Depending on whom you ask, companies that sell their products on a subscription basis are either companies that are veritable giants of growth, or firms that hide their business models’ inherent weakness in the form of short-term losses that are dismissed under the guise of investment.
Regardless of where you stand on this issue, companies based on the subscription model aren’t going away, if for no other reason than customers can’t wait three years between upgrades anymore. They need to move much more quickly than that and the subscription model accommodates that need for agility. Customers also don’t want to deal with the complexities of installing, managing and maintaining software anymore and the subscription model removes all of that.
In this post, we hash out the pros and cons of the subscription approach and argue whether its viability is just a matter of accounting semantics or a fundamental problem with the way these companies operate and the high cost of doing business in this fashion.
Wall Street has tended to look at companies with subscriptions models showing red ink on the balance sheet with a fair bit of disdain this year. But perhaps traditional accounting methods aren’t the best measure of a subscription company. As Tien Tzuo, founder and CEO of Zuora told me, our accounting systems aren’t set up to accommodate a subscription model. “Our accounting systems, don’t know whether a dollar is going to recur,” he explained.
So if you’re Dollar Shave Club and you have 10 customers paying $5 a month for a year, you don’t have the revenue on the books yet for the life of the customer, even though you know you’re going to be earning that money (10 x 5 x 12) over the length of the subscription. The traditional accounting system can’t take this into, ahem, account.
When you start looking at enterprise companies, the numbers are much bigger, but the same problems apply.
The idea that we haven’t known in the past when a dollar of revenue will recur is only partially true. If a company signs a large support contract that will be serviced on a recurring basis, and tells its investors, they model under that presumption. Certainly, there can be minor GAAP-to-non-GAAP dissonance at play, but even the most rudimentary acolyte of corporate finance can parse the difference.
The strong version of this, of course, is when your entire revenue stack recurs. This is when you may begin to discount short-term losses more heavily in favor of longer-term recurring revenue. But the concept is merely the amplification of a problem that we had previously.
Companies have long had recurring and non-recurring revenue streams. AOL has lots of the former in the form of dial-up customers. (AOL owns TechCrunch.) It isn’t that hard to model. And yes, SaaS revenue has a few other quirks to it, but it’s still recurring revenue on a contract, making its DNA quite similar.
I’ll allow the following: When you are a high-growth organization, the effects of recurring revenue can become overshadowed, in the short-term, in the following circumstances: When you have long-term clients that are hard to land, have high, positive dollar-churn over the life of the account, and very high gross margins. This is when you have the highest up-front costs, compared to lifetime-value of a customer, which most distorts growing recurring revenue that must be accounted for in its own fashion.
The biggest issue with the subscription model is that you have to spend a certain amount of money to acquire each customer and this cost rises substantially in the enterprise. At some point, usually I would suspect after a year or two depending on the customer size, you start making money and then it’s just pure income.
The problem for SaaS companies running on a subscription model is that there is a big upfront cost for acquiring customers that comes in the form of building a big-boy sales and marketing team. At some level, it’s not all that different from a traditional sales model and to get the best people you have to pay to lure them to the new kid on the block.
When you look at these costs on paper using traditional accounting methods, it doesn’t look very good for the company that is filing the S-1. There’s lots of red ink and losses, but there’s also money that’s coming, gobs and gobs of money if you ran your subscription company right. Perhaps if Wall Street began asking a different set of questions, it might begin looking at these companies more favorably. The money’s there, it’s just not accounted for because it won’t show up in the standard accounting methods for some time.
Maybe we need some new language and a new way of looking at this and if we were able to offer a line item for known recurring revenue in the pipeline, or if Wall Street financiers simply asked these questions during the company road show, it might flip how Wall Street sees these companies and what constitutes success.
What I want to know is why SaaS revenue can be so expensive. In the above, my colleague Ron Miller makes what I will call the common argument about SaaS revenue: Losses up front, party in the back.
Some large SaaS companies have spent more than three times their top line over the course of their life to build nine-figure annual recurring revenue. You can imply the cost of that. Not all companies have to burn so heavily to generate that amount of top-line, recurring or otherwise.
The corollary to that point is that the idea that SaaS incomes are future profit cannons is slightly doubtful — if that past revenue is now such pure profit, why doesn’t it support, and constrain future customer acquisition expenses, which contribute to gaping GAAP losses down the road, say, eight years, keeping revenue roughly equal to just the sales and marketing line item?
Even when the lines start to move toward crossing, the idea that SaaS revenue becomes a quick profit goldmine appears to be belied by the results of companies that we have the figures on; why is Salesforce still unprofitable, using normal accounting tools? Does its growth rate make its continued, and expected-to-continue, losses reasonable?
There are so many companies out there working with this revenue model, and the subscription model has so much potential because it isn’t a matter of getting you to buy something again and again. You have a service and so long as you are reasonably responsive to your customer’s needs, chances are you’re going to keep them.
Contrast this with the old model of selling boxed software (as one subscription model example) where the vendor spent a couple of years updating the software and the customer had to spend tons of money to upgrade. With the subscription model, the product is constantly being updated so customers are getting the benefits right away without waiting.
Since this model is more attractive to customers, it bodes well for the subscription companies who can count on continued revenue from happy users. You don’t have to sell them a new car. You sell them the service and your sales cycle is almost over. Of course, you want to add more licenses and schmooze for renewals, but the hard part is done once you’re in the door.
That should drive down the sales and marketing costs and over time, the cost to attain customers should level out. It may take a small leap of faith, but we are seeing it play out already as Zendesk bucked the trend and went public earlier this year and it’s doing just fine. Perhaps Wall Street needs just a few more examples like this before it finally comes around that subscription revenue is far more lucrative than selling one widget at a time.
This fits neatly under the idea that a SaaS account, once sold, becomes a quick profit source. The presumption I take is that once it has paid off its customer acquisition cost, which should take no more than one-third the life of the customer, gross margins give you wings and up go your margins. Again, I think that if this were the case, we would see different financial performance from firms that we have the figures on.
Let’s use an older firm to make our point: Microsoft now wants to sell Office on a subscription basis. The dollar amount is probably — measuring on an amortized basis for the Office In Box version — larger than what it previously took in. But the company now has greater server, bandwidth, and other costs. So does the margin situation really change?
And then there is the idea of continuous development; is it cheaper to build a product once in three years, ship it, and then restart? Or is it cheaper to ship every week? It probably varies for every company, but when every subscriber is now a continued research and development cost, taking part in other line-items as well, the idea of gross margin can, perhaps, become gross.
To summarize, SaaS companies can be great revenue engines, but we should be careful to dismiss all losses as reasonable investment. That, and there is nuance to long-term support and sales costs to selling software in this fashion that, for companies coming into the space from the Old World, might prove more costly than promised.
We look forward to SaaS kids shouting “3X LTV to CAC!” In the comments, as if we have never heard of the concept before.