Bread Financial Holdings Inc (BFH) 2017 Q4 法說會逐字稿

Good morning, and welcome to the Alliance Data Fourth Quarter and Full Year 2017 Earnings Conference Call. (Operator Instructions) In order to view the company's presentation on their website, please remember to turn off your pop-up blocker on your computer.

It is now my pleasure to introduce your host, Mr. Eddie Sebor of FTI Consulting. Sir, the floor is yours.

Thank you, operator. By now, you should have received a copy of the company's fourth quarter and full year 2017 earnings release. If you haven't, please call FTI Consulting at (212) 850-5721. On the call today, we have Ed Heffernan, President and Chief Executive Officer of Alliance Data; and Charles Horn, Chief Financial Officer of Alliance Data.

Before we begin, I would like to remind you that some of the comments made on today's call and some of the responses to your questions may contain forward-looking statements. These statements are subject to the risks and the uncertainties described in the company's earnings release and other filings with the SEC. Alliance Data has no obligation to update the information presented on the call.

Also on today's call, our speakers will reference certain non-GAAP financial measures, which we believe will provide useful information for investors. Reconciliation of those measures to GAAP will be posted on the Investor Relations website at www.alliancedata.com.

With that, I'd like to turn the call over to Ed Heffernan. Ed?

Great. Thanks, Eddie, and good morning, everyone. As always, joining me is Charles Horn, who is preparing himself for about 20 questions on tax. And we'll also update you on the fourth quarter, full year results. I'll give a wrap up of 2017. And then we're going to get right into '18 and the trends we're seeing in -- by business, and we're going to get after it.

So with that being said, I'll just kick it over to Charles right away.

Thanks, Ed. Revenue for the fourth quarter came in strong at 15% growth to $2.1 billion. For the year, revenue increased 8% to $7.72 billion, slightly soft to guidance due to Hurricanes Harvey and Irma, negatively impacting revenue at Card Services by approximately $80 million.

Core EPS increased 34% to $6.26 for the fourth quarter and 14% to $19.35 for the full year. Excluding the net tax benefit, core EPS increased 12% to $5.24 for the fourth quarter and 8% to $18.33 for the full year, better than our guidance of $18.10.

The recently enacted tax reform lowered our provision for income tax for the fourth quarter of 2017 and full year by reducing our net deferred tax liabilities, essentially reducing cash taxes that we would have had to pay in the future. Tax reform also benefits us moving forward, as we expect our effective tax rate to drop to 26% to 27% in 2018, and that compares to 33% in 2017 prior to the adjustment we just talked about. We also anticipate the cash benefit to approximate $100 million to cash flow in 2018.

Our net corporate debt ended the year at approximately $5.9 billion, while the corporate leverage ratio ended the year at 2.7x, consistent with the last several years.

Let's move on to the next slide and talk about LoyaltyOne. It was a mixed year for LoyaltyOne. AIR MILES met expectations, and we restored the profitability at the level we anticipated, while BrandLoyalty results were unexpectedly poor. We entered the year knowing that we needed to retool the AIR MILES business model following the breakage assumption reset at the end of 2016. We were successful in introducing small changes that allowed AIR MILES EBITDA margins to return to the mid-20s range.

On the other hand, AIR MILES issued came in short of expectations, down 4% for 2017. Softness in promotional activity, especially in the grocery vertical, caused the shortfall. We will need to stimulate promotional activity in 2018 in order to drive issuance growth.

Turning to BrandLoyalty. Revenue decreased 12% in 2017. The weakness is due to several reasons: one, key markets Germany and Russia were off due to the absence of events, such as the Euro Cup and Olympics, which reduced the number of programs executed; two, North American expansion into the U.S. did not manifest, as it's taking longer to establish the foothold here than we anticipated; and three, the rollout of Disney product offering was delayed to 2018. On a positive note, the backlog for the first quarter -- or first half of BrandLoyalty looks good, and that's 2018.

Let's go to the next slide and discuss Epsilon. Our goal for Epsilon entering (inaudible) percent top and bottom line growth. The strong finish to 2017 allowed us to hit the 5% revenue growth target. And on an apples-to-apples basis, we also achieved the 5% adjusted EBITDA target. The difference to the reported decline of 1% is the restoration of $29 million in incentive compensation versus 0 last year. Essentially, it's a 1-year grow-over that we had to play through, but we needed to restore bonuses in order to maintain highly-competitive compensation packages going forward.

The storyline for 2017 is all about the faster-than-expected stabilization of the technology platform business, which finished strong with 7% growth -- revenue growth in the fourth quarter, as the offering stabilized through the sale of more packaged offerings and record volumes through our Harmony digital messaging platform.

The other 2 big drivers of growth, auto and Conversant CRM, finished the fourth quarter and the year with double-digit revenue growth. Auto continues to benefit from ongoing penetrations in new OEMs, while Conversant CRM cross-selling efforts continue to gain traction, in particular, as the digital capabilities are now embedded in both auto and technology platform offerings.

Let's flip to the next slide and talk about Card Services. Facing the headwinds of the second year of credit normalization, Card Services still delivered with double-digit growth in both the revenue and adjusted EBITDA for 2017.

Looking at the key performance metrics. Credit sales increased 6% for the year with apparel brand partners showing consistent improvement as the year progressed, culminating in strong holiday performance. Tender share grew more than 160 basis points in Q4 and more than 100 basis points for the full year. Average receivables increased 15% for the year. The growth rate in the fourth quarter dipped slightly to 13%, as we divested approximately $800 million of noncore program receivables in December, which hurt receivables growth by about 2%.

Operating expenses were up 7% compared to revenue growth of 13%, so very strong expense leveraging. The net loss rate was 6% for the year, up 90 basis points over 2016. 60 basis points of the increase were due to lower recovery rates and 30 basis points due to higher gross losses. Gross loss rates stabilized as 2017 progressed, as we reached the end of the credit normalization cycle. The delinquency trend or wedge as we call it, closed within 10 basis points at year-end, excluding the transitory impact of the hurricanes.

Turning to the all-important holiday season. Overall credit sales increased 8% with strong performance from our recent vintages. On a comparable brand basis, credit sales were up 5% over last year. The growth was fueled by continued non-store spend, up 11%, which includes digital and web channels. Non-store sales growth was driven by both an increase in buyers and spend per buyer, and that's 3% and 2%, respectively. This growth came from an increase in omnichannel buyers, reflecting our continued strategy to help brands engage customers in all relevant channels.

Card Services continued to expand and diversify the portfolio brand partners, signing nearly $3 billion vintage in 2017. The new partners we have signed this year have been a great reflection of where the market is growing and what consumers value, brands like IKEA and Viking Cruises being valued for their product and experiences they provide, and Build.com being a strong e-tailer in a growing category for ADS.

A key initiative for 2018 is to develop a consumer direct funding program for our Utah bank. It would allow us to take saving accounts and other non-demand deposits, thus diversifying our funding sources and helping funding rates.

With that, I will turn it over to Ed.

Great. Thanks, Charles. If everyone can turn to the slide titled 2017 Full Year. This is sort of more of a commentary part of the call. I don't have a lot to add from a numbers perspective, but I would call out 2 items on the slide.

First up would be under Epsilon. As Charles mentioned, the goal for the year was mid-single digits top and bottom. On an apples-to-apples basis, we believe we accomplished that. Again, the delta between the reported number for EBITDA and the 5% growth rate is the restoration of the incentive comp or bonus program. So if you -- said differently, if you went back to prior year, there was 0 payout, which we don't believe is sustainable in this market. And therefore, we wanted to make sure we were highly competitive as we move into '18. But we didn't want to cloud the actual performance. And frankly, exiting the year at a plus 7 top line is a very good signal as we move into next year. So we think we're getting a handle on that sort of mid-single-digit growth rate and, overall, it feels good.

Card Services, as Charles said, even with 90 basis point increase in loss rates, still managed to deliver EBITDA net of all these provision costs and funding costs and everything else of double digit. So a nice job there.

And then finally, on the earnings per share, we had guided to roughly $18.10. We came in at $19.35. The differences are the following: overperformance would have brought us in at $18.33, so ahead by $0.23 and then a little over $1 coming from the tax benefit, which consisted both of the benefit itself, less about $12 million that we allocated to nonexecutive bonuses coming out of the tax reform act.

So overall, I think it was -- certainly, it was an exciting year. And hopefully, as we move into '18, it will be a little less exciting, a little more fun.

So let's now turn and talk a little bit about the businesses themselves. First, we talk about LoyaltyOne, which consists of our Canadian business, AIR MILES. As Charles, again, sort of highlighted, in my mind, we set out to do 5 things. And I think 4 out of the 5 came through. Most importantly, after the crisis last year, we stabilized the model with a number of changes that we made. And specifically, we came in, in the mid-20% EBITDA margins, which was our goal and very pleased to see that, that has stuck. Second, we renewed our largest client, Bank of Montreal, which is critical to the ongoing success of the program. Third, we wanted to make sure, coming out of the crisis, that we retained our clients. And we had 100% retention rate, so congrats there.

And then finally, we wanted to get the folks excited about the program again. And that comes through with active collectors. Active collectors are all the way back to pre-crisis levels by Q4. So we've set the table pretty nicely.

Probably there are couple of things that we still need to work on. One is, look, the model stabilized, but it has not yet kicked into adding to our growth rate at ADS. So we want to move it back now that it's been stabilized into growth rate mode, and that's the job for '18. The one goal where we came up short was promotional spend from our clients, which represents about 1/3 of all the miles that we issue. It was quite soft. And that's something that we're working on right now to make sure that it's not just the everyday spend that gets rewarded, but we want a big chunk of the promotional spend as well. So overall, I would say, the Canadian business did a nice job following the crisis the year before.

BrandLoyalty, unfortunately, had a -- frankly, there's no other way to put it. It was a very, very poor year. And I think we got a little bit comfortable with 3 straight years of strong double-digit rate of top line and EBITDA. The program just kept rolling along. And then we just had -- frankly, it fell off a cliff in '17. And we had very poor results. The big question for me was, was that an air pocket? Or do we have a fundamental issue with the program? Fortunately, we now have visibility on the fact that we've got a huge book of business for '18. And it looks like it's more like an air pocket. So we're looking forward to having a very robust snapback from BL in 2018.

All right. Let's go to Epsilon. Again, sort of that mid-single-digit organic top and bottom line is what we're looking for. That's what we brought to the table. That's what we're looking to continue into 2018 and start building up a level of consistency that people feel comfortable with. Again, for those of you who recall the end of '16, that's when -- there was a big question on our largest chunk of the business with Technology Platform, which is about over 1/4 of all of Epsilon Conversant, and that was a bit of a melting iceberg. At that point, it was down 13% year-over-year in Q4 of '16. The team did a wonderful job of turning it, not only getting the pricing at a level we want, setting up the office in India, also moving from a very customized model where things take 15, 17 months to get to market to a more standard package, where we're talking about 2 to 3 months. And as a result, what we saw is steady progress and actually exited the year at plus 7% growth for that business. So that's probably the big story there.

Also as Charles mentioned, our 2 big growth areas. Digital CRM and the auto offerings were both up strongly double digit. There was a bit of a concern in Q3 whether the digital CRM, which had fallen off in terms of growth rate, whether that was an air pocket. Fortunately, it was. And as a result, you saw a nice pickup in Q4.

On the negative side, I would say it's a bit more from a macro perspective. We've got a lot of folks since -- it's both a tech offering, but it's also a services offering. And from a macro perspective, we're going to be watching very carefully the impact of the tight labor market for hot skills. A lot of folks competing for this type of talent, and we want to make sure we're on the top of the list.

And then finally, we want to have more consistent, better visibility on the financial results. I think '17 was a good start, but there's still work to be done there.

All right. Card Services, as Charles mentioned, the fabulous wedge that we put out, I don't know, 15 months ago, which is our best predictor of future losses, as everyone knows. We started -- it was up year-over-year about 50 basis points, and that's where we sort of said, "Look, we think losses will be up about 50 basis points." By the end of the year, for me, it should have closed. That would have been a nice finish to the year. Unfortunately, we came within 10 basis points. We did have some noise from the hurricanes. But look, the trend's our friend at this point and it's effectively closed. That means that the outlook for '18 gives us comfort that stable loss rates will follow.

Growth rate in the book of business is quite strong, 15%. Frankly, I feel that's a good number for us to shoot for on a yearly basis. If you were to look back the last 5 years, I think the numbers were closer to 20%. But I think 15% is a very solid number. That gives us a lot of flexibility when it comes to signing new clients and also being very selective on renewals. We did have a record new vintage this year, close to $3 billion. For those of you who don't live and breathe in the card business, $3 billion essentially means if you took all the signings that we did over the past year, whether it was an existing file or whether it was a file starting from scratch and it takes 3 years to ramp up, after 3 years, all of these clients together will add roughly $3 billion to the portfolio. And that's how we look at it. And we have a number of vintages each year that are spooling up.

If you looked at the clients, it's interesting that they differ across verticals and across physical presence versus strictly e-commerce. Obviously, Signet, Build.com, Viking, Guess, Diamonds, Adorama and IKEA represents what we're beginning to see now, which is different from the more traditional model of mall-based soft good apparel. So we continue to see the interest coming from many different areas and that gives us a lot of comfort on what we're going to do going forward. Double-digit adjusted EBITDA growth. So obviously, that includes absorbing the 90 basis point increase in losses, and frankly, getting that done, I thought, was outstanding.

On the negative side, what did we miss? We missed the third-party recovery market, which I'll talk about in a little bit, absolutely plummeted. And as a result, it lowered our recovery rates dramatically. And if you look at that 90 basis point increase in net losses, we thought it would be more like 50. What it turned out to be was gross losses were only up 30. And so, we lost it on the recovery side when the market cratered.

And then finally, a more macro issue. Obviously, everyone knows about the retail environment, so we just need to keep our eyes open in terms of what's going on with our core clients as well as signing the newer ones.

Full year, decent year. Revenue up 8%. Core EPS up 14% with the tax bene; up 8%, excluding the bene. Leverage at the corporate level was very modest at 2.7 and that included the impact of funding over $1 billion for buybacks and capital for growth in the portfolio. So good generation of free cash flow for the company. Frankly, the biggest news that everyone's been waiting for is the stabilization of the credit loss rate. We're there, after 2 years of absorbing increases. And we're poised now to return to mid- to high-teens core EPS growth rate in '18, which is, quite frankly, our long-term model. So it's been a bumpy couple of years. And we're right where we wanted to be for the slingshot for 2018.

We are raising the cash dividend 10% to $0.57 per share starting in Q1. And with that, let's put '17 behind us and move right into '18 and talk about guidance.

So pretty straightforward, of course. There's always a little wrinkle, but the revenue itself looks quite strong. We're looking at double-digit revenue growth of roughly 12%. Against that, there is an accounting adjustment that does not affect cash flow or EBITDA or earnings per share. It's merely geography, but it does change the recognition of roughly $350 million in our business in Canada.

And I'd like Charles to just very briefly talk about it.

So essentially, this revenue standard has been out there for quite a while, but it was really in December when the AICPA came out with a white paper that basically said that if the loyalty program does not take inventory risk, meaning we don't own it, we don't carry the inventory until the point of redemption, we should record it net, meaning the gross revenue minus the cost of the product redemption to get to a net revenue. LoyaltyOne travel has definitely fit that threshold. So what it forces us to do is to look at the travel-related redemptions going into 2018. Recognize we'll now have a net recordation for it. To Ed's points, no impact at all in terms of your earnings -- any of your earnings metrics. Obviously, it bumps up your EBITDA margin appreciably, but it's purely just shifting to a net presentation. This was something that came very late in the year. We had not anticipated it early. Not a big impact overall to ADS in terms of revenue. And like I said, no impact to profitability.

Okay. So again, from our perspective and from how we're going to be presenting on earnings and stuff, we'll pro forma the nums and from our perspective, it's a 12% top line core EPS, even with the big bump-up in '17 from the $18.10 to the $19.35. We're still looking at 16% to 19% growth in 2018. If you want to go back to the original guidance, we said we wanted to do $18.10 this past year, bumping up to $21.50 for next year. The new guidance is -- we came in at $19.35 and we're going to bump up $18 to $22.50 to $23 a share, representing 16% to 19% growth.

All right. Big question, tax bill, tax bill, tax bill. Tax bill will provide us with $100 million -- roughly $100 million a year in additional free cash flow. And that's roughly $1.78 per share. We're going to take that $100 million and, obviously, flow through roughly 2/3 of it. And we're going to take between 30% and 33% and plow that back into the company. We're going to put it back into the company in 3 different buckets. The first bucket would be to accelerate some existing projects that we have on the board. The second would be to establish an innovation fund to get -- to make sure we're on the leading edge of the various digital payment products that are out there. And then the third would be to bolster some employee benefits, whether it's getting the 401(k) participation rate up or holding down any increases on the benefits, the premiums on various health care packages for the employees, we want to help out there. So it's -- we're trying to live the spirit of the agreement as well as what we believe is the appropriate thing to do. And we can talk about that during Q&A.

So let's go into specifically guidance. Let's talk about LoyaltyOne, which is our Canadian business. In BrandLoyalty, we expect high single-digit revenue, low double-digit adjusted EBITDA growth. AIR MILES, we talked about, has stabilized. Now we would like to see it begin to return to incremental growth.

Second, BrandLoyalty, as we also talked about, usually runs at a solid double-digit top and bottom. We did have a very poor year in '17, which means we expect a very strong snapback in '18. We think the World Cup, the Olympics and the Disney deals that we have, it's already looking like the number of deals we signed is going to be a record.

Epsilon, consistency, consistency and consistency. We want to see a minimum of mid-single-digit top and bottom. We expect the tech platform, now that it's turned, to actually turn into growth, along with auto and the Conversant CRM. Those 3 businesses are almost 70% of Epsilon/Conversant.

In Card Services, growth math, as they say. We expect roughly mid-teens growth or 15% growth in average credit card receivables. I get a lot of questions about how do you do the math. And if you look at the signings from '15 to '17, those signings are ramping. And let's say, combined, they add about $2.5 billion to the growth rate in 2018. Plus the core, we expect to be in growth mode -- modest growth mode, add another $1 billion. So you're at $3.5 billion. And then we have budgeted about $1 billion of what we consider folks that either are going bankrupt or we don't feel the renewal process is favorable to us. And so we put in a fair amount of squish there to get us to the 15%, which means it could be a tad bit higher or right at 15%. But that looks like a good number. Gross yields are very stable.

And then let's get into the credit stuff. Stable delinquency rates, best predictor of future losses. Again, we started '17 at about 55 basis points higher. By Q4, you get rid of the hurricane noise, it just about closed to 10, and that's sort of what we're seeing going forward here. We expect stable delinquency rates and that will lead to stable loss rates. And so if you look at the loss rates and the composition of loss rates, what you'll see is it's kind of interesting that the gross rates were up only 30% -- 30 basis points, sorry, last year, better than the 50 basis points we had initially forecast. The consumer is quite healthy. We're seeing very good behavior from the consumer. But really, where the 90 basis point increase came from, 2/3 of it came from, again, as I talked about, the third-party recovery market, which fell off a cliff.

The recovery market for sales of written off accounts got to be very challenging for us. And again, for people who don't live this, there's 2 stages. One is when you go delinquent, some people call that a recovery process for the first -- for the next 180 days. We've always done that 100% in-house. I'm guessing we always will. It's certainly one of our advantages. But what we're talking about is post write-off. So after 180 days, after we write off the account, we sell some of those -- that paper to the third party. That market collapsed. We normally recover between 22% and 25% of gross losses. So if you looked on the chart here, in 2016, 1.8% meant roughly 26% of gross losses were recovered. In '17, we received much less; when we sold all those accounts, our recovery rate fell to 16%. And therefore, it added 60 basis points to the net loss rate or 2/3 of the total. So similar to the past, we know we can do better in-house. That process started in earnest in late fall and will be complete by Q2 of '18.

All right. What's it all mean? We're trying -- Charles and I are trying to answer every single delinquency and loss question here, so we don't have 100 of them after the call. What's it all mean? Q1 will be noisy and will be elevated. The final hurricane flow-through will cause gross losses to be elevated, while we will also be refraining from any third-party sales of written-off accounts, which would normally have lowered -- which will, in fact, lower the recovery rate versus prior year. Basically, we want to recover those ourselves, which means it will come a little bit later in the year, so expect an elevated loss rate in Q1. This is noise only. The delinquency flow suggests stable gross losses, and in-house recovery process will be done by Q2. So after an elevated Q1 to these transitory items, Q2 through Q4, we expect flat to better versus prior year on loss rates and to make up for the Q1 noise with favorable comps in Q2 to Q4.

The 2018 summary is the 2-year credit normalization process is over. We are back as a company to mid- to high-teens earnings growth. Free cash flow will be very robust with over $1.5 billion. Right now, we've only allocated roughly $500 million of that to support regulatory capital and dividends. So over $1 billion of unencumbered free cash flow, and we feel very good about 2018.

And we appreciate your patience, and we'll open it up now for Q&A.

(Operator Instructions) Our first question comes from the line of Darrin Peller with Barclays.

Let me just start off on the card business. From an outlook perspective, the -- I know you're talking about mid-teens receivables growth, or the like again. And I know the year ended off with a slightly slower growth rate. I think it was closer to 10% in the December month. I just want to understand where the confidence is coming from in terms of the portfolio growth and the underlying trends in retailers you're seeing right now. And then just a quick follow-up on the BrandLoyalty -- on the LoyaltyOne business after, please?

Yes, sure. What we do is, obviously, we've got roughly 160 clients. Actually, a lot of the core clients that we had wound up doing a bit better in Q4 and holiday than, I think, a lot of people had anticipated. So we are seeing a little bit of traction there. Also what we do is we look at, Darrin, the vintages that we signed not just in '17, but '16 and '15 and see how they're ramping up and when they're going to layer into the '18 growth pipeline. And so you sort of wrap it all together and what we're seeing is even factoring in at roughly $1 billion that could go away due to a client going bankrupt or renewals that we don't want to pursue, we've looked back -- oh gosh, if you went back prior 5 years, our average growth rate has been 22%. If you layer in the vintages, what you're seeing with the ramp-ups of the IKEAs and stuff like that, the 15% looks certainly doable.

Okay. So it's really organic. I mean, in terms of the strength you're seeing, it's actually organic addition -- additional clients.

Yes. We have not factored in any acquisitions of portfolios in 2018.

Okay. I was just going to ask about -- I appreciate that. On the LoyaltyOne side, just to be clear, I mean, I understand BrandLoyalty has a ton of different new business coming on -- World Cup, Olympics, Disney, et cetera. Just give us more color on the overall confidence on that business, not just -- I mean, on loyalty and the AIR MILES side, specifically, in terms of the overall guidance, high single-digit revenue. What do you need for the AIR MILES business to do for that business overall -- LoyaltyOne, to deliver the guidance you're giving, high single-digit revenue growth and -- I mean, can you give us a little more confidence on what you're seeing there to give you the conviction in that?

Sure. I think the Canadian business, we've set the bar, I think, at a reasonable level. So the goal for '17 was stabilize, get the margins back to the mid-20s. For '18, frankly, Darrin, we're talking 2% to 3%-type growth, which should certainly be doable. The BrandLoyalty, frankly, is going to be the engine driving this bus. And based on what we're seeing in terms of the number of programs signed versus the folks who decided to take a pass last year and sort of wait for the Olympics and World Cup and everything else, it looks pretty robust.

Okay. All right. Charles, just last question and I'll turn it back to the queue. But on the tax side, it's great to see your guidance unchanged, backing out now the tax benefit implicit in your '18 guidance. Just curious in terms of the assumptions for the tax benefit. I mean, I think we had estimated a little bit higher than the $1.50-or-so you're guiding and I know there's a $0.50 reinvestment. But even with that, is there any other items we should just be aware of in terms of conservatism in your tax adjustments or tax estimates for the year?

Sure. So if you look at 2017, our effective tax rate was 33%. We're guiding to 26% to 27% effective tax rate in '18. To Darrin's point, it does assume that we'll get a little pressure from the States. You've read what's going on in New York, New Jersey, California. They could -- the states could look to grab some of that federal savings. So there is little bit of conservatism built in, in that the state rate could go up in 2018 versus '17, which is why we set the bar at 26% to 27% effective tax rate '18.

All right. So it's maybe another 100 basis points or so above what it otherwise would have been. Or...

I can't quantify it, Darrin.

Your next question comes from the line of Sanjay Sakhrani with KBW.

Maybe this goes back to the spirit of some of Darrin's questions. But Ed, maybe you could just talk about where you see the largest risks to your outlook. I know there's a lot of positive things happening underneath it all, but maybe you could just talk about what you're worried about to the extent that you are worried about anything. And then as far as tax reform, can that help Card Services and Epsilon as consumers benefit from that and businesses look to reinvest? I'm just trying to think through the impacts.

Yes. I think they're fair questions. Look, we don't -- the areas that we need to continue to show consistency, that would be Epsilon/Conversant. We've had a good run in '17. Frankly, I think we need to demonstrate we can do that again in '18 before people get really comfortable. From what I'm seeing, it looks pretty good. Obviously, the BrandLoyalty thing was probably the thing that took us all by surprise. And I can -- I see the number of programs we signed based on prior year. We know how that will flow into earnings growth. I don't think we're going to say much more until we actually just print the numbers because we missed so badly during '17. If there's one place of risk at the overall level, it's certainly not with the consumer. What we're hearing out of the thousands of folks who are in our collection areas is that the consumer is quite strong, quite healthy. So I would say, Sanjay, the only thing that would pop up to me, of our 160-or-so card clients, we're assuming a handful, and we watch them pretty closely, are going to have a very difficult time. A couple may even head towards bankruptcy. To the extent, on the retail side, that estimate is off and there's another handful that we haven't thought about, I don't think it will hurt '18, per se, but what it will do is it will make for a tough grow-over for '19. So from a macro perspective, it's more retailer-specific, a handful we've sort of identified. It's no more than that handful.

And on the tax reform, Sanjay, we would say, you've seen many companies come out announcing they're going to put more money in the hands of their associates, which could lead to card holder spending more money in 2018, which could be beneficial to our card operations.

How about Epsilon? Are businesses looking to reinvest some of the upside?

Yes. I think from some of the reinvestments, I would call out probably 2 areas. The first in terms of reinvestment would be -- we have always -- given the size of the card business today, it was a lot easier when the file was a couple billion as opposed to heading towards $20 billion. We will be rolling out and launching a new source of funding, which will be on the consumer deposit side out of our ILC facility in Utah. And obviously, that requires a consumer platform. It requires marketing. We'll have to come up with some fancy slogan or catchy slogan, I guess, that will get people excited about what's in their purse. And then the other area would be with the CRM business at the Conversant side of it. Given the growth rates there, we will probably -- not probably, we will absolutely be ramping up the size of the sales force effort and, perhaps, focusing even more on not just the large clients, but also mid-market. So those would be the 2 big areas I would see initially for investment.

Your next question comes from the line of Vincent Caintic with Stephens.

So first, I have to ask a question on credit. But on the in-house collection efforts that you've had so far, so I know it's early days, but given the 60 basis point headwind you had in 2017, what are you seeing so far in your experience that maybe can give us some encouragement? And then also, how you think about the level of credit reserves that you keep. We've had some good provision performance recently, but just kind of wondering how you think about credit reserves.

Sure. Great question. Yes, we went through this whole recovery in-house versus selling to third parties post great recession, where we raised the same type of issue and we had to bring things back in-house. All right, it's pretty simple math. At a certain level, it's more profitable to push it outside. What we have found, so far, because we're about 6 months into this process, is that much like we had anticipated, if I were to put a number on it, the recovery rates internally are probably somewhere around 20% to 22% of your gross loss rates. So against a marketplace that was paying us 25%, it certainly made sense to go outside. However, that 25% dropped all the way down to 16% during '17. So as we gradually move that up to the 20%, 22% in-house, that's what we're seeing. And that's why I think we feel very good about where the net loss rate trend is going to be this year. From a gross loss perspective, it's going to be flat.

In terms of credit reserves, we tend to carry 11 to 13 months in terms of forward loss coverage based upon the trends we see in losses. Beginning back in 2015, we carried up to 13 months coverage all the way through the third quarter of this year. We trimmed it back to 12 based upon favorable trends. Fourth quarter, we kept it at 12 months forward coverage, which is likely where we'll stay until we see it stabilized and potentially drop lower. The key thing to keep in mind in looking at the reserve quality or the reserve coverage, if the Signet file came on late in the year, close to $1 billion, the losses embedded with the file recorded net, so there is no reserve attributable to it. So you always had to focus on the reserve divided by the reservable AR, which would not include the bulk of Signet in that number.

Okay. Got it. That makes sense. And then separately on the -- I mean, the loyalty programs and kind of the loyalty updates, you talked about 15% AR growth in the card side and the signings' activity that you're getting there. I'm just kind of wondering how also that might translate into some growth on the Epsilon side. It seems like -- there seems to be more engagement on the overall loyalty offering that you have and if you can just give more perspective on that.

Sure. Yes. I mean, 100% of our card clients are on the loyalty platform developed by Epsilon. The various digital distribution channels, such as email, our card clients would use the Epsilon. We are, right now, beginning a bigger push to get the unique sort of identification platform that is Conversant CRM into the hands of our card clients. So we would expect that to be a big area for us to push. I would say, across the board, our card clients are asking not only for the type of insight into their existing customers, but they also want to be able to say, "Let me take a look at my best customers. How do I go out and find people similar to that, that may not be customers today?" And that's the beauty of what Conversant can do with its unique ID. We can take all those characteristics and then go find these folks and, on behalf of the clients, get something really exciting that could motivate them to show up, either online or in-store at our clients. And that's sort of a new area that we're really beginning to expand. So it's not just taking care of the existing best customers. That's obviously job #1. But it's now moving into, "Help me take a profile of my best customer and then go out into space there and find someone similar and then get them all excited about the brand." And that's -- that is an area that's getting our clients very excited because it's going to drive sales, and that's sort of the bottom line. So that's the big push.

Your next question comes from the line of Jason Deleeuw with Piper Jaffray.

So I just want to be clear on the credit recoveries. Sounds like we have some noise with the hurricanes maybe in the first quarter. But taking it in-house and then the changes you've made, I mean, shouldn't the recovery rate -- shouldn't there be a tailwind to your net charge-off rate? Or are you expecting that still kind of flat versus the 2017 level?

That's absolutely going to be -- when it's fully up and running, it should be -- you should think 20%, 22% versus 16% that we did in 2017. So from that perspective, if you had flat delinquencies, which we have, and you're going to see flat gross loss rates and you've got this tailwind on the recoveries, as the year progresses, it certainly suggests that we're going to be in decent shape.

Good. And then just on Epsilon. The technology platform revenue, it grew 7% in the quarter, so it was a nice improvement on the growth rate there. Can we sustain that level? And can you just help us understand kind of the drivers of that and then kind of the margin profile of that revenue versus the rest of Epsilon?

Sure. I'll let Charles do the margin. In terms of the product itself, what we basically had there was the traditional Epsilon product release, big iron-type databases and loyalty platforms that were highly customized for each and every client. And they were just massive, right? They dealt with tens of millions of customers of clients, all the bells and whistles in the world. And they were, frankly, pretty cool. The problem was they cost an awful lot of money. And it took anywhere between 9 to 15 months for this thing -- for these things to be delivered and cranked up for the customer. That just wasn't feasible anymore. We were frankly a bit slow to make the pivot to transition to a lower cost, quicker delivered product. And that's what took us a full year, maybe 15 months to do. We opened a large office in India, which helped with the price point. It brought us very competitive with some of the SaaS-based solutions out there. In addition to that, we, frankly, could no longer sit there and spend 6 months getting all the bells and whistles on these platforms. And we said, "Here's what we can deliver in a 2-month time frame." And so we sort of marginalized the offering. And lo and behold, with the new price point and the time to market being brought from 10 months to 2 months, we found a very receptive audience because a lot of these folks, frankly, don't want to run the things in-house. There's a big market for companies that want to run and they get SaaS-based solutions and they hire IT staff and they go do their thing. So it's also a very big market as long as we can hit the price point and we can deliver in a timely fashion, where they're saying, "Look, I don't want to be the ones responsible for figuring out how the watch is made. I just want to know the time." And so that's why I think this is something that, right now, certainly seems sustainable based on the book of business we've signed.

Yes, Jason, from an EBITDA margin standpoint, think of technology platform being in the 18% to 20% range.

Your next question is from the line of David Togut with Evercore ISI.

Could you drill down a little bit more into the mid-teens portfolio growth assumption for '18 for Card Services? More specifically, could you talk about your assumption for same-store credit sales and what the breakdown of that would be between expectation for same-store sales of your existing retailers versus expected wallet share gain?

You bet. I think, David, it's -- we are taking a fairly conservative view based upon what we've seen over the last year or so. So for the core clients, the folks who've been with us for 3 or more years, what we have seen is they've gone from probably negative 2% or 3%, to the last quarter, they were probably -- it just broke the surface of the water, probably 0% to 1%. So we're assuming same-store sales are probably not going to be better than 1% for those core retailers. With tender share, we should be able to get to somewhere between about 4% -- 4% to 5% in terms of sales that will -- sales growth on our cards from those core. And so as you know, you go back half a dozen years or so, that 5% used to be closer to 8% or 9%. But it looks like factoring in same-store sales of about 1 point, we'll get tender share, of course, and so we're thinking 4% or 5% and then the rest will come from the big books of businesses that we've signed in '17, '16 and '15.

Got it. And then just a follow-up on capital allocation. Charles, you highlighted $1 billion of unencumbered free cash flow. How are you thinking about capital allocation priorities with your stock at its current price, let's say, balanced against potential acquisition and/or divestiture opportunities?

I'd say, David, we're going to be flexible on how we use that. I'd say, right now, M&A is probably pretty low on the radar screen. We could use some for buyback. We could use some to pay down debt over the course of the year. Really, the focus Ed and I have is to make sure that all 3 of our segments are firing and back on track. And based upon those trends, we'll determine how we deploy that capital. So I'd say, conservatively, expect it can go to just pay down debt, delever a little bit. Even though we're in a very good shape now, it makes sense. We could do a modicum of buyback. Obviously, we have the dividend that's there as well. What I would say is just expect M&A to be a relatively low priority at this point. Really, the focus, as I said, is to get all 3 of our segments back firing.

Your next question comes from the line of Andrew Jeffrey with SunTrust.

Really nice to see Epsilon back in the mid- to high single-digit range. I'm wondering if you could provide just a little more color. Auto has been really strong. CRM is doing well. So I guess, a couple of questions. First, on auto, could you characterize how you think about the cyclicality of those -- of that end market versus sort of structural demand? And then within the CRM offering, are there any verticals, in particular, that you'd call out as being important drivers? And again, I'm just trying to get a sense for the sustainability of the growth in those 2 really important drivers of Epsilon.

Yes, good questions. On the auto side, what's interesting is the bulk of the work that we do, both for the OEM and for the various dealerships, the thousands and thousands of dealerships, really is more after the sale is made. So it's much more in terms of service. So that doesn't have the type of variability you'll have with everything is driven off of sales, new car sales. So we're the folks who will sit there and, hopefully, use predictive analysis and more and more of the telematics, the data that comes out of the dashboard that basically will say, let you know through a text or an email or a big card in your mailbox that says, "Hey, Andrew, it looks to me like you're about 1.5 weeks away from when it's time to get your oil changed. Come on over to the dealership and we'll give you a great deal." That's the type of stuff that we'll be doing. As more and more data comes from the telematics side of it, we'll be able to bug you all day long about your transmission level and all these other things that we'll be making sure you don't break down out there in California. So that's sort of it on the auto side. From the CRM side -- so the auto side shouldn't all be that cyclical. From the CRM side, look, the huge verticals, thus far, have been retail. Auto is certainly looking bigger. What else, Charles?

Financial.

Yes, financials, the banks for sure. There's a couple other verticals we're beginning to look at that hold promise. So we want to take the offering and move it from just sort of a retail-type offering to auto financial services. CPG looks promising. Those are the verticals we'd be focused on.

Okay. And just one quick one, if I may. In technology, to the extent you're winning more there and your win rates look better, are there -- are you -- are these competitive takeaways? Or are they buy versus build decisions that you're swaying?

I would say it's less competitive takeaways. I think it's -- the market continues. It's a very -- it's a growing market, right, as you know. It's really -- comes down to the CMO needs to decide whether she wants to have the thing built in-house and have a big tech stack, meaning let's get a SaaS-based solution and surround it with 100 tech people or now that we can have a similar price point to that offering and do it ourselves, is it okay to give up controls of the guts of this thing and just have the results pour through into her shop so he or she can just have a bunch of analysts to look at the results? So that's -- the sandbox itself is getting bigger and bigger, so I would say the vast majority of it are more and more folks who are getting into the whole concept of, "Hey, we can use this data in loyalty platforms to understand my customers and market to them." It's huge.

Your last question comes from the line of David Scharf with JMP Securities.

First question on the card side. Just a point of clarification. Ed, when I look at the components of your portfolio growth guidance this year, is the $1 billion subtraction for retailers that ultimately may become noncore, non-growing, is that just a guess and a cushion? Or are there specific programs you've identified?

Yes, I think it's a case. Think of it, David, we divested $800 million in December, so you know for the full year, in terms of average, that's going to be part of it. So basically, it means that there's another $200 million potential over the course of the year that we could look at and evaluate, either do a held-for-sale or could be a retailer that goes away. So said simply, the bulk of it was the divestiture we made at the end of December.

Okay. Got it. And the mid-teens, Charles, that's a year-over-year -- that's a 12/31 versus 12/31 guidance apples-to-apples, that's average.

Average '18 over average '17.

Got it. And then lastly, shifting to Epsilon, you noted the tightening labor market. Curious, you obviously took a lot of initiatives in recent years on the off-shoring side to address that. Do you feel like, based on the mix of skill sets that are needed, the composition of Epsilon as it exists today, is off-shoring as an option pretty much run its course? Or are there additional opportunities there?

Yes. Again, we view it a little differently as from off-shoring. It's actually an office that we built over there. These are Alliance Data employees just like the folks here. So it's very much keeping it within the family-type approach. I would expect we're going to have 2,000 folks over there. And no, I don't think it's run its course in terms of the cost model itself. The proportion that -- of our associate base that's over there versus here is probably getting to be at the point where we find it's comfortable. And I don't think there'll be a huge shift, either way, going forward. But in terms of the benefits on working across, both India and here in the states, we're just beginning to see that. So I would expect, while there's a nice chunk that we've picked up thus far, there's probably another chunk to go.

Okay. That's helpful. And just last question quickly on the AIR MILES business. You've noted, obviously, for several quarters the challenges in issuance growth associated with the grocery segment. Is there something structural that's changed among a number of the large sponsors in terms of how much they're willing to spend on promotional activities? Or did you feel like 2017 may have just been a breather on their part as they assess maybe the fallout from the expiry events? Trying to get a sense for visibility there.

Yes. I think, David, it's hangover from what took place in the fourth quarter of '16 around the expiry. I think that's really what it was. Obviously, we had a number of things come on this year. We need to get Bank of Montréal renewed, which we did around the third quarter; that should help. And I do think in the grocery vertical, we just need to get them back satisfied, get the consumers back in their stores and that, by itself, will lead to getting promotional activity back. But I would attribute it to purely hangover from the Q4 '16 event.

Okay. Thanks, everyone. We'll see you after Q1.

This concludes today's conference. You may now disconnect.