Predictable Revenue

A fundamental underpinning of SaaS is the concept of Predictable Revenue. I’ve always been confused about this thinking, “Why is SaaS based revenue any more predictable than any other revenue model?”

As I pose that question, I get a lot of pushback, “Dave, you don’t understand. SaaS subscriptions are very predictable, we have an ongoing monthly stream of revenue coming from the subscriptions….”

When I reply, “All you are doing is talking about cash flow, not predictability,” people’s eyes roll, they know I’ve gone off the deep end.

I then try another approach, “If predictable revenue is a key underpinning of the SaaS business model, why are so many SaaS companies failing in meeting their goals?”

I’ve been hammering very hard on this concept for a number of weeks. It’s not that I have anything against SaaS, I’ve co founded and sold a couple of SaaS based companies, I sit on the boards of a number of others. But we have this concept of what SaaS is that limits our thinking about how we can build and scale companies. As a result, we see too many software companies failing.

And I get frustrated. I see so many software and related companies that are limiting their ability to grow, because they follow the SaaS mantra. If they freed themselves from what the SaaS model represents and start thinking about things in a different way, they can achieve unimaginable success.

If we look at the “founders” of the SaaS concept, it was innovators adapting and assembling bits and pieces of other business models to create a new approach. But that model is no longer serving customers or SaaS companies well, so we need to innovate.

So let me nerd out a little about some of the SaaS underpinnings and how we might think differently about them. The most misunderstood is the concept of “Predictable Revenue.”

I’ll start with this, all revenue is perfectly predictable! Well, at least 99.9% predictable. Whether it’s an outright purchase, a subscription or anything else it’s perfectly predictable. That’s because Revenue is represented by payments we have received plus invoices that have been issued. And this is a precise number. Look at the P&L of any company, including SaaS companies and the top line you will see is “Gross Revenue.” The CFO and accounting staffs know the only thing they can report in revenue is payments received plus invoices for payment that have been issued. When we understand this, we understand that 99.9% of revenue is predictable because we already have it in hand or invoiced. (There are some invoices that a small number of customers don’t pay, but we will set that to the side.)

Now you will say, but Dave, look at the future revenue we can count on based on our ARR. That’s fair. SaaS has invented this metric that represents a contracted future revenues. ARR represents the rolling 12 month representation, MRR represents the rolling monthly representation. It cannot be recorded as revenue, because it hasn’t yet been paid or invoiced. And ARR can go up or down. But companies with other business models have similar metrics with similar future predictability, but it’s labelled as something else. For example, it’s often called backlog. This represents contracted orders for future delivery. Perhaps the product hasn’t been built or shipped, perhaps it was contracted to be delivered at a certain time. Perhaps the services have not been delivered. But backlog represents future revenues that are perfectly predictable. Or embedded product organizations have supply contracts, representing the amount and schedule for delivery, invoicing and payment.

The difference between the monthly subscription models and any other model has nothing to do with predictability. It only involves cash flow.

As a subtly, you will never see ARR or Backlog or Contracted Revenues reported on a P&L, because they don’t represent revenue. In investor reports, quarterly and annual reports, there is always a discussion of future revenues based on ARR, Backlogs, etc. But accounting standards do not allow these to be recorded in the P&L.

So the concept of the subscription model being more predictable than any other revenue model is a false narrative.

When we recognize this, it’s tremendously freeing for SaaS and aaS companies. One of the biggest challenges, particularly now when investors are unwilling to support growth regardless of cost model, is cashflow. Too many companies simply can’t afford to put in place the resources they need to deliver on future contract commitments, because they can’t pay for them. If we start thinking, “How do we accelerate cash flow,” recognizing that we can collect money in a way that’s different we can solve many of our problems.

If you look at the original subscription businesses, media–newspapers and magazines, while they call the subscriptions, while they might quote a monthly payment, they will only accept an upfront payment for the duration of the subscription. That’s a purchase model, not a subscription model.

Let me look at another aspect of the “Predictable Revenue” model. This will be much shorter, because it’s much more obvious. Much of this is based on our interpretation, or perhaps misunderstanding, of Aaron Ross and Marylou Tyler’s Predictable Revenue book. The underpinnings of that book are concepts behind lean manufacturing and the Toyota Production System (TPS) originated by Ohno, Toyoda, Deming.

The underlying principles of all of these is an assembly line mentality in workflow design. What drives productivity and efficiency is a continuous flow process, structured Takt times which set the pace of the process, and standardized work. In selling we looked at standardized work by starting to segment parts of the selling process. We put in place SDRs, BDRs, AE, and so forth. We outlined specific tasks/metrics for marketing, pre-sales, sales development, etc. Each “station” in the sales manufacturing line has very specific functions and responsibilities. In designing this approach, the goal was to standardize the work at each workstation, creating consistent repeatable processes. When the work at one workstation is completed, it is moved onto the next, then the next, until we get a PO.

While this design approach has an appealing logic, there is at least one thing that does not make this manufacturing model a model that can be easily transferred to selling. For this model to work, variation at any point in the process has to be eliminated. The thing that blew up manufacturing processes before TPS was the variation that occurred at any step in the process. The greater the variation, the more likely the assembly line would fail to meet it’s objectives. Each workstation was expected to get a certain set of inputs and those always, without variation, produced outputs from that workstation. The work done within the workstation did not vary. If bad or flawed materials were introduced into the process, the line collapsed. If one workstation failed to produce the required output, the workstation in the next step would fail, and that failure would ripple through the process.

In manufacturing, the way this was eliminated was to eliminate all variation. The work was done in a very specific way, with no variation. The materials and products on which the work was done fit very specific parameters with no variation.

And that’s the problem with applying the same principles in the selling assembly line. While we might be able to precisely script the actions of sellers, minimizing variation, each customer is different. Each customer has differing needs, understanding. And customers aren’t constant, they change. So while our qualifying step may pass along a perfectly qualified customer, something changes in the next step. What we expected of the customer in that step has changed. But the way our work was designed, we can’t adapt to that change.

Variation in manufacturing processes can be reduced to fractions of a percent, which makes these lean principles very powerful. But in selling the variation is very high, as a result the predictability of the outcome plummets.

In the early days of SaaS, recognizing this as a problem, sellers started targeting individuals, rather than organizations. While variation couldn’t be eliminated, it could be reduced substantially. With careful qualification up front, we could get prospects that behaved similarly, and run them through the sales assembly line. And we could get reasonable results. But as we started running small teams or larger groups of customers through this process, it was like herding cats. They did what they wanted to do, and kept changing.

This exploding variation, at every step in the process creates huge challenge. We see this in plummeting win rates. Only 15-20% of the inputs end up coming out of the assembly line with POs. And it’s taking longer and longer….

So what we thought might be so predictable in our assembly line selling process, turns out to be very unpredictable. Results plummet, though while trying to recover, cost of selling skyrockets.

But, again, we can learn from manufacturing. Early in the days of lean manufacturing and the Toyota Production System, manufacturers found certain kinds of manufacturing processes didn’t fit the principles well. Variability was very difficult to control. In those situations, manufacturers started creating “clustered manufacturing.” This would be a station, perhaps with several people working together, that could adapt what they were doing for the specific situation. Rather than do the exact same steps every time, they would adjust the work done, to accommodate the variability. Teams, working together, would figure things out, they would adapt and change what they did to produce the desired outcomes.

In selling, we are beginning to understand this. With the huge variability in each buyer and their situations, rather than rigidly applied steps, we are beginning to be more flexible. Many are seeing the SDR/BDR handoff to AEs is no longer effective. They are looking at full cycle reps. They implement basic processes, but train the sales people to adapt to the situation. Even in this, there are ways to reduce the variability, to improve productivity/results. Focusing viciously on the ICP is one. Having a strong process, but not rigid steps within that process is another. Focusing on the customer and what they are trying to achieve, rather than ourselves, is another.

I’ll stop this long explanation here. The reason I have gone through this is that while everyone wants “Predictable Revenue,” our understanding and models of Predictable Revenue have misled us, or have been very wrong. And sticking rigidly to these models limits our ability to adapt, respond, and grow.

Let me summarize:

• All revenue is predictable. But there are different cashflow models in each business model. Sticking rigidly to one cashflow model may not serve us as well as it should.
• Predictability in the selling process is a function of minimizing and eliminating variability with the inputs (customers). There are some areas where we can minimize this variability, but in reality they are small, and as choice and buying complexity increases, variability increases exponentially. As our manufacturing predecessors learned, the way to deal with higher variability is to provide great adaptability.

Within SaaS our misunderstanding of these terms and processes, and our rigid adherence to them have limited our abilities to innovate, grow and achieve. The market for software and related tools will continue to skyrocket, but the way these are being bought no longer fit a rigid model for how we engage our customers.

Outside of SaaS, we always have to look at where our models no longer serve us, adapt and change them.