My SoFi & PayPal earnings reviews sent today.

Missed revenue estimates by 1.5%, but met its foreign exchange neutral (FXN) revenue growth guidance. FXN revenue rose 1% Y/Y.

Slightly beat EBIT estimates and met $0.93 GAAP EPS estimates. EPS fell 6% Y/Y mainly due to higher promotional activity amid a more “cautious consumer.” Operating efficiency gains discussed later helped to offset this.

Starbucks reiterated the slashed guidance it offered last quarter. This includes low single digit revenue growth, slightly negative to flat global comp sales, 6% store growth and slight EPS growth. These were all aggressively cut last quarter. I get that this isn’t all that impressive, but it will likely bring some stability to tanking estimates – at least for now.

SBUX trades for 20x 2024 earnings. EPS is expected to be flat this year and compound at a 13% clip during the next two year period. Here’s how that compares to its historical norms;

As I’ve written about in detail recently, Starbucks has been struggling mightily with throughput and in-store service in the USA. Leadership told us that if there’s one thing we take from the call, it’s that the upside potential to drive better service and efficiencies in its stores, order flows and supply chain is immense. Starting with throughput, customers are placing orders, waiting too long and canceling before paying. This is wasting labor hours, raw materials and money. There are a few things it’s now doing to address this issue. First, it is implementing phase one of its “Siren Craft System” in all U.S. stores this week. The first part of this (which will be a staged roll-out) includes updated employee and manager training, some minor hardware tweaks and some process improvements at its stores. It also includes an added peak time employee to handle orders, which has been wonderfully impactful for throughput and margins at Chipotle. There’s very little CapEx for phase one and changes have been well-received out of the gates. The next parts of this rollout will feature hardware (like new blenders) and store model updates, which will involve CapEx and will reach 40% of its U.S. locations by the end of next year. It will be very strategic about which stores it chooses first, as 10% of its U.S. locations are customer service outliers that could use this help the most.

This is the main event for diminishing throughput issues, but there are other tailwinds here as well. It will retrofit its espresso machine to boost throughput by 15% with no quality sacrifices. It’s also extending mobile order pay (MOP) to non-loyalty members to cut time to service for that fulfillment avenue as well. It’s getting closer to its partners and helping them more directly with labor allocation, scheduling and more. This has driven down partner churn for the firm, which should mean better continuity and service. The Siren Craft system will be a big help here too by operationalizing this hands-on help.

All in all, these initiatives are expected to cut time to order fulfillment by 10-20 seconds. It has struggled to time this fulfillment time by a couple of seconds over the last few years. There are already small signs of things looking a bit better. Inbound customer service calls about orders taking too long fell 50% Y/Y during the quarter. It also updated app algorithms to improve order ready accuracy rates by 50%. Keep building on that needed progress.

Intense Chinese pricing competition continues. It’s not playing this game, which is why EBIT margin maintenance there has been better than revenue maintenance. It began to see some improvement in weekly sales there during the quarter, but it is in no way out of the weeds here. Encouragingly, its highest value loyalty members continued to grow there. Specifically, it added 1.6 million to reach 22 million while “customer connection scores” set new highs. This sounds similar to a net promoter score (NPS). And while things in China are tough, it is not slowing down store openings in the least. It has ample opportunity to open more stores with sub 2-year payback periods and gaudy year one return on investment figures. It’s committed to continuing to grow in that nation and doesn’t see this temporary weakness as altering its longer term market view. It’s playing the long game there. To date, it’s only in 900 of its potentially planned 3,000 markets there. Despite this optimism, it is toying with the idea of adding more strategic partners to lower overall capital intensity and risk. This is something I would support.

The firm’s cost savings efforts are ahead of schedule, as it thinks it will do better than its $4 billion in net savings over the next 4 years. There are a few sources of these “outperforming” efficiency gains. First, subtle tweaks to stores have yielded a 110 bps improvement to store operating expense intensity. We weren’t told where the gains came from, but the Siren Craft System should certainly build on the progress. It’s also working hard with its supply chain partners to negotiate better deals, without sacrificing quality. This netted another 100 bps in cost savings during the quarter. Finally, it sees G&A falling from 7% of sales to 6% of sales going forward (so another 100 bps of pocketed efficiency gains) as front-loaded store investments have now taken place.

Somewhat counterintuitively, new Starbucks stores around the globe continue to perform very well – just like in China. For example, a new store in Joplin, Missouri netted a year 1 ROI of 65% with 30% cash flow margins, a sub 2-year payback period and $2 million in average unit volume. It has been highly incremental to overall revenue. Starbucks is going to focus on tier 2 and 3 cities like this one, where it feels the opportunity is most untapped.

There are two pieces of the SBUX traffic story – loyalty and non-loyalty members. Loyalty represents 60% of revenue and is seeing healthy frequency growth. Non-loyalty is the remaining 40% and is seeing material frequency declines. Starbucks hasn’t let non-members use MOP until now, despite 25% of non-loyalty members wanting that implemented. It has finally made that change and expects to see positive traffic impacts, which should build on 7% Y/Y MOP transaction growth this quarter.

Loyalty members spend more than non-members and SBUX is most confident in its ability to convey promotional messaging via that app. And they’re right, promotional activity within the app carries better targeting levels and superior returns for the firm vs. any other channel. And there’s a ton more to do here, with just 14% of its transactions coming from promotions vs. 29% for the average competitor. It can boost that 14% while also collecting higher store revenues, as the campaigns have shown to uplift incidence rates for menu add-ons.

So? This is a large focus area for the firm. It’s determined to drive loyalty growth through MOP openings and better execution going forward.

Was this good? Not at all. But everyone in the world knew it would be bad and sentiment could not have been worse. When that happens, all of the bad news is usually priced in, and that’s what I think is the case here. Throughput gains are showing signs of working, macro won’t suck forever and China comps will begin to get extremely easy as all of this work starts to bear fruit. This is an iconic brand going through a rough patch that I view as entirely navigable. I see brighter days ahead.

Q4 2021 is when hefty CapEx to support cloud infrastructure initially began. Front-loaded CapEx weighed on quarterly FCF margin (OCF is pre CapEx as FCF = OCF - CapEx).

Reiterated 10%+ revenue & EBIT growth for 2025. It sees slower Azure growth for the first half of FY 2025 to build capacity to meet demand. It then sees growth accelerating during the second half of the year. It sees EBIT margin contracting 1 point Y/Y due mainly to gross margin pressure as it invests more in this infrastructure. It will control OpEx to ensure there’s not any further contraction. Finally, CapEx will rise Y/Y.

Microsoft trades for 31x FY 2025 earnings. Earnings are set to grow by 13% this year and by 17% the following year. Here’s how its multiple compares to historical norms:

This call was eerily similar to the last few Microsoft calls. It functioned to walk us through its GenAI approach and everywhere CoPilot is driving adoption and monetization potential. Some of this will be review for consistent readers. 

Microsoft is determined to “drive innovation across a GenAI product portfolio that spans infrastructure and applications.” It’s looking to be a full service provider of GenAI utility – between infrastructure, apps and a bevy of models to choose from. As of right now, the company is determined to make the necessary foundational investments to support all of the GenAI infrastructure and software demand that it sees coming in the decades ahead. As a result, CapEx will rise Y/Y to support future demand. Leadership did tell us that about 40% of this CapEx is flexible and can be eliminated if “demand signals” sour. Its desire to operate in every layer of GenAI also gives it an ability to allocate any potential capacity surplus across several use cases. It can deliver that capacity to clients, send it to whichever Copilot app is commanding the most attention, or just use it internally for its own productivity gains. There’s far more risk to underinvesting here for the mega-caps vs. over investing. Alphabet said the same thing last week. Here’s what CFO Amy Hood had to say about CapEx allocation across the GenAI layers:

“It’s incredibly flexible because we've built a consistent architecture with the Commercial Cloud and the Azure AI stack. Regardless of whether the demand is at the platform or app layers or through first-or-third parties, it uses the same infrastructure. So it's a long-life flexible asset.” – CFO Amy Hood

For now, it seems like Azure and Microsoft need all the capacity they can get. Azure’s growth came in at the low end of the company’s guidance range because of capacity restrictions creating a bottleneck to fulfill demand. Some macro weakness in Europe also contributed, but this was a material headwind. As a result of needing to catch up on infrastructure investments, it sees slower Azure growth during the first half of its upcoming fiscal year vs. the second half.

Microsoft Azure’s “share gains accelerated” in the firm’s fiscal year 2024. Pace of migration remained rapid and Azure Arc remained a go-to facilitator of those migrations. Azure Arc, which enjoyed 90% Y/Y customer growth, is its platform for enabling integrated multi-cloud access to a client’s apps and data. Most large companies don’t want the vendor lock or rigidity associated with using one public cloud. Most want many. This makes Microsoft a better partner for enabling that, which is especially important in the GenAI age of access to data being the difference between leading or following in an industry. Sticking with the data theme for a moment, Microsoft’s data platform (called Fabric) is now being used by 14,000 customers and was a key cross-selling tool in Azure AI deals. Specifically, AI customers using its data tools rose 50% Y/Y.

Summing things up… Microsoft’s massive data scale gives it the unique ability to train models more effectively than most; tools like Arc and Fabric make it easier to access all of this data; Azure provides the models, service and tools to build apps using this data; Copilot is the leading GEnAI software tool as we speak. It truly does create GenAI value in several areas, which is likely why it has been the most successful enterprise software monetizer of this new tech wave.

Azure AI is the firm’s suite of GenAI cloud tools to let developers and clients create, store and deploy applications and workflows in a fully managed fashion. It offers natural language processing, anomaly detection, speech-to-text and more with a large variety of GenAI models for customers to choose from. This Models as a Service tool, which saw 2x Q/Q growth and added clients like Adobe, is what enables choice of pretty much any mode you can think of. Paid customers for this product rose. In aggregate, Azure Ai customers rose 60% Y/Y to 60,000, with spend per customer also rising.

GitHub Copilot users rose 180% Y/Y to reach 77,000. Microsoft has now brought this devops company to a $2 billion revenue run rate, while GitHub Copilot is now larger than the entire firm was when Microsoft bought it. Good purchase. It’s adding Copilot Workspace for developers to automate part of the software creation cycle, which will invariably be a highly relevant addition for GitHub customers specifically.

Power Platform is ripe for GenAI augmentation. In essence, it helps customers build apps, and automate work in a low or no code manner. It’s easy to see how CoPilot could turbocharge this utility, which explains 45% Q/Q growth in Power Platform AI users.

Copilot for Microsoft 365 daily active users doubled Q/Q as customers overall rose 60% Y/Y. Most customers are promptly coming back to Microsoft to purchase more seats based on the product resonating and creating value. 10,000 seat customers doubled Q/Q. Copilot here can schedule and organize meetings and delegate to-do lists, among other things. Copilot Studio enables the seamless building of custom applications on top of Microsoft 365. Users of this product rose 70% Q/Q. If you’re noticing a theme of GenAI add-ons gaining rapid customer adoption, you’re right. And again, this theme in the world of enterprise software is quite rare.

This would be a fantastic quarter for anyone except King Microsoft. For this iconic enterprise, things were merely fine. The slight cloud miss is the main thing people will focus on, but that seems to be supply constraint-related more than anything. All in all, Microsoft continues to march on as one of the most impressive software compounders in history. This merely reiterates that now obvious opinion. Congrats to shareholders on some more wonderfully boring execution. Satya Nadella is kind of good at this… I guess.

I got a few requests to cover Lemonade after its earnings release tonight. The call is tomorrow, and I’m hoping to have a full review with Meta and Mastercard done for tomorrow night. No promises. I am a one-man band and this week is hectic. From a 30,000 ft. view, there’s nothing alarming here. It looks like the company sees a weaker Q3 and a stronger Q4 than the street expected. Annual guidance was maintained as was path to profitability; liquidity risk is gone; loss ratios show continued signs of maturation. I still have much more work to do on the report and I always reserve the right to change my mind. From what I’ve gotten to read so far, I like what I’m seeing.

As I said a few weeks ago, this investment remains compelling to me, but wildly speculative and volatile. For context, this afternoon’s move only gives up about half of its gains over the last 30 days. It will continue to violently swing around. Bears will continue to be loud when that swinging is to the downside. I’ll continue to dispassionately stay the course as long as fundamental data warrants doing so. For now, I wanted to include a snapshot. More to come this week.

Results:

Annual Guidance (Q3 light, which implies Q4 strong):

Balance Sheet: