Squarespace (SQSP): Solid Capital Productivity

, ,

I own Squarespace stock SQSP 43.56 +0.03 +0.07% and in many respects, it’s a great case study for lots of the things I have written about calculating and analyzing ROIC. So I’m going to run through it in this post.

What’s the bottom line? SQSP has a pretty attractive ROIC profile – here are the numbers (excluding goodwill and intangibles from invested capital):

SQSP is delivering economic profits, and if its LTV/CAC and subscriber retention hold up, it should continue to. I also like the company’s capital allocation discipline. Let’s get into the basics of the business and then the capital productivity numbers after that.

How Does Squarespace Make Money?

Typically, I would start a discussion of any company with a How Do They Make Money description, because that knowledge provides the underlying context needed to evaluate the capital productivity of a business. If one doesn’t understand what the business is selling and how it’s selling it, the ROIC is just abstract numbers and is of little help in forming any views about the future prospects for the business’ growth and profitability, which is the entire point of the ROIC analysis in the first place.

While putting this together for this blog post, I came across a good SQSP business description on the website of a venture capital firm that did a highly readable and cogent breakdown of SQSP’s S-1 shortly after the IPO in 2021. So for now, I would just point readers there and suggest focusing on the product and pricing discussions in particular. Basically, Squarespace is a SaaS business that sells websites on a subscription basis (say $200 per year) and also makes some money from commerce on those websites. It recently purchased Google’s domain business and has a payments product that just launched and will hopefully soon be a sizable, high-margin revenue contributor. The name of the game for value creation here is to retain subscribers (85% + retention rate), add new ones (via the highest LTV/CAC possible), and increase monetization of existing subscribers via new products over time. Competitors include Wix (WIX), GoDaddy (GDDY), and WordPress (private), which is the product this blog uses. So that’s how they make money; it’s not a particularly complicated business.

Squarespace Stock ROIC Analysis – Before & After Recasted Intangibles

The company has pretty limited capital expenditures and a small asset base, at least as measured via traditional GAAP. Its “investment” is made via the main three expense line items in the income statement: R&D, S&M, and G&A. This is where SQSP spends to acquire new subscribers and service existing ones. Practically speaking, these expenditures can be thought of as growth cap ex aside from the portions spent on maintaining existing customers, which can be thought of as maintenance cap ex (this distinction is accounted for in the recasting process). As such, SQSP is a classic example of a business for which recasting intangibles is necessary to have a view of capital returns that is consistent with economic reality rather than arbitrary GAAP conventions. But before getting into the recasting, let’s take a look at Squarespace’s ROIC without any adjustments, using the operating approach.

SQSP ROIC – Without Adjusting For Intangible Investment

Recall the operating approach to calculating ROIC:

Calculating Invested Capital For ROIC
(Operating Approach) 

Invested Capital = 
  Current assets (taking out excess cash)
- Non-interest bearing current liabilities
--------
= Net Operating Working Capital

+ Net property, plant, and equipment
+ Goodwill & other intangible assets
+ Other operating assets

And here’s the source data we are using – SQSP’s balance sheet for the last few years:

Then, here is the math to get to invested capital (the line at the bottom) for the period 2019 – LTM 9/23:

The first thing that jumps out is the low level of invested capital for a business generating about $1b in sales and $70mm in NOPAT LTM. This is the result of consistently negative net operating working capital, which stood at negative $340mm at the end of 3Q23. The main driver of the low net operating working capital is the large deferred revenue short-term liability, which represents funds received from customers in advance of the service being performed/revenue recognition. Conceptually, I think this is a legitimate source of working capital to include in invested capital: instead of borrowing from a bank, SQSP is borrowing from its customers. This is a great (free) way to finance a business when you can do it.1

The non-operating working capital investments of the company are also low and comprised of just a few simple (simple is good!) items: net fixed assets, an operating lease asset (offset by a similar liability under operating lease accounting cite rule), goodwill, and intangible assets. Most of the net fixed assets are depreciated leasehold improvements, and then there’s some computer equipment and furniture/fixtures. This is standard fare for an internet company. Just servers and what’s needed to house tech professionals – no factories, heavy equipment, tools, or vehicles to worry about maintaining and replacing. They ain’t making semiconductors here. Let’s hit the main non-operating investments one by one:

In summary, over the past few years, the main drivers of SQSP’s invested capital levels have been goodwill going up because of a large acquisition then down because of a writedown, intangible assets increasing substantially because of an asset purchase, and net operating working capital growing more negative. These movements caused invested capital to move around a fair amount, all while revenue (with a CAGR of about 17%) and NOPAT grew pretty steadily3. Here is how the numbers fall out, with 2020 ROIC being N/A because invested capital was still negative at that point (though they still generated $30mm of NOPAT!) before goodwill surged via the Tock acquisition:

When screening for investment ideas, one often comes across ROIC numbers that move around a lot and don’t leave us with any easy conclusions. That’s the case here and it’s one of the reasons why SQSP is a useful one to run through. If asked how I think about the company’s ROIC based on this choppy data, I would say: “It’s high enough to be interesting – keep going.” It’s still early days for this business – it’s far from steady-state where you could say yes this business reliably generates A% returns on capital, like you might say for Costco, for example. But what we do know is it clearly does not employ a lot of capital and seems to be about to grow nicely without it, which is especially interesting given the recurring nature of the revenue, the nice growth runway ahead given the relatively low number of existing subscriber and the many ways to add new ones and monetize them, and the customer captivity. So we would be intrigued enough to learn more, which is what we do next.

But Wait – Should We Be Recasting Intangible Investments For SQSP’s ROIC?

Yes. As previously mentioned, SQSP does almost all of its investing in its business – at least in economic terms – via the income statement. As such, adjustments need to be made for thinking about ROIC. I explained this process in a prior post using a pharmaco example and then applied it to Semrush and Hubspot. I am going to use the exact same approach here without going into as much detail this time around. If you want to walk through the steps, refer to the posts above.

Here are the assumptions4 I used for SQSP’s three main income statement “expenses:”

The rationals here are very similar to those for Semrush and Hubspot: it boils down to a ballpark estimate of what percentage of each is spent on new customer acquisition (capitalized) vs. servicing existing customers (uncapitalized and therefore expensed). I also assumed that the amortization periods were 2 years for R&D and 5 years for G&A and S&M respectively. Again, this is based on a very rough estimate of how long SQSP keeps a subscriber (an 80% retention ratio translates to 5 years) and that R&D expenditures are often speculative and don’t add value for more than a few years at best.

These assumptions drive the following (incremental) capitalized investment and amortization levels that are used in our recasted versions of SQSP’s NOPAT and invested capital:

You can see that we are going to be adding a very large amount of invested capital here, especially as SQSP’s expenditures via these three line items in absolute dollars grow substantially during this timeframe. We are amortizing over just 5 years, so when the GAAP expenses go up a lot Y/Y, the capitalized asset levels grow quickly. This also means higher annual amortization – especially for R&D (over only 2 years) which alone drives about half of the total amortization of these capitalized intangibles in LTM, for example.

All of these numbers are pulled together in the schedule below. The two key line items – Capitalized Intangibles and Amortization of Intangible Investment – tie out directly to the schedules above and are highlighted in red:

NOPAT, even net of the new intangible amortization being taken out, is of course a lot higher because we have pulled out the expenses completely. But look at how much the invested capital increases because of the capitalized intangibles that we added. Invested capital grows by multiples for each year, so while NOPAT went up, invested capital surged. As discussed elsewhere I believe the best measure of ROIC excludes goodwill and intangibles, so the numbers to look at here are in the bottom row in green which are also used in the ROIC chart at the top of this chart. The bottom line is that measured this way, SQSP clearly generates economic profits. It is also generating lower returns over time and generating about a 16% ROIC right now. But it’s still getting back well more than its cost of capital and a lot more than a typical member of the S&P500.

ROIC & LTV/CAC: Joined At The Hip

I think of LTV/CAC as kind of a customer-level version of ROIC that can be a useful check on any capital productivity conclusions I may be reaching, and it makes sense to use that check here. Both measures ask the same question: what is the return the company gets in terms of profit for the money it spends to get that profit? This is particularly true for a company like SQSP because there is not much going on at the company other than trying to acquire new and serve existing subscribers and harvest as much profit as possible from those subscribers via initial subscriptions, cross-sales, and getting more share-of-wallet over time.

Below is a screen grab of an LTV/CAC analysis from the Credit Suisse sell-side initiation on SQSP in 2021. Yes, this is outdated by now, but I couldn’t find any direct references to LTV and CAC by the company in its materials, so this seemed like a pretty good next-best alternative to provide instructive numbers other than my own rough calculations. Note how they flagged leveling and declining LTV/CAC (calculated two different ways). This is pretty typical for this kind of business as it grows, and it’s consistent with the declining ROIC. But they still model it at levels that are clearly accretive to value prospectively. For example for 2022 they show LTV/CAC of 3.5x – 4.3x. In the “based on unique subscriptions” version of the calculation (which uses only S&M expense BTW), the key numbers are LTV/CAC = 1,161/333=3.5x for 2022.

So that’s saying that for every dollar the company invests to acquire a new subscriber, it is getting back about $3-4, fixed costs aside. Below is a screengrab from a very recent Morgan Stanley sell-side piece with LTV/CAC comps for Klaviyo (KVYO), an IPO from a month or so ago. SQSP and KVYO are both SaaS businesses but otherwise not particularly comparable (though KVYO is comparable to HUBS which I have discussed recently). Still, I think this is worth looking at to get some big-picture perspective on LTV/CAC for SQSP:

Note that Morgan Stanley didn’t even include SQSP here for some reason, even though they did include a pretty direct competitor – Wix. Of course, this chart suggests SQSP’s LTV/CAC is low but anyone with experience using LTV/CAC will tell you how much arbitrariness and flat-out BS there is in these numbers especially when they are provided by the companies themselves (have you ever seen a start-up pitch a VC without a monster LTV/CAC in their deck?) Also, note that all of these businesses are in different stages of their life cycles (presumably there is more low-hanging LTV/CAC fruit for apps in their very early years), they have different expense structures with respect to the magnitude of fixed costs that are not picked up in CAC, and the levels of competition or tech risk vary dramatically. Most important, they have vastly different valuations.

From a qualitative analysis standpoint, I would also add that SQSP’s CEO constantly highlights the company’s heavy focus on ROI with its marketing efforts and relentless efforts to iterate and improve ROI with new initiatives and sales funnel tweaks. This is related to the fact that this business was self-funded and profitable very early on in its life, didn’t raise much outside capital, and is still founder-led. SQSP is also far simpler than most of the comps in the chart above, and these factors make me apply a lower discount rate to it. But I did vow not to talk about valuation in these ROIC posts…The bottom line here is that SQSP has a pretty good LTV/CAC in absolute terms which is roughly consistent with its attractive ROIC. From an investment standpoint, there are still many more questions to answer on SQSP, including developing a thesis on the trajectory of the company’s ROIC. But clearly, the company’s basic capital productivity profile qualifies it for more consideration.

  1. Note that I included restricted cash and investment in securities in short-term assets. The latter is not an operating asset and should probably be excluded, but I kept it as an extra measure of conservatism. ↩︎
  2. Here’s what SQSP said about the goodwill impairment related to Tock in its 2022 10K: The impairment charge during the year ended December 31, 2022 represents a non-cash goodwill impairment charge of $225.2 million primarily due to market values deteriorating subsequent to our Tock acquisition in March 2021. ↩︎
  3. See attached excel spreadsheet for the income statement and the NOPAT calculations. ↩︎
  4. Note that these expenses include stock-based comp which is non-cash. I treat SBC as a cash expense so I have pulled the SBC amounts out. ↩︎

Please email me with questions/comments/errors related to this post!