Flutter Entertainment plc Q2 FY2022 Earnings Call

Good morning, and welcome to the Flutter Entertainment half year results update call. Peter Jackson, CEO, who is joined by Jonathan Hill, CFO. (Operator provides instructions.) And now I will hand over to Peter.

Thank you very much. Good morning, and thank you for joining Jonathan and I for this call. Hope you all had a chance to watch our presentation this morning, which helps set out extra color on our half year performance and how well positioned the group is for growth. Before we go to questions, I'd just like to highlight a few key points. I've been really pleased with our performance in the first half of the year. We had positive momentum with revenue up 9% and AMP growth of 14%. Our U.S. performance was outstanding, and continues to exceed all expectations delivering a profit in Q2. This was driven by phenomenal execution with the sports betting market share of 51% in Q2 where we're number one in 13 out of 15 states. And it was also achieved despite the fact we've been leaning into customer acquisition in Q2, acquiring over 0.5 million new customers. As we laid out in our presentation, we have a clear line of sight to U.S. profitability in 2023, for the full year, including the cost of share-based payments. You'll have seen that we now expect the outcome of the arbitration process with FOX in October. Unfortunately, as this is a live process, we can't provide any further detail. However, as previously stated, we remain very confident in our position regarding fair market value. In the group, outside of the U.S., revenue and EBITDA in H1 were in line with expectations with strong performances in Australia, and in our consolidate and invest in international markets. And we're excited to welcome the Sisal team into the Flutter Group last week. In the U.K. and Ireland, our online performance in Q2 improved sequentially on Q1, and we are confident that the proactive measures we are taking from a safer gambling perspective has set the business up well for the future. Across the whole group, the second half of the year started in line with our expectations. Our advantages in scale and diversification position us extremely well to capitalize on the growth opportunities ahead. And with that, I'll now hand you over to Phoebe to handle questions.

(Operator provides instructions.) And our first question is coming from Mitchell.

Yes. Firstly, on the U.S. and just to pick up actually on one of your introduction remarks there, Peter, in terms of your Q2 performance and you generating positive EBITDA in Q2 is one thing, to do just given the step up in share and customer acquisition is something altogether different. I appreciate you've covered a lot of this in terms of the presentation. But can you comment on how your customer economics and your approach to customer acquisition more generally evolved in the period. Clearly, it has been well documented to a bit more benign competitive environment in the second quarter. If I'm reading the chart correctly on Slide 19, you seem to secure the step up in market share with no major change in the ratio of GGR to NGR. So it's question number one, around your approach to customer economics and customer acquisition in the U.S. And then secondly, in U.K. and Ireland online, I think you referenced as well in terms of the sequential improvement Q2 over Q1. If I've read it correctly, I think it improved from minus 26% external bonus and minus 11% sequentially. I wonder if you could just provide some context for that improved performance and how it positions you heading into the second half of the year.

Thanks, Michael. Look, I think in the U.S., the major point to make about our performance is that we've maintained our disciplined approach to customer acquisition. We've always been very much focused on looking at the ratio of our acquisition cost to lifetime value. And we continue to see advantages around the lifetime value of our customers because of our Parlay penetration and other benefits we get from our very strong product. And so we've remained disciplined with our acquisition costs. And look, whilst other people were pulling back from the market in Q2, we carried on leaning in because we could see that the lifetime values were being maintained. And so we just took as much business as we could. And as I just mentioned, we've acquired more than 0.5 million customers. And so I think we're very pleased with how we've exited the first half. We've got a much bigger business now in terms of customer numbers than we had anticipated. We actually ended up spending a bit more money on that. But we know that the value the customers are going to bring in is as good as it has been in the past. So the business is very well positioned in that regard. And that's what's driving the very strong revenues that you see. Jonathan, if you have anything to add?

I think that's the critical point here is we're building the embedded base and as it pertains to the slide showing the existing customers versus new customers. What we want to do is enter 2023 with as big an embedded base as possible. And we obviously explained the sort of the 1-year economics to you on our customers. We obviously don't make money off the customers in aggregate with acquisition cost in year 1. So you can understand why, therefore, we can move the revenues up, but aren't moving the EBITDA guidance because actually some of that high performance has been driven by that incremental investment in year 1. But obviously, that gives us an even bigger base going into 2023, which gives us greater confidence as we cast forward into — looking forward to hopefully getting to that EBITDA positive position in 2023.

And Michael, regarding your second question around the UK & Ireland performance. Yes, I'm sure Jonathan will have something to add to this as well. But you're right in terms of the improvements we've seen from a sequential perspective. And I think that it's important to recognize there's a few things going on. First of all, when you look at the comparative period last year in lockdowns, we know that our customers were more engaged with our products than they were at pre-COVID, and that has reverted. And so that's one factor. Secondly, we did move early and make a number of changes from a safer gambling perspective and we're starting to annualize those. And thirdly, I think I point to a number of product improvements we've started posting across the business. Jeremy, I don't know if you want to sort of talk specifically about those?

I think the other point is we're looking at the Gambling Commission data as well. And we think we really saw a step up in our performance in Q2 relative to the prior period — we've got some good momentum in the business. The product improvements coming through, which you'll have seen in the presentation, are clearly driving the performance, both in getting Sky Bet back on track in terms of building that product and the high engagement levels we've seen at the end of the last EPL season and then obviously some product improvements also on the gaming side on Paddy. So really, really positive on that and seeing gaming going back into growth in June year-on-year gives us real confidence as we go into H2, and also it also informs the comment we've made around not seeing at this point any impact in the UK & Ireland certainly from the sort of the cost-of-living question that clearly is exercising people in many industries, but we haven't seen that as yet. And gaming is obviously a pretty key indicator of activity levels. And when we look at things like our Tombola performance, which is obviously a very pure gaming business. And certainly, the gaming across our other brands, we're not seeing anything discernible at this stage, but we are keeping a pretty eagle eye on what's going on in the weekly data.

(Operator provides instructions.) Our next question is coming from Clara Canton.

I wanted to also ask a question on the U.S., and maybe sort of on competitive dynamics, realizing that some of your peers have sort of taken a step back. I'm wondering if this is a dynamic that you think is really going to be related solely to sort of non-launch periods? Or do you believe that when sort of the next queue of states or the next cohort of states comes online, is there a chance that the new launch costs could end up being a little bit less extreme also? And then the second question is on product. I wondered if you could talk about product evolution sort of not only to date and your success with Same Game Parlay adoption, but also product innovation maybe heading into the NFL and NCAA season on the back of the year.

Clara, good morning or good evening depending where you are. Regarding the U.S. and the competitive dynamics, it's very hard to talk about what our competitors are doing because we don't have visibility into what's motivating them. I think from our perspective, we're very focused on acquiring as much business as we can while monitoring the returns and paybacks that we're seeing at the moment. That's why in Q2, we continued to lean in hard because we saw abundant opportunities. It's worth remembering, and if you look at the statements in the presentation, a lot of our acquisition was coming from some of our earlier states. So this is not just something that's focused around recent state launches. There is a degree of seasonality, obviously, in the U.S. around NFL being a big push from a customer acquisition perspective. And clearly, a state launch also has an impact as well. But I think that it does appear to me that we're seeing our competitors acting with more discipline, which I think is a good thing for the industry. We have always been very disciplined. And we're starting to see, I think, people coming closer to the position that we're taking. Competitors may not have the advantages that we have in terms of being able to monetize customers as effectively and are having to pay higher acquisition costs because they don't have our DFS base to cross-sell into. From a product evolution perspective, we have the number one sportsbook app in the App Store. We have the benefit of our global organization helping develop and provide ideas to support the U.S. team. Slide 7 of the presentation shows a lot of the group benefits that FanDuel has access to. We have around more than 80% of our handles priced in-house at the moment. It's going to go to 90% in due course. And there's a lot of innovation around the product. The popular Parlays, the Same Game Parlays, more than 80% of our customers played the Parlay product in Q2. We've only just launched search. So there are lots of one percentage point differences that we're making across the product. When I look in places like Australia, we launched the Multi Revolution, which is what they call the Parlay product, back in 2016. And here we are six years later still evolving and developing the products. I'm very confident that the situation will remain the same in America with lots of innovation to come and us able to benefit from the expertise and innovation that the FanDuel team sees from their colleagues across the world. Going back to your first question, the start of the NFL season is such a customer acquisition point in the year for U.S. sports betting that we still expect there to be a very competitive start of season. And while people may have pulled back to an extent in Q2, we certainly expect and plan for it to be a highly competitive start to the NFL season. Again, I think it will be every single year. We can't be complacent by a very strong Q2 performance, and we've got to be really focused on winning again at the start of season and NFL because I'm sure everybody will come out with guns blazing again. So we need to stay pretty focused on that at this point in the team. We've got a great plan for the start of NFL. And it's against the context where there are elections in November as well. So we know that there's going to be a lot of noise from a media and advertising perspective. So Jonathan is absolutely right. It is the most competitive period, and there's going to be a lot of people out there with messages and attractive offers.

Our next question is coming from Ed Young.

Thanks for the level of transparency and disclosure in the presentation. It's really appreciated. My first question is on the U.S. on Slide 23, you've given very helpfully a breakdown on an estimated P&L in New Jersey. I appreciate you're trying to give an indication that mature states are significantly profitable. But I just wondered — in terms of the three buckets there, you've already flagged marketing will come down further over time. Cost of sales there is shown with a 59% gross profit margin; it's also still quite a competitive state and its promotions might be a bit higher and then other operating costs, but then presuming you're going to get some more operating leverage. So I guess what I'm asking is, is that a kind of implicit steer or guide to how we should think about FanDuel's overall profitability when those states are four years in other branded average? Or is that too much to read into it? And it's just really saying this is healthily profitable at that period? My second question is on the U.K. one of your peers had a comment on recreational customers yesterday and the chart around that. You put it on Slide 28, which again is useful showing that you're actually covered. I just wonder if you could help quantify what you mean by tier 1, tier 2 and tier 3, so we can think about to what extent it's possible a like-for-like comparison.

Look, let me just deal quickly with the U.K., and then we come on to the U.S. We don't provide the detailed breakdowns of tier 1, tier 2, tier 3 customers. But I can tell you that the base that we have in the U.K. is now very representative of what the U.K. looks like, whether you take slices of income, wealth, or other demographics. So we think that we've set the business up very well to be sustainable moving forward. Jonathan, I don't know whether you want to talk to Ed's point around the EBITDA breakdown for New Jersey?

What we're trying to do here, Ed, is not try and let you read too much into the forward view on this. What we're trying to do is say, should people have confidence that we're on a profitable EBITDA trajectory on the states from when they've started through to year four, and do we judge it on the basis of how we've allocated the OpEx. We've done it in a way that we think makes sense to give you guys a steer on the fact that we are EBITDA profitable in that state. Clearly, this will evolve further over time. And you're right, we do expect the P&L to alter over time and particularly on the marketing line. But at the end of the day, this state is still growing. And actually, when you look at some of the penetration curve data that we put in on Slide 20, I think you can quite quickly work out which state is New Jersey from that penetration. And you can see it's still going up steeply. So you've always got this issue of how much top line growth do you end up driving by investing in marketing and promo spend. But the P&L in year four is still driving significant top line growth, and you can see that from the penetration curve on the middle graph on Slide 20. So yes, don't try and read too much into it. I think we're just giving — trying to help people see the track we're on at this point with further opportunities for the P&L to evolve over the next few years. But obviously, we're doing a Capital Markets Day in November, and we can maybe touch a bit more on that at that stage. We're not going to extrapolate any more at this point.

Our next question is coming from James Rowland Clark.

A couple of questions, please. I just have sort of a slightly different way of asking Richard's last question is on Slide 23, on the left-hand side, you've got the chart indicating the operating leverage that you have in the U.S. with revenue growing at two times the operating cost base since H1 '19. Would you mind just giving us a little bit of help on how we should see that evolve over the next six months, and into next year? And then secondly, again, on the U.S., you've delivered an EBITDA profit of EUR 22 million in Q2. Could you just help us with how that looks for FanDuel versus Fox Bet, you have in the past split the two businesses?

James, I mean, I'll let Jonathan give us the details on that. I think the profit that we made in Q2 not only hit all of our U.S. business but also includes the costs, the investments we've been making to launch FanDuel into Canada. So the gross number for FanDuel is actually greater than the one you see on the front of the presentation.

You'll see in the interim release, we referred to the fact that Fox Bet was, I think, 3% of revenues and about 20% of the EBITDA loss in the half. So you can work out that it's $30-odd million of which you can take half in Q2 because it's a pretty flat business. So therefore, you can probably gross up the €22 million and get to a number closer to €40 million excluding Fox Bet. And then obviously, we've got, as Peter said, the investment of FanDuel in Canada, and we've leaned in. We did benefit from some positive sports results, but even without those, we feel that FanDuel is in a very strong position in Q2. On the point on Slide 23, I don't like guiding on very short term because if you had looked at this graph a year ago, you would have seen a different story, which would have been even more revenue growth relative to OpEx because we had a slight hiatus in 2020 as we entered COVID in terms of our OpEx growth. We pulled back quite strongly then we stepped up in 2021 then we get the annualization of that into 2022. So I think looking at half year is quite difficult. What I would say is in H1 this year over H1 a year ago, if you normalize, we had 71% revenue growth and about 46% growth in OpEx. So it's not quite the 2:1 ratio, but it's not far off. Over the medium term, it's a reasonable directional framework.

Our next question is coming from Kiranjot Grewal.

Just a couple of questions from me. I think in U.S., you've been talking a lot about being opportunistic during Q2, acquiring customers, while others are spending less. You're the biggest in the U.S. globally, and you have a strong balance sheet. Could you not be opportunistic to maybe spend more even going further on acquiring customers? And then secondly, around the customers in the U.S., how sticky is the customer in the U.S. versus those in more mature markets, say, the U.K.? Just trying to get a sense of the customers won in Q2. How much of that you'll retain as maybe marketing spend goes up by competition?

In terms of the two broad questions, we did spend a lot more than we had originally planned in the first half, particularly around Q2. Together with the team, we saw a lot of opportunities in the U.S. and so we pushed hard. I'm not sure that we could have spent a lot more in the subsequent way that we always try to do. I'm sure that there would have been some further opportunities, but it may not have delivered the returns that we wanted, but we have been pushing very hard. You're right, we have been putting all the investment that we can into the business. So we don't feel constrained about that. And in fact, there were a number of states that were planned to launch that we haven't seen. And irrespective of that, we went ahead and increased the investment.

To add one point, we maintain consistent discipline. We did exactly the same thing in the diametrically opposite situation in the run-up to NFL season last year when everybody was going pretty crazy with helicopter money and large free bets. We maintained real clarity and discipline in what we were doing. When we had the opportunity to lean in, we also kept the same level of discipline and clarity on what we were doing. It served us well in both situations, and we think it's the right way to build this business — a really clear, disciplined and focused way. We think we took some great opportunities in Q2 to bring in customers, particularly around our NBA product which is fantastic, and it gives us a great opportunity to bring people in, and we think we did that really effectively.

In terms of the stickiness of our customers, refer to Slide 21 where we show the cohort performance of some of our customers. It's quite hard to identify on a clean basis because of COVID, but last time we talked about the year 2 on year 1 performance where we're seeing an 11% growth in revenue per customer. On Slide 21, you can see that we've seen that accelerate somewhat — looking at customers in year 3, comparing revenues to year 1 for the same cohort, we've now seen an acceleration of that revenue growth. So there's now a 22% CAGR performance. If you think about other businesses over the years, whether that's Sky Bet or Sportsbet, these cohort dynamics show that even if you see some attrition in customer numbers, the revenue from those customers can grow year-on-year. We are continuing to see that trend in the U.S. market.

Overall, when you look back at the cohort charts for Sky Bet and Sportsbet, these cohort changes for the U.S. create a similar pattern.

Our next question is coming from Monique Pollard.

Just a couple of questions from me, if I can. The first is on the UK & Ireland online performance. Obviously, as you said, the percent revenue growth was down 19% in the first half, and I think it was down 11% in the second quarter, much better than the U.K. Gambling Commission data, which was down 23.5% and 22% in Q2. So it seems clear that you're taking share in that market. I'm just wondering whether you think there's less potential to keep the share gains into the second half given some of your new products like the build-your-own product was only launched relatively recently, so it should have more of an impact in H2 versus H1. Then second, on the U.S., given that New Jersey EBITDA margin, helpful in terms of getting insights to profitability in some of the immature states. Just wanted to understand whether EBITDA margins generally, you see them being higher, lower or relatively consistent in sports betting only states versus sports betting and gaming combined.

Thank you, Monique. In terms of the UK & Ireland online performance, you are right to highlight the favorable performance of our business in Q2 compared to the market and where we're taking share. A few factors: firstly, we're seeing the benefits of our more recreational-focused business. Our portfolio brands — Tombola, Sky Bet, Paddy Power — position us very well compared to the market. Secondly, we've been at the forefront of leading the market on safer gambling, and we've taken a number of proactive measures some time ago. It appears other operators are beginning to catch up. The Gambling Commission has been raising the bar, encouraging operators to be more responsible, and you are beginning to see that impact. Third, we are continuing to invest in developing our products. There are areas where we thought we'd set a bit behind, and I think we've continued efforts on delivering good products. When I look at June performance showing gaming up year-on-year, I think that is great validation of the work our teams are doing.

Monique, on your other question, if you think about it, it doesn't matter whether it's a single product or multi-product. If we're investing on a CPA-to-LTV basis that makes sense, we should deliver consistent profitability. There may be slight divergences depending on stage and investment strategy, and there will be a small impact probably from multiple products over single product, but I don't see it being a massive divergence. When we look across other states, we can see them on a similar basis moving into the EBITDA profitability that we're seeing in New Jersey. It's just they're further behind at this point in the evolution.

Our next question is coming from Joe Stauff.

Two questions, if I could, on the U.S. I wanted to ask around the dynamic, especially user growth, New York versus New Jersey, and Pennsylvania obviously, a big new state that's exploding you're doing well in. Were you able to grow users say in the second quarter in New Jersey and Pennsylvania? And the second question is as we think about the U.S. group and waiting for the arbitration and so forth to be settled, I'm wondering if you've framed out maybe the margin benefit or the additional costs necessary to add to the U.S. group if they weren't sitting within the total group in terms of the Flutter group.

Joe, in terms of customer acquisition dynamics, the best thing to look at is Slide 20. If you look at that, you can see the penetration rates for all of our states — they are all following a very similar upward trajectory. None of the states have reached a point where the growth rates are plateauing. Based on the dates of launch, you can infer which is New Jersey because it's the longest line, and you can see it is continuing to follow a very similar trajectory as it did in the past. So we're seeing very good growth rates and customer acquisition volumes from the earlier states, and they continue to grow. I think that's important. We are continuing to acquire business. In the first half, one-third of our acquisitions in the half were in pre-2021 states, so that tells you there is still huge growth there. We acquired 1.2 million in the first quarter and 0.5 million in the second quarter. Over one-third of those came from pre-2021 states. You can see FanDuel penetration in each of these states and those lines still going up. That's really exciting for us seeing growth across new and historical states.

On the second part of your question around the benefits of FanDuel being part of Flutter — they are significant. If you look at Slide 7, we've highlighted benefits like risk and trading capabilities, which are very significant. Superior economics around monetizing handle come from the group's risk and trading capabilities where so much is done in-house. The fact that our Parlay products are developed in-house and priced in-house means we keep that incremental margin and have a superior customer proposition. The group's technology and the global betting platform help FanDuel. We also have the best casino product in the world with our PokerStars business. We will increasingly bring more of that product and capability into the U.S. market. There are a lot of benefits, including the ability to transfer best practice and expertise across the group quickly.

Our next question is coming from Richard Stuber.

Two quick questions. First, your guidance for GBP 225 million to GBP 275 million loss in the U.S. for the full year implies about GBP 100 million to GBP 150 million of losses in the second half, which is similar to what you did in the second half of '21. To me, that seems still pretty conservative, especially given the GBP 16 million of profit you made in the second quarter. Is this purely a function of ramping up customer acquisition ahead of the NFL? Or are there other costs in the second half we should be aware of? Second, in your presentation, you discussed things like personalized generosity investment in Sportsbet. Is this something you've rolled out across other regions, in particular the U.S.?

Richard, the answer is yes. We expect to invest into the back half. We have more states live and more states live for the first time with NFL, which is an important acquisition point. We'll be investing positively and aggressively with our usual level of discipline through Q3 and the run-up to NFL and through Q4. We also have Kansas coming live, which we hope and believe will be in Q4, so that will add another state. So it's really just that investment and scaling up the business and making sure we're taking the opportunities before us.

On personalized generosity, this is a journey. We've made significant strides in applying generosity effectively in Australia for a number of years and have been improving it in the U.K. as well. When we launched the global betting platform into the U.S., we were able to bring across tools and capabilities that allow us to apply personalized generosity to customers. We have an internal platform — CPP — which is our cross-promotional platform that allows us to personalize that generosity. It's something we're doing and continuing to improve, always with reference to local taxes and cross-selling opportunities across sports.

Our next question is coming from Joe Thomas.

Two questions. First, you mentioned in the presentation that you're thinking about optimizing marketing spend in the U.K. Could you clarify exactly what that means? I'm guessing that means reducing it. Also, how does the news that you're getting rid of best odds guaranteed offers in the U.K. feed into that? Second, the press has mentioned a variety of management changes and some job losses at Flutter. Are these changes that had always been planned? What do you think will work better and why are they being done?

Joe, the U.K. division is made up of the historical Paddy Power Betfair business together with Sky Betting & Gaming. We always said at the time of the merger that we wouldn't rush into integration: we would take our time and be careful because we wanted to maintain momentum. The changes you're seeing now are a reflection of that integration process. There will be some job losses associated with integration, and we're seeing that now. There are opportunities to share best practices around marketing and generosity. We need to reflect some of the changes in market dynamics and some of the changes to lifetime values as a result of safer gambling measures, and all those things are being reflected in what you're seeing in the U.K. division. The work around tech platform, people, and optimizations around marketing and generosity are all part of that. Regarding the best odds guaranteed changes, we continue to review how we apply generosity to make sure it's applied to the right customers and products. It's a standard part of reassessing our approach in the market.

To build on that point, it's not just about reducing spend; it's about optimizing. First, identify areas where spend is suboptimal and therefore reduce. Second, consider how much of that can be reinvested if we can find effective ways to drive growth. So there's a two-step process: identify and reduce inefficiency, then reinvest where there's clear opportunity.

Our next question is coming from Simon Davies.

Two from me, please. Firstly, the Sisal deal takes you into the international lotteries market. Do you see that as a strategic opportunity for the group and something you want to invest in more broadly, given it's a relatively early stage of digital development? Or is that possibly something you might exit? Second question, returning to the U.S., what are your thoughts on the timing of the launch in Massachusetts and how material would that be in terms of start-up losses? And do you have any thoughts on the likelihood of California opening up over the next 12 months?

Simon, we only just closed the Sisal deal last week. The international lotteries operation they have is interesting — there are retail networks and lotteries in markets like Morocco and Turkey. We think it's an exciting business and we're looking forward to understanding more about the opportunities the team have been pursuing. It's exciting on a standalone basis and as a way to get into new markets where sports betting may not be available initially, but could follow. On Massachusetts, we don't have specific live dates yet.

We think Massachusetts is probably more likely in 2023. I wouldn't guide specifically around the loss in Massachusetts. What we've said is we've got a view on 2023 and the states that are going to launch, and the overall mathematics of the size of the retained base versus the new base still means that even with launches we might expect in 2023, subject to California, we expect to be EBITDA profitable in 2023.

Massachusetts would add about another two percentage points to the U.S. population covered by sports betting. California is the big one — it's a very large market and will be a tough fight. We'll have a better sense after the election season; ask me on the 9th of November and I'll tell you how we're feeling about it.

Next question is coming from Daniel Politzer.

Congrats on the U.S. profitability. Just to clarify: it sounded like you spent more in the second quarter on sales and marketing than you previously planned but you've seen the flywheel continue to work. So as you go into the third quarter, I just want to confirm you're now planning to spend more on sales and marketing than you previously had. Is that a fair characterization?

I wouldn't necessarily say that. We remain very agile in making decisions based on what we see before us. What we saw post Super Bowl and into Q2 was an opportunity to invest. We have plans and still expect the competitive market to be broadly similar to what we previously expected. We expect people to step back in having possibly conserved some powder in Q2, so I wouldn't necessarily extrapolate Q2 into Q3. We will judge the competitive dynamics and our paybacks and adjust our spend based on the CPA-to-LTV metrics we always use. So we may increase or decrease quarter to quarter based on those metrics.

As a quick follow-up on iGaming and M&A: how do you think about growing the iGaming business? Do you have a target share versus the high teens you're currently at? If you think about growing it, how do you weigh organic growth and product improvement versus doing something inorganic via M&A?

I think our share is around 20%, which is pretty impressive. Around half our customers have come direct, which shows the strength of the FanDuel brand. We've started scaling up resources for the FanDuel casino business: we brought Asad to lead it; he started in May, so it's starting to have an impact. We've brought other folks over from Paddy Power and PokerStars. The focus now is getting the basics in place. We have a new casino strategy and are excited about bringing expertise from elsewhere in the group to the U.S. business. We'll talk more about that at Investor Day in November.

Our next question is coming from Andrew Tam.

Just staying with the U.S. for a minute. In terms of the two-to-one operating leverage, it sounds like that's a rule of thumb. To what extent does that improve going forward as new state launches happen and start to fall away? Should that not improve over time? And the second part: given you leaned into customer acquisition in Q2, how should investors think about that relative to prioritizing states like New York?

Some costs just move with scale. Customer operations costs will move with scale. As we do new state launches, we need regulatory and compliance resources for each state. Some things are variable and activity dependent. This historic rule of thumb is a reasonable framework to think about the business. If we can make the ratio improve over time, we'll look to do so, but it's a sensible directional way to think about it now.

On leaning into customer acquisition while being selective about New York: New York has particular tax levels and we are thoughtful about the application of generosity such as free bets and other offers in that state. More generally, we continue to see robust lifetime values and saw opportunities to pick up additional business in Q2. We're agile with our marketing spend and chose to invest more in Q2 based on what we saw.

Got it. How should we think about retention versus acquisition spend?

It's a broader topic and we'll cover more at Investor Day in November. But we are investing in retaining customers. We referenced earlier the spend on generosity and personalized generosity and the ability to utilize tools from across the group to improve retention and cross-sell.

Our last question is coming from Ivor Jones.

Two things about Italy. Longer term, is Sisal a nice growth business or is there a step change coming as it becomes part of the group and you integrate the other brands? Narrow question: Sisal is performing well — is that because retailers recovered more quickly than you expected to get back to expected levels? Or is it outperforming your expectations as we get into July and August?

We're very pleased with Sisal's performance. The business only closed last week, so we're early in the integration. Online penetration in Italy is relatively low and we expect it to continue to grow. COVID accelerated online penetration and many customers who went online have stayed. We have seen a bounce back in the first half as retail reopened versus lockdowns last year. There are synergies expected from the transaction: we can utilize our risk and trading capability in Sisal to save cost and improve the product; our Poker business can utilize the retail network for deposits; and we can think about cross-selling poker into the Sisal base. The gaming content is another important element. Because of the time to get approvals done, we've been able to line up the integration, and we're excited to get on with it. Okay. Look, thank you very much, everybody. Appreciate all of your time. Sorry, we've nearly used the entire hour. Got some great questions, so thank you.

Thanks all very much.