. Our -- as I pointed out earlier, we did -- the only number -- aside from QuoteWizard, the only adjustment made was for mortgage revenue, not for VMD in that guide and not for EBITDA. So we'll continue to see strong VMD and EBITDA. There shouldn't be a huge differential in terms of that growth rate in Q4 and QuoteWizard's contribution is similar with respect to percentage.

I think on a percent basis, at the midpoint, there seems some deterioration on the margin side. Is that just seasonal? Or what's going on there?

There's always a little bit of seasonal in there, yes, every quarter. That has nothing to do with QuoteWizard and we're operating off of it. If you want to talk about sequential, we're coming off of a 39% peak number there.

Okay. And then just still on the cost side, can you maybe shed some more light on the cost of revenue jump in the third quarter? And how we should think about it going forward?

It's what we just discussed on the advertising side that Doug just went into detail on, the cost of revenue increase.

Our next question comes from Stephen Sheldon of William Blair.

So you've gotten a lot of questions or mortgage. But just given the degree of the decline this quarter, I just wanted to ask him if anything has changed in your view within the competitive environment that has had any impact there.

No, I don't think it has. We -- obviously, we're in the same, I would call it, same competitive environment with the other aggregators. We continue -- we believe we continue to take share from -- on the lender side. We continue to see great growth there. And so I have not seen -- we have not seen any significant competitive pressure. There's always innovation in the business. But the good news there is the tough mortgage market benefits LendingTree, I would say, disproportionately than others because we still have the ability to go out in the market because of our deep lender network. And then, as I said, once we got our new mortgage experience up and running, that's a game changer with respect to capacity. So for example, our lenders will see roughly -- we'll see many fewer leads coming in the front door, but they will be much more highly qualified, which means they will open up the floodgates, increase demand and then we'll be able to market into that. And that's why we're putting so much effort on that new experience.

Got it, that's helpful. And then I know you'll provide more at the Investor Day. But just at the high level and more qualitatively, how are you thinking about adjusted EBITDA margins heading into 2019, kind of excluding the impact from QuoteWizard? Would you expect some pull back next year given the boost you've gotten this year from lower variable marketing expense as a percentage of revenue, which could go up next year and a potentially more favorable demand environment? Or will the leverage from headcount additions this year and marketing efficiencies maybe let adjusted EBITDA margins still trend up some next year?

Yes. So Steve, it's a good question. I think we're certainly not -- we will definitely get some operating leverage, which will be great. As we pointed out, we expect mortgage to grow. We -- but this is not, when you look at our margins being excess of 20%, they may go up modestly, but that's not a deliberate strategy. You shouldn't -- we get asked the question all the time, "What's the natural margin of the business?" I don't think that what you're seeing this year is necessarily. Something that you should pencil out as an improvement into next year. We're still very much in market share gain mode. And so we're going to go after dollars, as we've talked about. That may not necessarily translate into -- we're going to operate the business in the same EBITDA margins that we have for some time. So we'll get some operating leverage, sure, but we will be in growth mode and our margins will be what our traditional margins have been. Part of this is that we've obviously made marketing decisions in the light of the macro environment in mortgage. And so that is resulting in a year in which maybe the top line growth is not as strong, but the margin expansion is there. We look forward to getting back to an environment where we're going to see top line growth.

And by the way, that is -- just to add on to that, that is a -- you'll see, if you look at LendingTree over the years, you will see some variation in margin percentages as the macro environment changes because you only -- because you market up to your last profitable dollar. And however, the VMM dollars and EBITDA dollars keep climbing through that. So that's -- those are the numbers that I tend to look at because individual percentage margins can vary based on channel mix and individual demand and individual products.

Our next question comes from Eric Wasserstrom of UBS.

Just one more question on mortgage, which is similar to the other questions I asked last quarter, which is just about the relative depth of this transition for the lender base versus prior cycles. And I think your response last quarter was that it's deeper and therefore, maybe longer-lasting. And I just want to see if anything about that perspective has changed.

Yes, I don't -- the longer-lasting, I'm not sure I would say this is definitely, "deeper" given the statistics I put about consumer benefit if you come in for a refinance and you can't save money either in a lower payment or, I guess, you can always -- most of the time, you could say by lower payments stretch out your term probably not the best consumer idea. But if you don't have a financial benefit, obviously, this doesn't make sense to refinance and those numbers have been weighed down.

The good news, though, in mortgage, and I've seen this play out over 20 years, it's not the absolute value of interest rates that drive refinance and purchase volume, it's more the refinance line, it's the rate of change. So you can have low interest rates and when the 10-year knocks down a little bit, you get a flood of refinance volume because consumers can save money from the old -- from there now, "high rate." So even historically low rates, you can still get good refinance volume as those rates bump up and down and you just adjust your marketing mix. I've said before, the mortgage business function is very much like the hotel business and travel lenders will fill up capacity. They will -- if they've got excess capacity inside of their shops, they want more volume. If they can't do it profitably, then they keep their bids down. But as they improve their conversion rates and their technology, those bids move up us they're doing in October. And then as I said, we market into it.

So just in terms of the ending of shakeout, what would be your assessment?

The ending, I mean, timeframe or...

Yes, like the -- I guess, the baseball reference to like which inning.

Good question. I have given up predicting interest rates. And so I think I'm going to maintain that. I have absolutely no clue. All I know is that the playbook for us changes from one interest rate cycle to another, and we can -- we flipped the switch very strongly. But if you got to read on where rates are going, please let me know because advance warning always helps.

Eric, the only thing I'd add to add to that is some of the new experiences we're going through should change who on the lender side benefits and can operate on the network. And thus, we don't -- you don't have to call the bottom of the cycle for our traditional lender client. We should benefit from that network expansion before that occurs...

Right, got it. So you abruptly anticipated on the operating...

Your next question?

Yes, my next question, which was then, what is it then that's driving the expectation about next year? And it sounds like it's more of a transition that you put in place, it's not so much an expectation about a bottoming of the broader cycle, is that correct?

That is correct.

We typically set on mortgage plan based on what the MBA looks like and we better set the numbers. But the VMM can grow through both cycles as long as we continue to make conversion rate improvements. And then to J.D's point that you just said too about new types of lenders and broaden base. As we switch over the mortgage experience, it levels the competitive playing field on the network side, and that's basically what you see as these less automated lenders can actually compete against the Quickens and the loanDepots, et cetera, who have highly automated factories, been doing this for 20 years. And then that helps to improve their conversion rates, which really improves aggregate conversion rates. Because it's very similar to the long tail effect that you saw with search engines. And the more you can boost up the lenders who are in sort of that long tail, you then improve your unit economics overall.

Our next question comes from Kunal Madhukar of Deutsche Bank.

Question on My LendingTree and the growth there slowed down on a quarter-over-quarter basis. And the comp was tougher, but the annualized revenue per My LendingTree account declined significantly sequentially. What's behind that? Is there some seasonality there? Or is that -- the promotions are not as attractive?

Oh, no, I think -- I mean, our growth rate declined from, I think, like 100% to like 70% sequentially. And I think it's simply more of a factor of your -- we're marketing less. So therefore, you've got less traffic flowing through the system. And a lot of our traffic -- a lot of the My LendingTree sign-ups are coming through, sign-ups from the core LendingTree platform. And so that growth rate is -- we're incredibly happy with it. And then as the monetization improves, you can continue new marketing. As I've said, we've got a really, really robust pipeline of syndication deals and expect to hear more from that in the coming weeks and months. But no, I'm very, very encouraged about the progress we've made with My LendingTree and I wouldn't swipe the growth rate difference.

Yes, And Kunal, the only thing I'd add to that is just take a look at -- the thing that we're encouraged by is the app downloads. And if you look at up our presence in the App Store that's improving, the downloads are improving dramatically. The quality overall of the people opting in. We -- what we want this to ultimately be is not just a base of members, but also people engaging in the app. And so from a quality perspective, we think it improved dramatically in the quarter.

And a quick follow-up on the guide. And I know you've discussed the revenue guide previously. But by my math, and my math maybe off, I'm getting like a mid single-digit pro forma year-over-year revenue growth for the fourth quarter. Is that the trough that we should expect?

From a growth rate perspective? Yes, we certainly do not -- that is up against the difficult comparison from the fourth quarter a year ago with a very robust mortgage business. We certainly are not operating a -- that type of growth business in aggregate. So in that respect in terms of revenue growth, yes. It is unique. It is a unique comparison both the third and fourth quarter are tough comparisons when you consider the change in the revenue days for the mortgage business.

Our next question comes from Rob Wildhack of Autonomous Research.

J.D., from your commentary earlier, it sounded like we could infer that variable marketing dollars in mortgage were up year-over-year. Is that correct?

Variable marketing dollars year-over-year, no. We've seen -- I'm not sure which commentary you're referring to, I apologize. Which statement are you referring to?

I mean, I'd have to go back and check, but we can do that off-line. Maybe more broadly, I wanted to ask about the sale of ad inventory. Do you consider that to be like a, "lever" that you have when flexing marketing spend? If you do, how far down the list of options is a decision like this?

Yes, on the ad selling, I would -- again, that is a sort of a substitute/complement, not really a lever. Basically, you do face syndication deal with, let's say, a CNN where we're going to put LendingTree rate tables, widgets, app download, opportunities, et cetera, et cetera, on a site like CNN. You put the LendingTree clickable units there until the last profitable dollar. And then whenever you can and then whenever it's not profitable, it's easy to sell that excess inventory out to third parties. So I wouldn't see it as a lever. And quite frankly, it's not something I necessarily really focus on because over time, as LendingTree monetization improves, those ad units will be absorbed by LendingTree.

And Rob, to answer -- to try to answer your first question just because we always want to do that. No, the dollars from mortgage in aggregate were certainly down. The percentage is flat quarter -- yes, from Q3 of last year.

Our next question comes from Hamed Khorsand of BWS Financial.

Could you talk about how fast you can onboard these acquisitions? Are they quickly generating traffic and revenue from coming on to your platform? Or are they still on a stand-alone basis?

Hamed, great question, I really appreciate that and it's something we probably should have focused on. We are able to, for the most part, predict what synergies we're going to get before we even do the deal because we can see what either changing the brand name is going to do or, for example, with -- we bought one SEO business and those guys are now running SEO across the whole company. So we're seeing synergies happen very, very fast. And I got to go hats off to our team here who really, really digs in everywhere, from HR, to finance, to operations. And I would like to call out the intrapreneur without the headaches. Every time we bring one of these guys in, we say, "Listen, we'll handle your capital. We're going to handle all those things that it takes to run a business from a corporate standpoint, and all you got to do is go grow your business." And then we put people. We have great integration teams. And, knock on wood, all of these acquisitions have been very, very accretive for us. You can't hit -- you can't get 1,000 in the M&A world, but I'm just thrilled with where our team has done it. And the synergies, if anything, I think we underestimated with some of our earlier deals.

Okay. And then as far as -- J.D. made a comment about peak margin on VMM. Is it becoming dilutive being in so many different products that you have to advertise each of them individually? And when do you start to do it in a more of a platform setting?

So yes, you try to do both. So for example, you might run an ad that features somebody buying a home or somebody getting a mortgage or somebody getting a credit card, but typically you try to also at the end talk about all of the loan types so you can bring in traffic that way. But no, you -- the more products you have, I'd like to call it, the more marketable events you have. And the more marketable events you have, the more you could optimize where you place those events across the Web or across TV. So the more products, the better. And we do have some ads that run where we talk about everything and you can expect to see some more of that. But you just basically do whatever works and you try to mix in individual products spots to tell a specific story, but then also try to say, hey, but we're also here for everything.

And Hamed, I'd actually say it's kind of the opposite. All of these nonmortgage businesses are benefiting from draft traffic. And so Doug may reference to it before in terms of somebody coming in from mortgage and going somewhere else. Like every Internet company, we're getting smarter in terms of tracking this and using data science to do that, right? So what's very evident to us is that all of these other businesses are benefiting from draft traffic, from not just mortgage, but from each other. And that's the benefit of an acquired company coming on to the platform is that they get that traffic. And so the aggregate brand is contributing to them in a pretty material way. And I think we'll get smarter about this over time and obviously deals with attribution models, and we've got to pencil it all out. But it's really exciting, actually, when you see what benefit they can get from being part of the platform.

So is your cost going up? Is that why you lower dilutive in your commentary or peak?

No, no, no. When we said peak, what we were referring to is the fact that VMM is 39%, was highest since, what, first quarter '15. That's all we're referring to. Now as you know, we manage the business for dollars. And so that VMM percentage may range from the low 30s to the high 30s. I'm simply referring to the fact that at 39%, that was a high number relative to recent history. There was -- I didn't. No, I'm not saying its dilutive at all.

Ladies and gentlemen, that does conclude today's question-and-answer session, I would like to turn the call back over to Doug Lebda, CEO, for any closing remarks.

Thank you, operator, and thank you all for joining today, and thank you for the really, really thoughtful questions. I just like to close with one comment. Around LendingTree, we always talk about trying to make sure we all think and act like owners. And we don't need to just put ourselves in your shoes. Everybody here is an owner in LendingTree equity, and we think about it in the same terms that you do.

One of the things that I think about as an investor that I'm just thrilled with as I look over the years, is the resilience of our business model. For those of you who've been around for a long time, you remember we sold our mortgage company. We thought -- people thought, "Oh, my gosh, the sky will fall in mortgage." We've had various mortgage rate changes over the years. Remember, a few years ago, the personal loan business was going to completely evaporate and go away. We had pressures in card or worries about card. And through each and every one of those, LendingTree has been able to grow through them all. Each time, it gives us a lot more confidence in our business and in our ability to execute. That we can not only survive, but also thrive through different macro environments. Our lender network is strong, our team is executing incredibly well, and I am very confident and optimistic about our future, and I look forward to sharing our long-range plan and lots of new information at Investor Day in a little over a month. Thank you all very much, and we'll talk to you soon.

Ladies and gentlemen, this does conclude today's conference. Thank you for participating. You may now disconnect. Everyone, have a great day.