Good morning, and welcome to The New York Times Company's Third Quarter 2019 Earnings Conference Call. [Operator Instructions] Please note, this event is being recorded. I'd now like to turn the conference over to Harlan Toplitzky, Vice President of Investor Relations. Please go ahead.

Thank you, and welcome to The New York Times Company's Third Quarter 2019 Earnings Conference Call. On the call today, we have Mark Thompson, President and Chief Executive Officer; Roland Caputo, Executive Vice President and Chief Financial Officer; and Meredith Kopit Levien, Executive Vice President and Chief Operating Officer. Before we begin, I would like to remind you that management will make forward-looking statements during the course of this call, and our actual results could differ materially. Some of the risks and uncertainties that could impact our business are included in our 2018 10-K. In addition, our presentation will include non-GAAP financial measures, and we have provided reconciliations to the most comparable GAAP measures in our earnings press release, which is available on our website at investors.nytco.com. With that, I will turn the call over to Mark Thompson.

Thanks, Harlan, and good morning, everyone. Well, Q3 2019 was our best ever third quarter for new, digital new subscriptions, indeed, the fourth best quarter in the history of our pay model and a very encouraging quarter for the company as a whole. We now have more than 3 million subscriptions to our digital news product, more than 4 million total digital subscriptions and just under 5 million total subscriptions. We also saw a number of other significant positive developments in the business, to which I'll turn shortly. We did slightly better than expected in digital advertising, though as we warned in our last call, Q4 is going to be challenging too. Overall company revenue grew by 3% in spite of the 7% decline in overall advertising revenue. The advertising decline combined with content and digital product development investments meant that adjusted operating profit decreased from $54 million in 2018 to $44 million. Let me begin with the digital subscription story. Overall, including M.I.T. Cooking and Crosswords, we added 273,000 net new digital subscriptions in the quarter, of which 209,000 were subscriptions to our core news products. That 209,000 was 46% more than the equivalent net adds in Q3 2018, which were themselves more than 1/3 higher than the adds in Q3 2017. So we're clearly seeing acceleration in our digital new subscription model. What accounts for it? Well, first, one of the liveliest news environments any of us can remember, after a quiet start to the quarter, there was a dramatic pickup in the news cycle in the second half of August and through September. The volume in intensities continued in the present quarter and look set to persist into 2020, which is itself, of course, an election year. The weight and complexity of the stories, impeachment, Brexit, Boeing…

Thank you, Mark, and good morning, everyone. As Mark said, we remain pleased with the progress we are making as we continue to execute against our strategy. Adjusted diluted earnings per share was $0.12 in the quarter, $0.03 lower than the prior year. We reported adjusted operating profit of approximately $44 million in the third quarter, which is lower compared with the same period in 2018 by approximately $10 million. Total subscription revenues increased approximately 4% in the quarter, with digital-only subscription revenue growing 14% to $116 million. On the print subscription side, revenues were down 3.3% due to decline in the number of home delivery subscriptions, a continued shift of subscribers moving to less frequent and, therefore, less expensive delivery packages as well as a decline in single copy sales. This decrease in print subscription revenues was partially offset by a home delivery price increase that was implemented early in the year. Total daily circulation declined 9.1% in the quarter compared with prior year, while Sunday circulation declined 7.9%. At the end of August, the Starbucks retail chain discontinued the distribution of all newspapers, including The New York Times, at its corporate-owned locations. This had a meaningful impact on copies, accounting for approximately 1 percentage point of the decline. Subscription revenue was immaterially impacted by the loss of this distribution channel. Quarterly digital news subscription ARPU declined approximately 11% compared to the prior year and approximately 3% compared to the prior quarter, as the impact from the large number of newly acquired subscribers, mostly on the $1 a week promotion, was significantly larger than the benefit from existing subscribers whose promotional offers ended and graduated to higher prices during the period. This quarter also included the conclusion of the 52-week promotional period for the first $1 per week cohort…

[Operator Instructions] And the first question comes from Alexia Quadrani with JP Morgan.

Just a couple of questions. Mark, thank you for the color on the subscription growth in your comments and the drivers behind that. That's very helpful. You mentioned that the news cycle has been elevated, and as a consumer, it definitely appears to me that elevated news cycle has continued into the fourth quarter. I'm curious if you would agree with that statement in terms of thinking broadly about the drivers of sub growth in the fourth quarter. And then, and, Roland, you talked about the natural moderation in ARPU going forward and all the drivers behind that. I'm just wondering if we could begin to see some evidence of that leveling off in the fourth quarter.

Well, I mean, the way I think about the news cycle goes as follows. If you look back over recent quarters, we've seen different levels of news. Right now, we're seeing a lot of both very significant running domestic stories, the impeachment process, the runners and riders to go on Democrat side in terms of presidential candidates and so on, and some very big international stories as well. But I want to say that over the last, let's say, 2 to 3 years, we've seen really very strong numbers coming out of our model and a background picking up of the numbers of net new subscribers we've been able to add to our news product, which has kind of pressed on, in a sense, through different levels and intensities of news. So although I think we are heading, it would appear -- you can't be certain -- heading, it would appear into a very, very busy news period. And as I said in my remarks, my sense, I'm a former journalist and editor myself, is this is going to blur into the election year. We're not that far away from Iowa and the start to the primaries process and so forth. I don't think our model by any means and the growth we're seeing entirely relies on news. I think that we've demonstrated very broad appeal for The Times, international appeal for The Times. And I think the underlying acceleration is not solely dependent on the news cycle.

So regarding the ARPU, I think these 2 questions are linked. And we expect to continue to acquire large numbers of new subscribers in the future. And again, as I mentioned, we plan to continue the $1 a week promotional offer because we believe it is working really well for the company. And so as that happens, the number of subscribers we bring on is expected to be much higher than the number of folks that are transitioning to higher prices, and the mathematical result of that is pressure on ARPU. However, as I mentioned before, we believe that, that decline will begin to moderate as more and more folks step up from that $1 a week to higher prices. And also, as Mark mentioned, we are planning a rollout of a price increase to tenured subscribers sometime in 2020 and that will also take some pressure off ARPU.

And then just a follow-up, a clarification. I think, Roland, you said higher levels of investment in subsequent quarters and costs, is that versus Q4, where it's more moderate or versus full year of 2019? I'm assuming it means it's going to be -- cost increase will be lower in Q4, but then kind of get back to elevated levels in 2020, but...

Yes, so it's the latter. The comment was relative to Q4, which is going to be a low -- a very low growth percent.

Yes. If I can just add, and you should have gleaned this from what Roland and I said, our view is we're building a machine, both a journalistic machine and a digital-product machine, which is scalable. And we do take significant comfort from the way we were able in Q3 to drive the numbers of new subscribers without increasing marketing costs from Q2. So you can begin to see there at least one part of the improving operating leverage, which we think our model will give us. We're obviously very focused on building a product and a set of capabilities, which enables us to continue to significantly scale the number of subscribers. But with the -- over time, the growth of costs beginning to kind of -- the gap between the costs and the revenue progressively widening as the model scales.

Thank you very much.

Thank you. And the next question comes from Vasily Karasyov with Cannonball Research.

Good morning. Congratulations on the good quarter. Wanted to ask you this question, early in the year, you were talking about increased focus on driving subscriber starts with product organically rather than paid starts. So I was wondering if you could give us an update how that progressed in Q3 and how you use it for different cohorts of subscribers. And the second part of the question, Roland, can you help us understand in more detail the impact on the P&L of paid starts versus organic? Is it just as simple as moving between marketing expense and production costs?

Sure, I'll let Meredith start and then I'll follow.

Yes. The simple answer to the first part of your question is, yes, I think what we've just seen in Q3 is that our thesis of the product itself is doing more of the work to get people to make a habit with us and then ultimately pay and stay is working. I think Mark and Roland just alluded to the fact that -- I think Mark said it in his remarks throughout the call -- that we had a higher percentage of starts coming just from the organic engine. So I think that's really beginning to work. I'll say there were three major changes that we've been working on for some time that began to land in the third quarter. The first one was what Mark described as the pay model change, where we now have the majority of people coming to us, being encouraged to register and log in to get to read more than a small number of stories. The other one is we're beginning to feel the effects of a restructure in our digital product organization that we did a year ago that I think is just bearing fruit now, which is just allowing us to frankly advance our digital products faster and more effectively. So more testing, more experimentation. And then the last thing I'll say is that the product experience itself, the way that people follow stories, engage with news, understand what is a brand-new story, what is an exclusive story, we're getting much better at how we do that between products and news, and I think we're just at the beginning of that.

So regarding how that shows up in the P&L, I want to point out there's probably two dimensions to this. One is kind of the simple dimension, which is, as we scale back on marketing costs and we continue to invest in our newsroom and we increase our investment in digital product development, you'll see some switch from costs that currently show up in SG&A to be much more heavily weighted to production. But the real difference that really is going to affect our P&L for years is that marketing -- direct marketing expenses drive starts in the period that you spend the money. Investment in product and journalism drive starts for multiple periods. As we make the journalism better, there's more journalism, as we have a better digital manifestation of that journalism, as we have better business rules and an access model around that journalism, we believe that will drive starts for many periods to come. And so if you think about the three areas we repeatedly talk about investing in, which is the journalism, the digital product and the marketing, we've also talked about the gearing of how we would grow those areas changing over time, and we believe we're starting to cycle out of a high-growth period of time for marketing and ramping into a higher growth period of time for our product -- our digital product development, with a continual investment in our journalism.

And the next question comes from Craig Huber with Huber Research Partners.

Mark, you mentioned early in your prepared remarks that the new cycle was significantly higher, I guess, in the second half of August and into September. Did you see a commensurate pickup in your digital subscription growth in that period versus, let's say, the first half of the quarter was?

Sure. And perhaps, Craig, it's worth saying first of all that the idea that you have a relatively quiet kind of early summer with a kind of back-to-school intensification of news as you head past the middle of August, that's a, I would say, there's some seasonality about it, which is those 2 were particularly sharp this year. And I want to say that there was a pickup. We won't disclose numbers, but there was something of a pickup over the quarter. I also want to say, though, that we talked about this change in the access model, that kicked in pretty much at the beginning of July, but had some weeks before it's, it kind of geared up and its effects became really apparent. And it's, and we haven't done the detailed analysis to try and separate the 2 factors. There was a second,. We think that, that change also built real momentum in the model as we went through the quarter. So there's 2 factors to state, I'd say.

And then it's worth saying on the second one that getting people to register and log in and return is a long-term engagement strategy. So we are building a population of people who will convert over a long period of time, which is, we're happy to see that happening. So it's not just that at the point of friction where they're asked to register and log in that they convert; it is also that they begin to engage. We are more effective in engaging them when they register and log in and return. And then they convert sort of over a period of time. So there's an assumption that it gets better over time.

It's worth adding, though, another benefit is that we can track them between devices. Because they're logging on to different devices, we can see cross-device use. And we're quite heavily tracking how these cohorts who arrived on this are tracking. And I want to say we're still seeing positive effects from the cohorts who registered and logged on back at the beginning of July. We're still seeing now months later better coefficients, as it were, in terms of their virtuous behaviors and propensity to subscribe and so forth for those cohorts many, many weeks later. So all of this stuff is encouraging for our model.

My next question about the paywall. Is it still 5 free articles out there in the U.S. but also international? And how do you sort of think of that going forward?

It's actually dynamic now. So we're essentially experimenting and training models to get at what is the right cocktail of engagement for, to get somebody to ultimately pay and stay. And so depending on what your journey is, how you come to us, what your prior behavior with us was, you'll see a different number of articles. But it's basically now varying between 5 and 10. But depending on what journey you take, you may see a different number.

Then maybe also, if you could kindly just talk about sort of your outlook for marketing spend. You've talked in the past about it's getting more and more efficient. Might, are you sort of reaching a ceiling here in terms of how much you could spend where it's not helping you as much if you add another, say, $5 million or $10 million to it?

I think Roland basically answered this question, and I'll just say we're very pleased in this quarter to see improvement in the organic engine and its ability to drive starts. I think, and I think over the long haul, that is a mechanism for finding operating leverage, but we don't rule out that there will be periods where we'll want to spend more. And I'll just tell you, we've now seen, this was one of them. But when we see opportunities to spend kind of into demand that has a positive effect on the model. So I would say, I wouldn't call it a trend yet, but I think over the long haul, what you've been hearing us say repeatedly is that the plan is to get the product itself, which is the UX and the journalism, to do more of the work. It's also worth saying that as you look at the mix of spend in marketing, one of the things that's been really pleasing is we're making our way, we're still making our way up the funnel, and Roland said earlier, sort of investment in digital product work in journalism is long arc work, and it pays off in big ways over a long period of time. I would say the same of brand work. So the more that we can shift the spend that we have to middle- and upper-funnel work, middle-funnel work, getting people to engage, brand work, getting people to sort of think or feel a particular way about The Times, that is an investment in the long-term health of the business.

And we're tracking lifetime value subscriber acquisition cost ratios very, very closely.

Yes.

We're getting better at projecting lifetime value. One of the reasons we're so pleased about the early news on retention of the $1 a week customers is that, that solidifies our view about lifetime value. And as Meredith says, what we think we can economically sensibly spend the variable direct marketing cost to get the subscribers, we will do. We're encouraged that we've seen some weeks now where we've been able to get great numbers without having to put as much money in.

Guys, is that 5 to 10 free articles across all devices, because you're obviously moving towards asking people to register before they access the content. So is that sort of 5 to 10 free articles across all devices out there as opposed to per-browser, per-device like it's been historically?

So just to clarify, when I said 5 to 10, we're talking about how many stories you get after you register. So if you're anonymous, you basically after one story now, you have to register and log in to get more. And then what we're doing is training a model to understand based on your particular journey, what's the right number of articles after you register and log in that gets you to engagement and ultimately to pay and stay.

And my final question, if I could. What should the size of the newsroom be at, say, the end of this year or right now versus where it started the year at? Obviously, you guys increased that outside the U.S., you're talking about Australia and Canada, but maybe you could just flush that out a little bit.

Yes. So let me go first. Meredith may want to add to this as well. I mean, look, the question of how big a newsroom you need, there are many, many factors that play into that. This is ultimately a Board, a publisher and Board decision. And one of the main things we look at is what does it take to cover a really complex world better than anyone else and deliver the best journalism in the world. And so this is not purely, as it were, a near-term economic question. What I want to say is we do not believe that we're going to have to scale newsroom costs for, as it were, endlessly to continue to feed subscriber growth. We think that, for example, the international target we talked about -- I said earlier, we hope to quadruple the number of international subscribers by the mid-2020s from 0.5 million to 2 million or more. We don't believe that we're going to have to, as it were, populate the entire world with -- we already have over 30 New York Times bureaus around the world. We do not believe we're going to have to massively scale that kind of on-the-ground journalism to achieve those numbers. We're going to do it through superior kind of personalization and focus on subscribers and their propensity to subscribe and by improving pricing and marketing tactics in chief. So it's -- this is not -- it's rather unlike, for example, the television screeners who do tend to find they need significant local scripted content to drive subscriptions in a given market. So this is another area where we think, as it were, we can achieve, if you think of the newsroom costs relative to revenue that we can achieve over time, improving operating leverage. I want to say, though, The Times wants to remain the best news provider in the world and where we need to put more boots on the ground anywhere on any area to achieve that, we -- our model basically is you invest in great journalism, and from that great journalism, the rest of the -- entire kind of business strategy of the company and the appeal to subscribers flows from the quality of the journalism.

I think that's right. And I would just say, I think we've gotten to where we are because of the continued investment in journalism over the last very long, long period of time. So when we talk about the product itself becoming the primary engine and an improving engine of getting people to make a daily habit with us and ultimately become subscribers and stay as subscribers, we assume continued investment in our journalism and continued investment in the digital product experience, and that's where we're focusing that investment. I think if you look at across the portfolio, we've made a sizable investment in our journalism in audio and that's really paid off. And I think you're going to continue to see us invest across a breadth of topics, as we've done with new products.

I'm sorry, I appreciate that. Where does that put the total newsroom size, though? Thank you.

We can't give a direct guidance on that. I mean, we're currently around the 1,700 mark in the newsroom. I can't really add to our previous remarks.

Okay. Thank you.

Thank you. And the next question is from John Belton with Evercore.

Thanks. I have 2 questions. First, on Facebook News. Can you just talk a little bit more about the decision to participate in that product? And maybe compare and contrast that decision with the one ultimately not to participate in Apple News Plus? And on Facebook News as well, you mentioned a multiyear license fee here. Will the agreement have any impact on operating expenses going forward? My second question is for Roland on the balance sheet. Any update on plans on what to do with the maturation of the sale leaseback agreement? And how do you think about the optimal capital structure going forward? Thank you.

I'll take the first one, John. We decided to participate because we saw this, as Mark alluded to in his remarks. It was a substantial increase in what we've seen from a platform for the right to use New York Times content. So I think the amount of money Facebook was willing to commit to us represented a step-change in what we've seen from a platform before, and so that made sense. And then I would say, secondly, and it's really important to say this, it was consistent with our business model of driving subscriptions via direct relationships. So what Facebook is ultimately getting from us is largely links and short summaries that send people back to our owned and operated platforms. So we think that's ultimately good for our top of the funnel and helps drive the business model that we have. As to cost, which I think is what your second question is about, the answer is no.

So, John, as you know, when we, in essence, buy back the -- our condo interest in the building, we'll be debt-free at that point. In essence, the revolving credit facility, one way to look at it is replacing the liquidity that we lose with the completion of that transaction. However, none of that really changes our outlook, our current outlook, which is that having a conservative balance sheet has served us very well in an industry that's in transition, in a company that's in transition. And really focusing on having maximum flexibility to invest when we want and how we want, independent of the vagaries of the industry or the vagaries of the marketplace, we believe is important to the long-term share value for all shareholders. Because we're still in a period where we believe there's plenty of opportunity as we continue to invest in the drivers of growth and the drivers of our growth strategy. So being a little repetitive here, but we still think there's plenty of room to invest organically. We continue to look for potential small to midsize acquisitions, all which would be in service of our growth strategy. So for now, really not a change.

And the next question comes from Doug Arthur with Huber Research Partners.

Couple of questions. Meredith, can you just update us on your progress in the podcasting product? And also, in terms of The Weekly, do you feel that you're getting some leverage off The Weekly in terms of new subscribers to digital?

That's a good question. Yes. I'll start on The Weekly and just say, we have always envisioned it to be a very good marketing tool for The Times in addition to everything else. So one big story at a time certainly is bringing Times journalism and very deep reporting into homes that we probably weren't in before. It was big part of our choice to go with FX and Hulu in combination. So I would say broadly, yes, it's bringing new audiences to The New York Times. On audio, the -- I may make you go back and ask the question again -- but I think you're asking so where do we see running room, and I would say kind of everywhere. And Mark got on this in his remarks, The Daily audience continues to grow. We don't see that moderating in any way. So as the audience grows, the ad business grows, and that's happening at a really nice clip. The Daily is also serving as a great distribution mechanism for new programs. So The Daily was basically the parent of 1619, which I think we're 5 episodes into, might be 6 at this point. 1619 has been a huge hit, launched into the feed of The Daily. I think it immediately went, because it launched into the feed of The Daily and then because it was so good, immediately went to the top of the charts as far as listened-to podcasts. So it's, our sort of audio strategy of having The Daily as an envelope to send other things out into the world, is really working. And then I would say, and I alluded to this before, we're continuing to invest pretty aggressively into our audio newsroom, and we have an expectation that we will put more product into the world, as we did with 1619, with that audio newsroom. And I think in the near term, it's very good for our ad business. There's a lot of demand for high-quality audio. And I think in the short, medium and long term, I think that will stimulate our subs business in direct and indirect ways.

Okay. Great. And then, Roland, I'm wondering, in terms of your cost guidance, obviously, it's very preliminary for 2020. I mean, you sort of alluded to the flexibility to step up marketing costs if there's an opportunity. Obviously, you're going to continue to invest in content. I mean is it, and the reference to 2020 cost is made off of the fourth quarter guide. So can we sort of think at this point of sort of mid-single-digit cost growth across the board? Or is that, is it too early to say that?

Well, I mean, we, our tradition is not to guide past the quarter, but not to directly give guidance, but back on the marketing question, most of this was about discussing the lower cost guidance for Q4 than we had in previous quarters, and we don't think that's going to continue. Specifically, on the marketing, on the direct side, we believe the efficiency of our machine here is getting better, and that will require less marketing dollars to bring in the same amount of subs. However, Meredith touched on this, the brand aspect of marketing, I mean, and we will keep the option open, that when there are opportunities that we believe that an injection of brand spend will be beneficial to the top of the funnel and, therefore, the long-term growth of the company, we're going to do that. But all said, back to the 3 areas of investment, what we see is a bit of a pivot from the investment in marketing to a much more of a heavy-up on our investment in digital product development. So you can triangulate amongst those numbers and see. We'll continue to spend and invest in the growth in 2020.

Okay. Great. And, Meredith, just a follow-up on the audio. Can, are there any sort of subscriber numbers or unique numbers you can give us in terms of any new benchmarks you've hit there?

I'll say 3 things about it. One, we may have said on a prior call, I don't remember, but we crossed 1 billion downloads for The Daily. So I think...

We talked publicly about, I think, regularly now more than 2 million listens a day. And I think now, something like 12.5 million listens over a month. And you'll know, Craig that the, one of the things that's interesting and, I think, exciting about The Daily is the demographic: 3/4 of this audience is 40 years old or younger, 45% is 30 years old or younger. So in terms of, as with the broader health of our brand and our funnel, bringing in a large number of millennials who are typically completing the show, which means 20 minutes at the moment, often on iTunes or Spotify, but on a smartphone. So engagement of 20 minutes or more by millennials on a smartphone to a serious journalistic product is, I mean, I don't think anyone in the world is doing that, to be honest. And it's a really exciting thing for us.

The only other thing I'll add, I shouldn't pick favorites from our marketing efforts, but we really like the audio ads that we're running for our own products, essentially to drive subs inside The Daily. So they're generally our journalists talking about how they do their work and those are having a very positive effect. And you can expect to see us do more of them.

And the next question is from Kannan Venkateshwar with Barclays.

First, I just wanted to get a clarification on the international numbers that you guys provided, the 500,000 subs. Meredith, if I'm not wrong, last quarter, you had mentioned 16% of your total subs were international, which would imply roughly about 570,000. And I know promotions rolled off internationally as well. So I just want to make sure we're looking at apples-to-apples, when I look at the 16% of total sub number last quarter versus the 500,000 that you guys quoted this quarter. And then I have a couple of others, which I will follow up.

Yes, it's, I think it's set at 16% of new subs. Don't forget we've got Cooking, Crosswords in our numbers as well, which I think answers the first question.

Okay. So I guess from a trend perspective, is it fair to say that you guys are growing despite the promotions rolling off internationally? That's the clarification I just wanted to get.

Yes. Yes.

Okay. Okay. And then second...

Have you got that first point, which is, we're talking about 0.5 million over around 3 million, which is, pardon me, my mental math says that's about 16% as opposed to the higher number, which includes Crosswords and Cooking?

Sure, sure. Yes. We got it. And secondly, from a guidance perspective for the next quarter, what I wanted to understand was obviously there's this promotional roll-off coming, and you guys are expecting it. And obviously, you have a plan to deal with it. So in terms of assumptions for revenue growth next quarter as well as how you guys are thinking about sub growth, if you could just help us think through the variables and whether you expect numbers in Q4 to replicate the strength that we saw in Q3? And related to that, I guess, is when you look at this $1 a week promotion, obviously, it's been a workhorse for you guys over the last year or so, how is the yield on that trended over time as that promotion has matured? Are you attracting the same number of subs that you did initially? Or has that eased off a little bit as it matured?

I think I'll take a stab at the first one and Roland may add some and take the second one. Just on the drivers of the business, which is what I think you’re asking about, how should you think about future quarters and particularly the next quarter. I think it goes back to something both Mark and I said, which is we're beginning to build a flywheel here, which is essentially, can we continue to have a large anonymous audience? And I think there's news cycle in that. I think there's also our own ability to stimulate things like search and social promotion, can we get that audience through the application of friction or, in some cases, the ease of friction, can we get them to become a registered user of The New York Times? Once we've done that, can we get them to log in, return and log in, and then can we convert them? So I think that answers the question of sort of how should you think about the numbers going forward. Much of the model is tied up in that.

Thanks, Meredith. Just 1 or 2 observations for me. I mean, firstly, if part of your question, Kannan, is, are we -- as we're seeing offer fatigue, in other words, an initial -- an offer was initially attractive and is beginning to wane in its attractiveness, the answer is absolutely not. There's no evidence at all. On the contrary, I think there's some evidence that people are kind of discovering about it, and we're getting new subscribers because of the offer. The other thing -- the next thing says there is some complexity there because, obviously, what we've now done is we've changed part of the geometry of the pay model and that's interacting with the numbers as well as the simple offer. So we've continued to have -- as we have done now for some years, had part of any month when we are on a higher price and part of the month where we're on a lower price, and typically, the lower price is that $1 a week offer. That's continued. But we've also now got a much, much larger and rapidly growing cohort of registered, logged-in users who were able to interact with it more effectively and get higher conversion from. So there's a lot going in the model, but the -- and as per yield, what I want to say is the -- we've said that the $1 a week, so as we're through that first year, retained very much in line with previous cohorts, there's no difference in their retention. Clearly, a really important question has been how well would they retain when you get to a point where they're being asked to pay a higher price. And I don't want to add to the remarks we made in Roland and my scripts at the beginning of this call, but to say the initial signs are very encouraging. I mean, we will know more as we go through some more months and see more of these guys going through. But I would say in terms of what the ultimate yield from these subscribers is going to be, which is a mixture of, obviously, price and retention, I would say the signs are very encouraging.

Yes, the only thing I would want to add is where this all comes together between the number of folks we have signing up, their retention rates, how they ultimately step up and all the other moving parts, the percentage of international, if you go back to our guidance for the fourth quarter, we're saying we believe we can grow that digital subscription revenue in the mid-teens once again, off a bigger base. So when that all comes together, we're still growing the revenue, and we're growing it without that percentage increase reducing.

Got it. And one last question, which is on the tenured subs, I mean, how are you guys thinking about what tenured sub actually is? What percentage of your base is that and how you're thinking about that process of stepping up? Is it gradual? Is it -- is there some kind of a product change in order to implement that? If you could just help us think through.

Yes. We won't to say much about that, but I would just say time and engagements. So if they've been with us a long time, the best predictor of someone's retention is how long they've been a subscriber, have they been with us a long time and how engaged are they.

Okay. But any sense you can give us on what -- how big that base is?

I don't think we want to do that, Kannan. I mean, I want to say our assumptions are pretty commonsensical. We're talking about people who have been subscribers for a long time. And we've got, arguably, nearly 170 years of experience with a certain class of subscriber. We know that beyond a certain point subscribers retain very, very well.

And this concludes our question-and-answer session. I would like to turn the conference back over to Harlan Toplitzky for any closing comments.

Thank you for joining us this morning. We look forward to talking to you again next quarter.

Thank you. The conference has now concluded. Thank you for attending today's presentation. You may now disconnect your lines.