Today’s Piece is Powered byAries:
1. Our Take on the Stock Comp Debate
a) When Stock Comp is More Appropriate
Stock compensation accounting is a hot topic on Twitter. We wanted to share our take on the matter, when equity compensation is more palatable, and how we think it’s best to account for it.
Growth companies are typically voracious users of cash to fund operations and future prospects. Especially for this type of company, a cash pile is always finite and access to debt capital is materially more expensive than a company like Apple. So? This is when hefty stock compensation is most understandable. Electing to compensate executives and employees partially via stock is a great way to preserve that cash pile without it becoming a bottleneck. It effectively extends the flexibility of payroll. Furthermore, equity compensation (especially with long-term vesting timelines) is a great way to align shareholder and employee interests. If an employee base’s compensation is tied to company or stock performance, that will motivate the staff to provide optimal performance.
Ask yourself as a shareholder of CrowdStrike, Datadog, Monday or some other growth stock which would you prefer:
- More dry powder to fund R&D, marketing, buybacks or M&A with 2-3% annual dilution while being able to attract & retain the talent required to sustain high growth? Or…
- Less dry powder restricted growth, marketing or M&A spend with limited ability to attract & retain talent but with 0% annual dilution?
For companies compounding profits well over that 2-3% annual dilution estimate, the answer is clear. We want them to have the talent they need, we want them to have the flexibility to allocate capital and we are ok with accepting a bit of dilution if the firm can deliver rapid profit compounding. Consider an example with two firms (A and B). Both start at $10 in FCF with 100 shares outstanding. Firm A compounds FCF at 30% for 3 years and dilutes shareholders by 3% per year. Firm B, with more liquidity restrictions compounds FCF at 20% for 3 years and dilutes shareholders by 0% per year. At the end of the planning period assuming no capital appreciation, Firm A generates $22 in FCF annually with 109.3 shares outstanding. This represents FCF per share of $0.20. Now consider firm B. At the end of the period, firm B generates $17.30 in FCF with 100 shares outstanding. This represents FCF per share of $0.17. As FCF per share is a key driver of shareholder value and accounts for dilution, Firm A’s strategy is preferred.
With a commitment to stock-based comp for growth, there are 3 important considerations:
First, stock comp intensity must diminish over time as a company matures and growth slows. If that doesn’t happen, 2-3% dilution annually will more materially cut into returns. It’s the recipe of elevated comp, no downward comp trend and slowing growth that makes hefty dilution “icky”. Second, there are limits to how much stock comp is appropriate. That limit rises with faster growth and higher quality re-investment opportunities, but it’s still there. When, for example, we see some firms diluting shareholders by 5% annually while other aggressive equity comp users are closer to 2%-3%, that does spell some concern -- especially if there’s no consistently brisk downward trend in intensity. You’ll notice we used dilution comparisons vs. “stock comp as a % of revenue” comparisons as we think that’s much more relevant and valuable (explained next section).
The third consideration is perhaps the most important. The leadership team must be effective and successful in allocating capital. If you don’t trust a team to do so, you should find another firm to invest in.
b) Stock Comp Accounting Best Practice
There are two ways to account for stock compensation intensity: Avoiding the add-back in our FCF calculation and tracking FCF per share growth. As per the example, we greatly prefer the second method but will walk through both briefly here.
The first method is to ignore the add-back on the cash flow statement. The cash flow statement starts with net income before making a series of non-cash adjustments. Stock comp hits net income but is then added back on the cash flow statement to arrive at operating cash flow (and then subtract CapEx to get FCF). FCF - stock comp can give a decent sense of comp intensity between firms, but it’s not the correct way to account for it. Why? It goes back to the idea of stock comp not being a cash expense. What is the true expense? Dilution. Forgoing the add back completely ignores the dollars actually flowing onto the balance sheet to fund organic growth and M&A.
To us, it makes much more sense to simply track FCF per share growth rather than FCF alone. This fully accounts for all dilution and eliminates the noise from a certain kind of equity type called performance share units (PSUs). PSUs are accounted for under GAAP rules in a somewhat misleading way when looking at “stock comp as a % of revenue.” Why?
If, for example, a company like SoFi issues PSUs at strike prices that don’t hit under the vesting period, those shares are never created. REGARDLESS, this still counts as GAAP stock-based compensation dollars and still impacts net income. This can make stock compensation look more aggressive than it actually is. A company cannot add back that previously incurred stock comp despite it never actually being awarded. That is another reason why tracking the dilutive impact of compensation offers a more accurate, reliable sense of stock comp intensity. Stock comp can be a ridiculous 40% of sales for this reason while dilution remains more reasonable.
c) The Pandemic’s Impact on Stock Comp Intensity
The pandemic was a strange time. A large cohort of growth companies vastly over hired during the period. I can count on two fingers the number of growth companies off the top of my head that didn’t over hire (The Trade Desk & Revolve). Firms assumed pandemic trends would be more durable and revert less abruptly than what actually played out. So? Demand growth slowed while payroll growth continued at a rapid pace throughout 2022 for many. Stock comp intensity as a percent of revenue skyrocketed. The flip side of that has been unfortunate yet necessary layoffs as well as product/geography rationalization across mainly the tech industry to right size workforces. Much of this right-sizing happened throughout 2022 and 2023 which will lead to slowing payroll and compensation growth starting this year and going forward for many of the firms that got too excited.
Aries is a brokerage offering low cost tools like free option flow data and transactions to investors across the globe. It distanced itself entirely from a payment for order flow model to pave the way for operations in countries like the U.K. The interface is slick and intuitive and today, they’re looking for smaller investors to directly invest in their fund. If you’re interested in hearing more, you can pre-register to learn directly from the founding team here.
2. Meta Platforms (META) -- WhatsApp Show me the Money
a) WhatsApp
WhatsApp has nearly 3 billion users. It’s THE go to messaging platform in key geographies like India and throughout the planet with inroads now being made within North America. The financial potential for this app has been mammoth for years… but now Meta is finally aggressively committing to letting monetization ramp.
WhatsApp is in the process of rolling out a slew of paid features -- some are now available with others in the works. WhatsApp for Business tools like click-to-chat links, shoppable personal links and multi-device customer service have arrived. Amazon uses the customer service tool in India today. Flipkart and BookMyShow (popular Indian ticketing service) also lean on the app to conduct operations; India serves as the app’s most promising geography for now. But it’s not just India. Overall WhatsApp Business accounts have quadrupled to 200 million since 2020 for a CAGR North of 58% during that time.
This week, to more meaningfully extract value from this explosive growth, Meta announced a roster of new paid messaging tools for organizations to more closely and directly connect with their consumers. Messaging is clearly a large part of WhatsApp’s future and we expect releases like this to be frequent. Considering its unparalleled reach, it’s easy to see how businesses like Shein would be eager to line up to use it. There are several levers to pull to turn this app into a monetization juggernaut. That incremental revenue opportunity is nearly entirely ahead of Meta, and the company has decided it wants to take advantage today.
Other monetization examples include:
- WhatsApp Pay for money transmittance (which MercadoLibre is currently testing).
- Dipping into ad-based monetization.
- Potential premium features for a consumer subscription (not announced, just speculating).
There are two immensely exciting growth vectors for Meta beyond Reels. Neither are Virtual Reality -- although that could potentially turn into a 3rd eventually. First, its AI, chatbot and supercomputer investments should continue sharpening its discovery engine to juice engagement and ad targeting to boost return metrics. This should mean more impressions and higher CPMs. Apple’s impact on its business continues to diminish as these investments bear fruit. With Meta’s unique open-source approach to AI models, it’s easy to see how excited developers eager to build on top of models like Meta’s could enhance utility and monetization further. It could also roll out a premium consumer subscription if it wanted to, but that has not been announced or hinted at as of yet. The second opportunity is WhatsApp and inning one of monetization has now begun.
b) European Union (EU)
The Verge published a very interesting article on Meta this week. The article discussed a new ad product that Meta will test in the EU this year with Android developers. The ad product comes with a bit of a different aim vs. what Meta has taken in the past. Interestingly, it will allow Facebook and Instagram users to download apps directly from within the sites.
The EU’s Digital Markets Act (DMA) recently passed is what is opening the door for this release per the article. DMA requires Apple and Google to permit payment and downloading methods aside from solely their own. This enables sidestepping hefty take rates with Meta not planning on charging any in-app fees as part of its new release. Meta thinks its more favorable fee structures along with native app downloading (not bouncing from page to page) will augment developer conversion and could make this a viable alternative to Google and Apple over the long term. Microsoft is planning a similar launch in the EU as app buying competition ramps.
c) More Meta News
- The company is debuting a game pass for its Quest headsets. The subscription costs $7.99 per month or $59.99 per year and comes with 2 titles per month. The gamers maintain access to all previous monthly titles as long as they remain active subscribers. This is one way Meta will look to make Quest a slightly smaller drag on profits going forward. It will not try to generate gross profit from hardware sales like its new competition Apple does. It will sell it at cost and monetize elsewhere.
- Citi came out with a positive note on Meta platforms. It sees quarter to date ad load for Reels improving (with more improvement in June) and online advertising bottoming. It also sees ad revenue growth of 14% for 2024 which is well ahead of sell side consensus growth around 11%. What was their price target? We don’t care… but also know that some of you do. It was $360.
- Meta Verified is launching in Latin America this week and will roll out globally this year. Using Bank of America’s 12 million 2023 subscriber estimate, this should quickly grow to $1 billion in incremental EBIT (nearly 4% of total 2022 EBIT) while continuing to grow thereafter. A nice piece.
3. Snowflake (SNOW) -- Snowflake Summit 2023
a) Snowflake Summit 2023 -- Embracing the Generative AI Opportunity
Microsoft and Nvidia Partnerships:
Snowflake announced an extension of its partnership with Microsoft Azure. The partnership will work to integrate Azure’s AI models and apps into the Snowflake ecosystem to feed those algorithms more data for better seasoning and compliance. It will also include a joint go-to-market approach to drive adoption of these newer Azure use cases and new Snowflake tools discussed below. Snowflake’s cloud agnostic approach helps it to effectively tear down data silos for broader cohesion and more valuable insight gleaning and data querying. These capabilities paired with Azure’s models and compute infrastructure should lead to some exciting application development going forward.
Snowflake also announced a new partnership with Nvidia. The two will work together to create generative AI use cases and applications powered by the unleashed access to a client’s full dataset. As part of the new relationship, Snowflake customers will gain integrated access to Nvidia’s NeMo (its framework for building generative AI models).
In the world of generative AI there are three key ingredients: AI models to create use cases and workflows, chips to power those flows and data to season AI models to a point where they’re actually valuable. Microsoft (and Nvidia too) bring the model expertise to the table, Nvidia brings its chip talents to the fray and Snow completes the equation with complete access to data paired with powerful tools to put that data to work. This news combines some of the best in breed players from all three ingredients.
Snowpark Container Services:
Definitions:
- Snowpark: Snowpark is a developer tool allowing for source code creation with a diverse array of languages to pick from. Snowpark allows for usage of these languages from within Snowflake’s environment to utilize all of Snowflake’s available data and tools to sharpen the building, monitoring and other processes. Building apps from within Snowflake allows for lower data transference costs and more simplicity. It also gives developers access to Snowflake’s real-time data cleansing tools to fix quality issues immediately. 30% of all SNOW customers were using the newer product as of last quarter.
- Container: In the context of Snowflake, a container is a grouped piece of software. It combines an app or workflow and all of the data and code it uses (AKA its “dependencies") into one package. This allows for disparate source code languages to work cohesively on the platform to power apps. Containers are portable for any platform supporting Docker (a platform allowing for usage and transference of containers) and flexibly scaled up and down.
Snowflake extensively discussed a new tool called Snowpark Container Services. This allows for the operation of software containers natively within Snowpark’s ecosystem for things like data analytics, Gen AI apps and so much more. The big refresh, per leadership, is “bringing apps to the data” and allowing developers to deploy and scale their work from Snowflake’s infrastructure.
This combination is set to accelerate app creation with less time and energy. By conjoining apps and data to one location, things like security posture and compliance are made easier by eliminating data access bottlenecks for programs in operation. Now Snowflake not only tears down data silos with its cohesive, multi-cloud data lake, but it’s eliminating more fragmentation by serving as the single destination for data access and usage as well as app development. This unification is expected to give Snowflake customers a larger cost and complexity edge.
As part of this announcement, the Nvidia partnership will include access to its GPUs and software from Snowpark.