Not very long ago, when people talked about software companies adopting or transitioning to the subscription business model, the discussion often focused on its relative merits. Was it better business to follow the traditional model, where vendors sold a perpetual license to a “box” of software — which the customer would then run on its own on-premises server — for a single up-front price? Or was it better to sell a subscription, where the customer periodically paid smaller amounts over time, under contracts granting them the right to use software that was typically cloud-based? The debate is over now; the subscription model won.
Few, if any, new software companies — and there is a virtually incalculable number of startups every year — are coming to market with the packaged software model and installing functionality on customers’ servers. At the same time, today’s software industry is the scene of an exodus, with hundreds of vendors that have used the old model for years or decades transitioning to the new one.
But the movement is even wider than that. Businesses of all kinds are starting to sell products via subscriptions, everything from men’s shavers to kid’s toys to gourmet foods. Cisco Systems is looking to sell a cloud-based management service with its switches. For the 2017 Major League Baseball season, the St. Louis Cardinals are selling a ticket subscription service that lets fans attend as many Cardinals home games as they want for a monthly fee of $29.99.
The subscription-based business model isn’t new, of course. It’s old, in fact, having been a primary model for newspaper and magazine publishers, for example, far into the past. But it’s the software businesses that are leading the way in this new revolution to capture recurring, predictable revenue from customers.
Software-as-a-service (SaaS) providers have been around since the 1990s. But the massive, relatively recent shift in the software industry has been driven by enabling technologies. Most prominent among them is the emergence of robust and inexpensive platforms — Amazon Web Services (AWS), the market leader, followed by Microsoft Azure and Google Cloud — for developing web applications and creating data centers that can scalably deploy those applications to customers. (See “Moving Up to the Public Cloud.”)
“These capabilities continue to improve exponentially,” says Tyler Sloat, CFO of Zuora, which offers a comprehensive billing and finance platform for subscription businesses. “When we moved our data centers five years ago, those services weren’t an option for us because they were in their infancy.” The move required a six-month planning process and buying a lot of equipment. “Today, we can spin up a new data center [within AWS] overnight,” says Sloat.
The subscription model’s success is showing up in revenue growth across industries. Zuora has created a “Subscription Economy Index,” populated with sales data from 350 companies that, as of November 2016, had used Zuora’s software for at least two years. According to the index, sales for subscription companies grew 900% faster over a 5-year period than did sales for S&P 500 companies. Subscription-based sales also grew 420% faster than U.S. retail sales and 500% faster than the U.S. economy.
Those are eye-popping statistics, to be sure. Still, finance chiefs at subscription-based businesses face challenges, as the software companies that have led the way know all too well. Among the nagging issues are teaching investors and resellers about cash flow and revenue timing; managing the sales mix as a company moves from the one-time sale to the subscription-selling model; and dealing with complex new accounting rules dictating how to recognize sales revenue.
Educating key audiences, like sell-side analysts and institutional investors, about the economics of the subscription business ranks high among the difficulties finance chiefs encounter. “We’ve been publicly held for more than five years, and I still have to spend a lot of time helping investors and analysts understand the difference between a subscription model and a perpetual model,” says Paul Auvil, CFO at Proofpoint, which offers cloud-based email security and compliance services.
The company’s current guidance for 2017 projects about $100 million of free cash flow, Auvil notes. “But, if you compare that $100 million to $100 million of free cash flow generated by a company selling [boxed software] for a living, there’s a vast difference in value,” he says. “Because unless I have a problem with customer churn, I’m going to get that $100 million again next year. And I might get new customers. Whereas if you’re a box company, your boxes may go out of vogue next year and your free cash flow may drop to $50 million.” Auvil adds that, when he does non-deal road shows, he typically meets with about 20 to 25 investors each day, and “half of them still don’t understand this.”
Another finance chief concerned about analysts and investors is Thomas Tuchscherer of Talend, which offers an open-source data-integration platform. The company went public in July 2016, and the CFO has been dismayed over analysts’ use of a metric called “calculated billings.”
The widely used metric is generally defined as revenue for a particular period, plus the sequential change in total deferred revenue as presented on the balance sheet. It’s seen as a way to back into an estimate of the value of new bookings made during the period, or “new annual contract value (ACV) bookings.” Tuchscherer calls ACV “the leading indicator of future revenue performance.” Hence it’s the metric analysts most want — but it’s a non-GAAP metric that few subscription companies divulge.
Why don’t they? “If you disclose it one quarter, there will be an expectation that you’re going to keep disclosing it, and it becomes just one more thing you’re going to be held accountable for,” says Zuora’s Sloat. “It’s OK if you have a great quarter, but then later when you don’t have a great quarter, you’d rather not disclose it. In the subscription business there’s a lag between bookings and revenue, so if you miss one quarter, you might still make it up in the next one.”
Sloat says calculated billings, which many see as a proxy for new ACV bookings, is a misleading metric. That puts him squarely in Tuchscherer’s camp. “Many analysts and investors agree that it’s not the best metric, but it’s a habit and it’s hard to move them away from it,” he says.
Tuchscherer recently wrote a blog post decrying the use of calculated billings. The metric, he wrote, does not adequately take into account discounts that customers may be granted in exchange for signing a multiple-year contract.
Also, he says, if a vendor intends to do repeat up-sells or cross-sells to existing customers, it’s good practice to align the end-dates of new subscription agreements to those of existing agreements. However, that may result in “stub-
period agreements” that are shorter than one year, which could impact billings and short-term deferred revenue.
Tuchscherer rattled off several other deficiencies of the metric before concluding that “calculated billings could only be a good indicator of future performance if all other factors remain constant, which is rarely the case in the technology industry.”
Other software companies may have additional constituencies to educate. For example, in the cybersecurity field, where Proofpoint plays, value-added resellers are most familiar with perpetual licenses. “When they close a deal, [traditionally] they get a 20% to 30% cut” for a one-time purchase, says Auvil. But when Proofpoint lands a subscription deal, it’s for a year. “We don’t have a big tidal wave of cash coming in, and I can’t give them 30 points of margin,” says Auvil. “Maybe I can give them 10 or 15 points.”
In other niches, big subscription companies such as Salesforce.com and Workday don’t have this problem because they sell directly to customers.
Auvil says Proofpoint has done a lot of work to help resellers understand that subscription revenue “is a gift that keeps on giving. You’re not going to make a lot of money up front, but you can get that 10 to 15 points of margin year after year.” The effort has been ongoing for 4 or 5 years, he notes, and “some of the bigger channel partners are realizing that the age of big-ticket perpetual licenses and boxes is slowly ending.”
Perhaps an even bigger challenge for Proofpoint, Auvil notes, is the ceaseless need to re-earn customers’ business, year after year. The subscription model, he says, is a double-edged sword. “Every year, a customer has to make a decision about who to buy,” Auvil says. “And in our world, cybersecurity, it’s pretty easy to move from Proofpoint to someone else.” Such a switch might only take three or four weeks, he says.
That benefits customers, because the only way for a vendor to keep them is by delivering both world-class product efficacy and excellent customer service. “That’s true in many cloud-based services businesses,” says Auvil.
Companies that transition from the perpetual license to the subscription model face a special challenge: convincing investors that it’s a good idea. The main emphasis when going through the transition is to “come out on the other side as quickly as possible,” says Andrew Miller, CFO of PTC, a maker of software for product manufacturers. That’s because, for almost any company making the move, financial results will inevitably plunge in the short term, as smaller sales replace big-ticket ones. It can also put a dent in market capitalization, as investors struggle to understand the company’s changeover.
When Miller arrived at PTC in 2015, the company was already planning its transition. He brought to the table the experience of having conducted two business-model changes in the past.
With the help of management consulting firm McKinsey, PTC conducted market and price-elasticity studies. “The main thing we learned was that 75% to 80% of our customers definitely preferred subscriptions,” Miller says. “We also learned the [optimal] ratio of pricing between the two models and what features of the subscription model mattered to customers.”
The transition also required a realignment of sales compensation plans to favor subscription sales, including different incentives for the 25% of PTC’s business that went through channels.
As it turned out, the pace of the transition exceeded expectations. The first-year goal for the new sales regime, launched at the beginning of the company’s 2016 fiscal year on October 1, 2015, was for 25% of new bookings to be subscriptions. But subscriptions ended the year at 56% of new bookings, which was actually ahead of the company’s second-year target of 45%. In the fourth fiscal period, the figure hit 70%.
PTC now expects to reach a final goal—85% of new bookings sold as subscriptions—in fiscal 2018, which starts this coming October. That level is based on the company’s analysis of the overall market; it does business in about 90 countries, including a number of markets in Asia, “where there are different cultural buying behaviors and business rules in terms of capex versus opex budgets,” says Miller.
The company also is offering a conversion program enabling customers that previously bought a perpetual license to switch to the subscription model. Such customers have paid an average of 25% more for subscription services than they had been paying for maintenance under their perpetual licenses, according to Miller.
For some subscription-based companies, another obstacle will be accounting. Accounting Standards Codification 606, the new revenue recognition standard slated to take effect December 15, 2017, presents the potential for a pounding headache.
The new rules won’t much affect pure-play SaaS subscription companies, other than probably requiring them to amortize sales commissions over longer periods. Things are likely to prove dicier, though, for companies like PTC and Talend that have hybrid models (i.e., they offer both subscriptions and perpetual licenses).
Under the new standard, companies will be required, for accounting purposes, to allocate the overall transaction price of each perpetual-license agreement among the vendor’s individual contractual performance obligations. In addition to the software, those obligations could include maintenance, product upgrades, call-center support, and implementation services. The proportion of the transaction price allocated to each obligation is to be based on the price at which the company would sell that good or service on its own.
If a standalone price is not readily available, the company can choose from several estimation methods. But any way you cut it, performing the transaction-price allocation will be filled with guesswork for some companies.
Under existing revenue recognition rules, software companies are already required to estimate the fair value of each separate deliverable, which many satisfy using a method known as “vendor-specific objective evidence.” That has relatively little impact on companies that sell everything as part of a subscription (except, in many cases, their implementation and training services).
But the new standard may indeed affect the financial reporting of companies like PTC and Talend. The CFOs of both firms say they are still assessing what that impact will be. But to be sure, there’s no fun to be had here. “It’s painful,” says Tuchsherer.
That’s particularly so given that he doesn’t understand the rationale for the rule, as it applies to Talend. “We sell a subscription to customers, and we don’t sell the software license separately from the support and maintenance,” he says. “It’s one package that includes everything. That’s how customers perceive the value of what they’re buying, and that’s how they actually buy it.
“So,” he adds, “having to carve out different elements of our subscription revenue, which frankly will be a very subjective process, is not very rational or logical.”
Asked whether it might force Talend to rethink how it packages its products and services, Tuchsherer pauses for a few beats. “Well, we don’t want to,” he says. “It would be very bizarre if a new accounting standard dictated or influenced the way we sell to customers. Yet, it may have an impact. I don’t think so right now, but it could.”
David McCann is a deputy editor of CFO.
Most software buyers like the subscription model because it provides so many advantages. Generally, a buyer:
Additionally, concerns that data is less safe in a cloud than in servers controlled by the buying company have faded over the past several years.
One company that doesn’t need to be convinced is PTC, a maker of software for product manufacturers. It has traditionally sold software via the perpetual-license model but is currently in the midst of a transition.
“One thing we learned in our market testing,” says CFO Andrew Miller, “was that, in every segment and geography we’re in (with the exception of small segments in a few small countries), regardless of customer or deal size, subscription was far preferred over perpetual, as long as you get the pricing right.”
From a purely financial standpoint, the vendor benefits from a recurring revenue stream from each customer, provided the quality of its products and services is good enough that customer churn remains low.
The vendor then doesn’t have to “kill what it eats,” in the words of Steve Love, CFO of Dialpad, a provider of cloud-based telecommunications services — that is, it doesn’t have to reach periodic revenue targets solely by attracting new customers. For successful subscription companies, revenue therefore becomes much easier to predict.
The only significant downside is for startup subscription companies, which at the outset will likely be taking in only small chunks of revenue from a modest customer base. That means they need a lot of up-front investment in the business before the cash really starts to roll in. —D.M.
Featured image: Thinkstock