In 1998, Yom Kippur fell on September 30th. For most of the Jewish community, the date of the most important holiday of the year was no different than in other years. For me and my Jewish CEO boss, though, as officers of a public software company, September 30 was a tough day to be out of the office, sitting in synagogue atoning for a year full of sins. It was the last day of the third quarter of the year and we had more deals we needed to close to finish the quarter strong and report numbers to Wall Street that justified our high-flying profile as a recently public Internet commerce software company. By sundown September 29th, when we left the office for the onset of the holiday's traditions and presumably focused on higher order matters, we had not yet made the quarter. Going offline without knowing our fate resulted in one of the most miserable 24 hours in synagogue I can remember (and I am somone who usually enjoys being in synagogue!).
When my CEO and I got back online after sundown September 30th, it became evident that the final handful of deals that we needed to close to make the quarter had slipped out. A few weeks later, we "pre-announced" that we were going to miss the quarter – one of the worst speeches I ever remember being a part of. Our stock naturally plummeted.
We were victims of a lot of problems, many of our own doing, and I can hardly blame Yom Kippur and the holiday's inopportune timing on our missing the quarter. But many years later, I began to appreciate that one of our core flaws was our business model.
We priced our enterprise software in the form of a perpetual license. As a result, the full revenue for each deal was recognized in that quarter as soon as the software was shipped. This allowed our revenue to skyrocket from $1.8 million to $22.5 million in one year, the year we went public at a billion dollar valuation (ok, it was 1996; everyone went public in 1996 with a billion dollar valuation), and then $61 million the following year. But the downside to our business model was that we did not have hardly any recurring revenue.
I later came to realize that recurring revenue is magic.
Since my harrowing experience, I have become a zealot about recurring revenue. When I discuss business models with entrepreneurs and investors, there is a varying appreciation for why recurring revenue is so special. Recurring revenue business models are not a little bit better than non-recurring models. They are 10x better. At Flybridge, we have added "business model", with a particularly weighting towards recurring models with high gross margins, as one of the important evaluation criteria when we make investment decisions alongside market and team, which are the two canonical criteria for all venture capital firms.
Before explaining why they are so magical, let me define a few types of recurring revenue models. Many jump to the assumption that SaaS (software as a service) is the only recurring revenue model, but there are actually a few you can choose from when designing your business model:
- Consumable - the classic recurring revenue business was invented by Gillette: get cheap razors in the hands of shaving consumers and then perpetually sell them expensive razor blades. Keurig has a similar beautiful model with its coffee machines – keep selling those consumable coffee containers and your business never loses its value. 3D printers, with their consumable resins, have a similar business model.
- Subscription – this is when you have a subscription contract for a period of time, typically annualy, and charge yoru customers for the service or content pro ratably over the course of the period. Magazine subscriptions and software subscriptions (often often called SaaS) fall into this category. SalesForce.com basically invented this model for software companies. Your cell phone provider, Netflix and Hulu are other examples of successful subscription revenue model businesses.
- Transaction – this is where you charge for transactions that occur over and over again. The credit card companies and other high-frequency payments-based businesses, such as PayPal or Stripe, are examples of this kind of recurring model. Uber is another nice example of this since securing transportation tends to be a recurring transaction for many professionals.
- Rental – finally, when you borrow an asset, such as an apartment or a car, you are signing up for a recurring charge so long as you continue to borrow that asset. This creates a recurring model as well. Data storage companies have this model as do many cloud services, such as Amazon's AWS. Amazon is renting you their assets – powerful computers and endless data storage. Amazon also has software and analytics that you are subscribing and so have a doubly powerful recurring model.
Here's why recurring revenue is so magical:
- Predictability. When you have a recurring revenue business model, you rarely miss your monthly or quarterly numbers by more than 10-20%. Your forecasting process is much more accurate. At the beginning of the quarter, you start with a base to grow from rather than begin at zero. In a SaaS or subscription software business, you can predict your churn rate and new business closings to determine your growth rate. The management team and the investors are thus rarely surprised by major fluctuations in your results. As discussed below, this predictability has many downstream benefits.
- Visibility. Because of the nature of recurring revenue models, you have clear visibility into what is coming in the next few quarters. You know where you stand well in advance. In a recurring revenue model, if you take the last day of the quarter off, you will not tank the company because you have so much visibility into your business, you are rarely surprised about what happens on that last day. For example, by mid-year, two of our portfolio companies with transaction-based recurring revenue models, Bluetarp and Cartera, have already confidently predicted they are going to meet or exceed their plan for the year and are working on what they can do to impact 2015. If they are off, they will know it well in advance of any of our portfolio companies that are non-recurring in nature.
- Expense management. Predictability and visibility means you can manage your expenses more precisely relative to your revenue. One of the hard things about lumpy revenue models is that until literally midnight on the last day of the quarter, you don't know how you did. Which means it is hard to ramp up or down expenses smoothly to match revenues. Ramping expenses up and down is a sticky process because it usually involves people and there are many friction points, delays and costs as well as externalities (such as morale) when you try to rapidly ramp down expenses in a quarter as a result of lower-than-anticipated revenue.
- Valuation. Because of the predictability and visibility factors, valuation multiples are radically different for recurring revenue businesses than any other revenue model. Terry Kawaja did a wonderful analysis of advertising technology company valuations and the positive impact on multiples that exist for SaaS and programmatic companies (such as our portfolio companies tracx and DataXu, respectively) as compared to non-recurring advertising technology companies. When we analyze the public company comparables for our portfolio company, MongoDB, we are always amazed at how much higher those comparable companies (enterprise SaaS leaders, like Palo Alto Networks, Splunk and Workday) are trading as a multiple of revenue (often 8-12x) as compared to other public companies that are not blessed with such a magical business model. A recent investment banking analyst report I read showed that companies with SaaS software models averaged a 6x revenue multiple, twice as high as the 3x revenue multiple that perpetual software companies average.
To be clear, recurring revenue models are not perfect. It is harder to ramp to 10x year over year growth. You do get plenty of lumpiness in bookings of new business, which translates into higher or slower growth rates over time, depending on performance.
But despite these downsides, it is clear to me why there is such magic in recurring revenue models. It looks like Yom Kippur once again falls on the last day of the quarter in 2017. With the majority of my portfolio companies having recurring revenue business models, I am not going to sweat it.
Great article. Have you checked out John Warrillow’s recent research in Automatic Customer? I’m talking about the fact that recurring revenue increases valuations and multiples. As an entrepreneur I am astounded more entrepreneurs don’t get this. Since one’s exit can generate up to 50% MORE wealth if planned properly (if you take the folks at Exits.com by their word) ,and since, ceteris paribus, $1 of recurring revenue is worth more than $1 of unit sales (given the former model has higher multiples)..shouldn’t more people be talking about this? When I really “got” that it was like a lightening bolt. All revenue is NOT created equal.
Royalties, baby. Not new products. Lol. Yeeeaaah.
You have share the great article thanks for you
It’s funny – I was just talking to somebody about this the other day – the fact that back in the late 90’s / early 00’s all customers knew the “end of quarter” dance – that they could get their best discounts by waiting until the end of the quarter. So our monthly revenue each quarter was like 10% in month 1, 15% in month 2 and 75% in month 3. EVERY QUARTER. Made for highly inconsistent quarters and lots of stress in month 3. No matter what we did to try to reverse this, it always happened.
And I totally remember 9/30/98. Although I didn’t enjoy synagogue as much as you apparently do, so it was a double whammy for me! 🙂
Great article. Jeff, I also just realized the power of recurring revenue. In gaming, recurring revenue features are also very lucrative. Subscribe to x membership to receive x amount of currencies at x discount monthly.
Here are some learnings after recent A/B testing with subscription model for my project.
1. Some customers are sometimes put-off by monthly payments. But introducing a free trial (i.e. 30 days free)) helps alleviate their concerns.
2. It’s easier to build recurring revenue models in the States than other markets. Some markets just simply don’t support automatic charges.
3. Evan, we also tried a x price every three months. But at a higher price range. But I found that more customers cancel within 3 months. So I dropped the price, changed payment to monthly, offered a first month free trial. LTV rose and cancellations dropped. Resulted in higher monthly revenue.
Thanks Jeff, I’d agree after founding one company without recurring revenue and 2 with. What are your thoughts as to how the duration of the revenue plays into whether it is considered to be recurring or not? For instance, my software developer ‘school’ has tuition that flows in from customers for 3 months per customer.
Hi Evan – I think of recurring revenue as having timeless duration. The customer may cancel at any time, but if they don’t, it just keeps going and going. Higher education, for example, has a four year term but is not recurring. After those four years, you are done and gone. Same with your developer school. Alums who donate, though, are recurring. They keep donating year after year after year. My 81 year old Mother has been giving money to Wellesley College, her alma mater, for 60 years! That’s an impressive duration-based business model.