Embrace the fear (or SaaS best practices) and move forward
Insights the Inovia team keeps going back to from SaaStr Annual 2017
A lot of amazing ideas surfaced this year at SaaStr — some of which are shaping a new set of powerful best practices. This is a walkthrough of the learnings that remain top of mind. All sources provided at the bottom.
Growth & Capital Efficiency
First off, Kristina & Byron at BVP did an awesome job contrasting Dropbox’s 8-year path to $1Bn ARR with Slack’s growth curve… It’s indeed a wowzer!
An important observation here is that there is more than one trajectory towards building a billion dollar SaaS business. This is reflected in BVP’s chart below which provides helpful benchmarks to keep in mind:
If you’re a founder you’re probably thinking “sure but what about burn?”. Turns out that at an early stage (ie ARR <$30M), there is a lot of noise around this question.
One important thing to note here is that formulas commonly used to compute core SaaS unit economics such as CAC Payback, Churn and LTVs are ok (at best) in low churn environments.
As customer size decrease — from enterprise, to mid market, to SMBs and all the way down to consumers — new dynamics emerge which make it difficult to interpret the output of standard CAC Payback, Churn and LTVs formulas.
For instance, looking broadly at US SMB behaviors, we can see why using formulas which assume a constant churn rate across time can be misleading — SMB churn is not constant and also not linear (see below).
To answer questions such as “when should I step on the gas with sales and marketing dollars?”, we’ve found empirical data from customer cohorts to be far more helpful than standard formulas.
In his blog, Dave Kellogg illustrates how using common CAC Payback formulas can result in significant errors (in the 10x range) compared to modelling out what is going on:
To build useful unit economic models, the best practices we are seeing is to look at empirical data using customer cohorts. Jonathan Hsu at Social Capital has done an excellent job at laying out how to get this done in practice.
Probability models like these ones are also useful in estimating customers’ expected lifetime values — without having to make assumptions such as constant churn.
When churn is material, the key point here is that expected lifetime value changes across a customer’s journey. Using probability models, expected lifetime values can be computed for different stages — an important distinction for growth stage companies that are scaling:
Sales & Marketing
An inevitable topic at SaaStr was sales & marketing. Kudos to David Skok from Matrix Partners for delivering one of the most rigorous presentation of the entire event.
His session on how to establish repeatable sales processes is definitively worth exploring, especially when it comes to direct sales and why Productivity Per Rep should be monitored by rep and cohorts of reps, as illustrated in the charts below:
Another related best practice we are seeing is to forecast bookings using bottoms up Productivity Per Rep assumptions.
Zooming in on per rep data, Mark Roberge (ex-HubSpot CRO) also had an “Aha!” moment in his early days as CRO. He figured out that some of the customer churn was driven by how reps were compensated. When looking at LTV per salesperson, a question emerged — “why is there such a wide range?”:
The fix there was to (i) align reps compensation with LTVs (vs. being compensated for adding accounts that would churn quickly after being onboarded) and (ii) reorganize the company’s go-to-market strategy by buyer personas (as opposed to functions).
This resonated — it’s easy to forget how internal processes may not be aligned with what the customer wants. When going through these exercises, always put yourself in the mind of your customers (and ask yourself if you like receiving cold calls for example).
Product-Market Fit
A couple of interesting leading indicators of PMF were the subject of multiple conversations. The ones we keep an eye on are:
- Logo retention;
- MRR upsells (Quick Ratio is also great);
- MRR growth persistence; and
- NPS (often overlooked in the early days, but powerful).
Make sure you monitor them and that they point in the right direction prior to scaling.
CEO Wisdom
Peter Gassner’s fireside chat was probably my personal highlight. He was able to explain how Veeva pulled off its extreme capital efficiency through a very simple mindset: “Why bother if my customers are not willing to pay more than what it cost me to service them?” — simple ;) Even when considering that Veeva was built on Force.com (which helped keep engineering costs down), the Veeva story is amazing — check out how they compare relative to other successful SaaS IPOs below.
One of his other words of wisdom were on how to think about forecasting. In its first year, Veeva’s forecast horizon was one quarter — what’s the point of going beyond that in the land of extreme uncertainty? Once they passed the one-year mark, management at Veeva then extended the forecasting horizon in years two and three to 12 month plans. They are now looking at three year plans post IPO.
SaaS community
“Scaling Together” felt incredibly appropriate as a theme when considering how open & collaborative the SaaS community has become. It’s really fascinating to see how insights such as the ones explored above inform the decision process of so many great management teams.
We get to meet and work with a lot of amazing people at Inovia — a big part of what we do is setting folks up for success. This is how we’re humanizing venture capital.
Our SaaStr CEO dinner in SoMa 🚀 Scaling together really feels like magic.
SaaStr content highlights and sources
SaaStr preso
- Bessemer Venture Partners, The State of the Cloud for 2017
- David Skok, Matrix Partners, 12 Key Levers of SaaS Success
- Mark Roberge & Michele Law, Sales mistake that can kill your SaaS business
Other sources
- Peter S. Fader, Bruce G. S. Hardie, Probability Models for Customer-Base Analysis
- Dave Kellogg blog, CAC Payback Period: The Most Misunderstood SaaS Metric
- Jonathan Hsu, Social + Capital, Cohorts and (revenue) LTV
