Over two years ago, we wrote a two part series about SaaS Metrics that has continued to spark interest in the subject. Since that article was written, Scio has consulted with well over 50 SaaS and Cloud companies through our SaaS Strategy Sessions, Workshops and most recently a program headed by FUMEC in Mexico.
This background has given us a broad view of the issues entrepreneurs should consider to plan and operate successful SaaS businesses. There are four common categories that these issues fall into:
- Business Model Scalability - Finding a scaleable, repeatable business model is key and very often lowest on the list of priorities.
- Operational Requirements – Most companies are laser focused on end user functionality. This isn’t bad in itself, but the fact is that SaaS businesses have to plan to manage operations with the least amount of cost and friction possible if they want to reach their full potential. Understanding the operational aspects of their business that can be built into the application and services is critical to avoid situations where business growth and cost control is constrained by operations that cannot scale efficiently.
- Managing to Metrics – The idea of metrics is widely accepted, but in practice, very few early stage SaaS companies have build the data feeds and internal procedures necessary to have the data for reporting and management. Without planning, it is very hard to back the data out of applications that are not set up to provide it.
- Product Planning – SaaS product planning is part business roadmap and part customer management in a carefully orchestrated system that produces what users need rather than simply complying with their wants. When it is done right, it yields a service that can continue to grow with the user community over the long run and produce a competitive edge that is very hard to overcome.
Recently, one of our partners, Rich Chapman of Softletter (SaaS University) asked us if we had developed a spreadsheet people could use to model SaaS metrics. With our client experiences in mind, we realized that we have talked a lot about using modeling to test business assumptions, but hadn’t provided any insight into how someone could build a simple modeling system and what data feeds and metrics dashboard might include.
To solve that problem, this article includes the Scio SaaS Metrics Workbook. This is an Excel Workbook with two spreadsheets. The first sheet has sample data included to highlight scenarios discussed in this article. The second sheet is blank with just the formulas so that you can use it without having to go through the process of cleaning out the sample data. So – if you want to “play along” as we discuss these metrics, you should go ahead and download the spreadsheet now.
A few caveats apply to this example spreadsheet:
- Every business model is its own special case. These metrics and the formulas they use are generally accepted and useful for modeling purposes. They are not however, a “one-size fits all” approach. For a deep dive into SaaS Metrics – I strongly suggest you look at Joel York’s excellent articles on the subject and Bessemer Ventures PDF that came out last Fall.
- A spreadsheet is not a complete approach to actually managing a SaaS business to metrics. Unless it is integrated with actual data feeds, a spreadsheet will require far too much manual data management and will eventually be discarded in the ordinary course of dealing with day-to-day issues. The best approach is to build the data feeds and display them a dashboard in the application. Unfortunately, for most implementations this means feeding some information into the report either from accounting processes or manual input. But, automating the reporting as much as possible is critical if you actually want to use metrics for managing your business. This means planning to provide the data as a part of application requirements. If you don’t, trying to get the data without the “hooks” built in will often prove impossible.
- This spreadsheet uses a 12 period window. The length of the periods don’t have to be months, although the Customer Acquisition Cost Ratio is generally managed on a quarterly basis. In a dashboard approach, a walking window could be very useful to see trends over time. The width of the window used is a decision you have to make to fit your business model. Of course, in the best case, a dashboard would use graphs of the key metrics with drill down reports just a “click-away” for more detail.
- Green, Bold entries should be feed directly from application data. This means your application needs to be built to capture and report on this data. If you don’t have these feeds at a minimum – it will be very hard to manage to metrics. Line 44 may bring some questions. Isn’t total revenue the same as renewal revenue? No, not in all business models. If you have transaction based revenue or utility-based features, your total revenue is different than your total renewal revenue.
- Blue, Bold entries need to be entered either directly or feed from accounting processes. This includes sales and marketing costs and operational costs. It is “fussy” to have to do this manually, but if you can develop a process and flow, it can be just be part of regular cost reporting within your organization.
- Clients are the entities that pay for the service. In a business-to-business SaaS application, this will generally be a company with multiple users. End-users are just that – individual application users who may or not also be clients depending on the business model.
So – with that in mind – let’s look at some SaaS Metrics….
Committed Monthly Recurring Revenue (CMRR) and Average Recurring Revenue (ARR)
Because the SaaS business model is predicated on a steady flow of recurring revenue, having a clear picture of your current standing is key. Using this figure, you can also derive the Average Recurring Revenue per Client which is helpful to understand if clients are generally adopting larger commitments or the average size is decreasing.
You will also see in this section, a set of formulas we use over and over. This is the Percent of Change from period to period and the Average Change. This is not a true trend line analysis, but it gives you an indication of where the metric is heading over time. This is important because in the day-to-day management of a business it is easy to miss the trends that are actually driving your business. In the example you can imagine a sales team crowing about their 28% growth in the 8th period and their 15% growth in the 12th period. But in the end, the average change of 6% tells a more important story of how recurring income is growing over time. Average change or trend analysis is a good way to smooth out the spot variations and help the team concentrate on the larger picture.
Retention Rate (RR) or Churn
When CMRR is growing, it can be easy to ignore the reality that many of the new clients don’t stay around long enough to pay off the cost of customer acquisition that comes from sales and marketing expenses. Startup companies often ignore this, hoping “we’ll make it up in volume.” The truth is that churn represents a steady cash leak that can eventually sink a new company. Most companies try to manage to 90% retention or better. In the spreadsheet example, you can see that the company is bouncing along the bottom with an average RR of 88.9%. Although income may seem good enough now, if the situation continues the cash burn could become a constraint on continuing product development. If there is underlying user dissatisfaction with the service, this will begin a spiral that the company may never recover from.
In Trial (IT), Time to Close (TC), and Close Rate (CR)
While not every SaaS business has a trial option, since most do, it is wise to have a way to track effectiveness of trials in bringing new clients onboard. Trials do work as a sales tool. The truth is that most SaaS buyers are already actively using Internet-based services and expect to be able to “try before they buy.” In some cases, the sales team may set up special trials for key customers, but generally these metrics do not track those special implementations. The point of looking the data coming from trials it not just to evaluate how many sales come out of them. You can find out a lot about how well the application fills the needs for specific verticals (if you have a way to collect data from prospects), if the application has as much of a “rapid uptake” as you suppose, how many testers typically participate in evaluations, how much data is generated (which can quickly become a push to sign a contract) and many other facts about the decision process. So – if you have a trial option in your business plans – set up data feeds that will help you adjust the application and the trial process to be more effective in closing sales.
Average Deal Size (ADS) or Average Revenue per Client (ARPC)
It is easy to confuse this metric with CMRR. Depending on the options available, deal size can vary depending on the period covered by each individual renewal and changes in the package purchased (number of seats, feature selections, prepaid transactions, etc.). This metric can help sales evaluate if clients are buying larger packages, changing feature sets, etc. Because the SaaS model is heavily biased toward recurring revenue, this figure generally does not include first time sales.
Average Revenue Per User (ARPU)
This metric is valuable in business models where a client buys multiple user IDs. Users typically require available support resources and in some instances may have access to features that burden utility-based resources like storage and data transfer. If you don’t track ARPU, you don’t have any way to determine if your revenue per user is beginning to making the pricing base unsustainable. Where multiple income streams are involved, it can be useful to track that income separately and use data to determine what kind of users are contributing the most to these revenue sources. Multiple income streams, beyond a simple subscription model, can be very valuable to a SaaS business if the subscriptions contribute enough to provide the base revenue necessary to sustain the business. Depending on variable income streams (rather than subscription income) to provide necessary operational income and profit however can be very difficult.
Cost of Service (CoS) or Cost to Maintain (CtM)
The cost of services and operations to maintain the application and client instances. This includes hosting charges, hardware and software renewals, support, staff operations, and outside services – everything it costs to operate. In a traditional hosting model, the costs tend to match the highest level of operational needs expected regardless of load variance over short periods. In a fully virtualized model, the costs can vary on a utility basis if the application is tuned to take advantage of elastic infrastructure capabilities. Regardless though, the operational costs are what they are – and that is what we are trying to model and view here. Here the running change is a key indicator again because it can offer insight into trends you might not see otherwise.
Average Customer Acquisition Cost (ACAC) or Average Cost to Acquire (AtC)
This is a “front-loaded” cost because you have to spend money on sales and marketing ahead of revenue. In web-based sales, it is hard to know the “lead-time” from the first moment a prospect comes in contact with the service through to sale closure. Only in the case of special marketing campaigns that include promotional codes can you actually begin to see how the sales cycle works. In the spreadsheet example, we’ve taken the simplistic approach of counting the previous period sales and marketing costs as the contributing factor for the sales in the following period. You could modify this to set a period that reflects your sales model better, but until you actually know the cycle, this is a good starting point and modeling assumption.
This metric also points out another factor: Most business-to-business sales models for SaaS include different subscription periods so you cannot simply assume that everyone is renewing every period. If the option is available, some may purchase on a monthly basis, some quarterly, some yearly, etc. In these cases, a separate report for the tracking the renewal period choices made by different market segments could be very valuable. These reports can answer questions like, “are larger enterprises actually buying the longer periods as we suppose?” “Is this market really acting like SMB’s and cautiously buying every month?” “Do clients tend to buy longer periods overtime?”
Customer Acquisition Cost Ratio (CAC)
This is a complicated metric and there are several ways you could look at it. The point is to aggregate costs and trends to help determine how long it will take to pay back the your sales and marketing costs on based on your current customer acquisition. The aim is to reach a ration better than 1. This means your costs are being paid off in less than one year by the revenue driven from clients. If the ratio dips to .5, it means it is taking two years. If the majority of your clients aren’t staying two years, you are losing money while acquiring clients. In a startup situation, when you are still carrying your initial development costs, this may be a perfectly acceptable situation – as long as you have enough cash on hand to carry your off the “runway.” So, in that way, you have to know your situation to know how to interpret what it is telling you. The example we have given is good for a “running company” but doesn’t reflect a startup.
Other metrics to consider…
Customer Lifetime Value (CLV) is an often cited metric but one that is hard to model in a spreadsheet like this. We’ve intentionally left it off, but that doesn’t mean you should leave it off your list of requirements. I suggest you look long and hard at the literature before you attempt to implement it though. You need to understand what you can determine in your data model and what is useful from a business reporting point of view before you set this metric for reporting.
There are lots of other metrics you might want to use. Largest Client % can be very important, particularly in situations when the total client load is not large. If a single client provides more than a third of the total revenue – it is time to start thinking about how to acquire a lot more clients to offset the risk that client presents if they “walk.”
If professional services are part of the revenue stream, it is important to be able to capture their contribution separately and to see if the balance between PS sales and application sales is sustainable. I’ve seen several models where the two sides of the business operate hand-in-hand, but seeing the contribution of each separately can help you understand how much of your resources you should be devoting to each.
This is our current modeling tool. We are using it now in our consulting practice and will continue to improve it based on feedback from clients and responses we get from users. It is daunting to look at – so I suggest first playing with some of the green and blue inputs to see what happens. If you are serious about managing your business to metrics, you will soon find yourself on the blank sheet running scenarios. And if you have questions or comments - please let us know!
Recent Comments