Ryan Lieber, Snowflake: selling in stealth, category creation and expansion
Snowflake's Employee #40 and first SDR shares his learnings from Snowflake's formative years
Hey friends! I’m Akash 👋
Welcome to Missives, where I write about startup strategy across software and fintech. You can always reach me at email@example.com.
Thank you for reading! If you enjoy these Missives, please share them with your friends and colleagues 🙏🏽. Wishing you a great week!
Current subscribers: 3,120, +35 since last week
Join 3,000+ founders, operators and VCs getting smarter about SaaS and FinTech
Today we’re speaking to Ryan Lieber. Ryan joined Snowflake in 2014, when the company was still in stealth mode, as the first SDR hired in San Mateo, California. He has seen hypergrowth over the last 9 years and in 2017 Ryan took the opportunity to support Snowflake's expansion into the EMEA market and made the move to London to help open the first international office. Following the development of a number of successful partnerships and acquiring new customers, Ryan became responsible for Snowflake's presence in Israel, where he developed a $12M ARR business in just two years. This journey led to where he is now, leading Snowflake's Startup Program across EMEA. The program's mission is to partner with the next generation of data-intensive startups to help them develop their applications on top of Snowflake’s Data Cloud, creating scale and future-proofing their business.
Thank you Ryan!
Ryan, you’ve held multiple roles at Snowflake since joined 9 years ago as the 40th employee. Can you give us a quick tour of your time at Snowflake?
I feel very fortunate to have been on this ride for the last nine years. Snowflake was the first company I joined after graduating from the University of California Santa Barbara in 2014. Part of my degree was technology entrepreneurship, and I always knew that I wanted to work for a startup. Life works in wonderful ways, and in hindsight, probably the best thing to happen to me was not passing my second Oracle interview while interviewing while at university.
I joined Snowflake in September 2014 as employee number 40 and was hired as the very first Sales Development Representative (SDR). I was responsible for prospect outreach and different pipeline generation activities. This was extremely difficult at the time given that Snowflake was still in stealth-mode, did not have a working website, and companies still really didn’t trust the cloud back then.
After finally experiencing success as an SDR, I was promoted to become our first Inside Sales Representative, and was responsible for selling into SMB organizations for all states and countries East of Tennessee, including handling international inbound leads. This meant being in the office from 6:30am to field calls from Europe.
In mid-2016, our VP Sales Chris Degnan (now Chief Revenue Officer) shared with the company that we would be opening offices in London and there may be opportunities for people to go over and help open the EMEA region. This was immediately of interest to me given that I was at a point in my life when I was looking to leave California. Once Chris announced this, I immediately built a business justification for why I should be the one to go open the office. We had the meeting the same week and once we got into the meeting, he effectively told me to pack my bags. I didn’t even get the opportunity to present my business case.
I moved to London in March 2017, where we had hired a team of three to start our EMEA operations. I describe the next 12 months as my X-Factor year because I did everything and anything that needed to be done. This ranged from cold-calling, running sales trainings, arranging lunch deliveries, handling partnership calls, and my least favourite: I was our EMEA IT person.
My most challenging, but rewarding role lasted from 2018 to 2022, where I was responsible for establishing Snowflake’s presence in Israel. I was the only Snowflake sales person selling into Israel for two years, taking it from zero dollars in revenue to an ARR of $12M. I truly believe that if a company can successfully establish and sell its product in Israel, then it can be sold anywhere.
Once Snowflake decided to hire a sales team in Israel, I thought this would be the perfect time to follow a passion I had developed which was selling and partnering with startups. This is what led me to my current role as Snowflake’s Startup Program Manager for EMEA.
Snowflake is a textbook example of category creation, ushering in the cloud data warehouse category. Can you share some insights from your first years relating to how you approached sales when doing category creation? Was it an education sale or an execution sale?
Category creation is a hard thing because it requires a lot of rejection, evangelizing and learning reinforcement, day in and day out, which is not for everyone. But the most difficult activity of all is trying to over-come misperceptions and fear, especially when it comes to displacing a legacy industry/technology - on-premises databases - which people have become so accustomed to and built their careers on. Getting these individuals comfortable with moving their precious data to the cloud was one of the hardest tasks. In 2014, the vast majority of people were storing data and running compute on-premises in Postgres, mySQL, Teradata, SQL Server, Hadoop, Oracle etc.
At that time, Snowflake began as what many refer to as a cloud data warehouse, yet we were much more than that, and had much grander product ambitions. Fortunately for us, our new platform was truly innovative and incorporated many unique and valuable USPs, such as being the first company to separate storage and compute which eliminated the contention between the amount of data that can be stored and query performance, and the ability to natively ingest semi-structured data. Our competitors lacked these differentiators because they were not truly cloud native; however, they did have market share, brand awareness, and executive relationships, which was just as difficult to fight against.
We figured that if we could find individuals that understood the importance and significance of these technological improvements, then we could make them early champions of Snowflake. In order to identify these individuals and companies, we had a lot of conversations about what roles within a company would receive the most value from Snowflake, but also which industries suffer the most from not being able to easily ingest semi-structured data.
After thousands of calls and outbound emails, we noticed a pattern of media, advertising and gaming companies who suffered the most from spending time and engineering resources trying to create data pipelines that efficiently structured and made queryable JSON, Avro, XML, Parquet, and ORC data. Ultimately, we decided that companies who were trying to process clickstream and mobile data would be the early adopters of Snowflake. Snowflake’s first customer HUMAN (formerly WhiteOps) is still a customer to this day.
One of the early sales directors gave me a line summarising how to approach selling into a new category. He said “don’t try to sell the religion and the bible. AWS is already trying to sell the religion of the cloud.” The way to interpret this is that if I was speaking with someone who didn’t believe in the cloud, I shouldn’t waste my time with them because they’ll never believe in utilizing a technology built on the cloud. This phrase can be applied to any industry.
To answer the second part of the question, it was both. The sales process 100% started out as an education sale because no company had ever separated storage and compute, nor had the ability to natively ingest semi-structured data. Most prospects didn’t believe us when we told them about these USPs. Our ultimate goal was to get them to a demo, where we could actually show them Snowflake and our unique differentiators.
Once the prospect was bought into the technological capabilities, then it turned into an execution sale. The initial people we were educating were not typically decision makers. After the demo phase, we would ask who will ultimately be making the “Go” or “No Go” decision, and then ask to be introduced to them. In the early days, it required assistance from our CEO, founders, VP Sales Engineering, and VP Sales to properly close a deal, regardless of the actual deal size. At that moment in time, we just needed happy, referenceable customers. Part of the execution sale was selling the value of being an early Snowflake customer and sharing the long-term vision of the company.
You were the first SDR hired by Snowflake’s current CRO. Snowflake is widely recognised as being a world class GTM machine under Frank Slootman’s leadership. Your colleague Lars Nilsson, VP of Global Sales Development, gave some excellent insights on Snowflake’s approach to sales development. Could you share some of your learnings around things like the optimal SDR:AE ratio, ways to set SDRs up for success, metrics for success, and compensating for the right incentives?
I am relatively far removed from the current structure of the SDR org, but I can share what the original organisational structure was like. When I joined, we had a VP Sales, and three outside sales reps covering the East Coast, Pacific Coast, and Pacific Northwest/Rockies. In order to give the early sales team the highest chance of success, they needed to sell into the territories that had the most advanced cloud adoption. In addition, we had three inside sales individuals to target the SMB sector. My role was to try and set meetings with individuals responsible for making decisions related to a company's analytics and architecture in order to help build a qualified pipeline.
An important quality that all of these sales individuals had is that they were all hunters and were coachable. Those were the two of the most important characteristics at the time.
Our CRO Chris Degnan’s metric for success at the time (and still is) was that these sellers needed to go on 8 sales calls a week. If sellers could generate their own meetings, and therefore opportunities, there would be a compounding effect that would ultimately result in success. Especially in the early days, Chris believed that any meetings set for the sellers would be a bonus, and that they should be expected to generate their own pipeline. If a seller can not break into accounts and set their own meetings, then that was a huge red flag.
As an SDR, my expectation was that I would arrange six new meetings per week. At the time, we were all about meeting with anyone who would speak to us. In hindsight, we spent a lot of wasted time meeting with consultants and individuals with the wrong persona. Quantity not quality was the motto at that point in time. But to be fair to ourselves, we didn’t truly have product market fit and needed to speak to a wide range of people in order to gain feedback. Fortunately, this model has drastically changed and Lars has created a very efficient and productive SDR org.
On the compensation side, an important decision that Snowflake’s leadership made was to adopt a consumption-based compensation model where sellers are still paid on a percentage of the deal, but an important part of their compensation is on how much the customer is actually consuming. This model helps better align the success of both customer and the seller.
Can you give us some examples of adversity you faced early on in Snowflake’s GTM and how you rectified it? Examples we see commonly are low conversion on MQLs to SQLs, sales cycles lengthening, buyer inertia, among others.
We just discussed how being a category creator often brings up skepticism and fear from potential customers, especially when they perceived that you were asking them to “give up” control of their data. In reality, Snowflake’s platform was much more secure than storing data on premises. Trying to overcome this perception is what stands out the most in my mind. We achieved this by having multiple security-focused meetings with the people at Snowflake who created our security capabilities. In the early days, it was a team effort at every stage of the sales process. When hiring these early technical folks, it is important to let them know that there may be an element of customer-facing activities which will be required.
As important as it is to know which opportunities to pursue, it is just as important to know which opportunities to walk away from. As a seller, saying no to a potentially huge deal is incredibly difficult to do. We had to do this a couple times in the early days, particularly when it came to trying to displace large on-premises environments. The larger the customer, the more feature demands they would have, the more engineering resources the opportunity would require. Snowflake had to be brutally realistic about its chances of successfully completing a large and complicated on-premises migration versus diverting precious engineering resources away from product development. Ethics aside, just because you see a whale while fishing, doesn’t mean you should try to hunt it. Could your boat even store the whale?
On the sales side, once we had successfully launched out of stealth mode, announced our Series B funding, and the adoption of the major cloud vendors businesses (AWS, Google, and Microsoft) started to accelerate, Snowflake started to gain more notoriety and brand awareness. Our “we’ll speak with anyone” approach shifted to be more selective.
In mid-2015, we began to place a greater emphasis in Salesforce hygiene and tracking. This time period was when we started adopting the terms MQL and SQL. What stood out was how the conversation rate was directly correlated to the relationship between the SDR and the sales rep they worked with. Oftentimes, sales reps do not spend enough time on educating and aligning the proper expectations with their SDRs. The most successful SDRs with the highest number of SQLs, were the ones who were targeting the most relevant ideal customer profiles (ICPs) that had been pre-agreed on with their aligned sales rep.
The conventional wisdom is that you need a pipeline coverage ratio of 4-5x to have comfort about hitting your projections. Organisations can measure the health of their pipeline in a number of different ways - can you give us a sense of how Snowflake analysed the robustness of its pipeline?
I’d say that ratio is correct, and if achieved, will help sellers get much better sleep at night. As Chris Degnan always told us, “pipeline cures all”. That is the absolute truth. Poor pipeline tends to steer reps to make suboptimal decisions and embellish forecasting, which culminates in a trainwreck at the end of the quarter.
From Day 1, Snowflake implemented MEDDPICC as its sales process. The famous John McMahon created MEDDPICC during his time in leadership roles at BladeLogic and BMC Software. John would later come to sit on Snowflake’s board and help provide guidance on sales strategy. Hiring leaders who have experience in truly being honest when it comes to MEDDPICC is how a company can analyze the robustness of your sales pipeline, but most importantly, help understand where there are gaps. Not being honest about the different aspects of MEDDPICC are what results in deals slipping at the end of the quarter or financial year.
In order to further reinforce and document MEDDPICC for every deal, we added these fields into our Salesforce Opportunity page. Whenever a deal is being reviewed, managers can quickly review these aspects of the opportunity in order to truly understand the health of a deal.
Another area of your expertise is attaining loyal first customers. There are several schools of thought around this topic, such as underselling initially, customer advisory boards, and more. What were some of the ways Snowflake developed strong logo retention in its early days (aside from the cost to rip and replace)?
Snowflake’s number one core principle is Put Customers First. This principle often guides decision-making when it comes to what capabilities to build and how, sales strategy, event creation, pricing decisions etc. I also believe this principle is a strong contributor to why Snowflake has a NRR of 142% and NPS score of 72.
In the early days, I believe we had sellers who weren’t just fantastic hunters, but also were exceptional at building champions within accounts.
Our definition of a champion is someone at the customer who will sell on behalf of Snowflake when we’re not in the room. The core foundations for champion building are based on trust and reliability. From our first CEO, to our founders, to the sales engineers, everyone demonstrated these qualities to the customer. Providing these initial customers with access to our brilliant founders who were highly regarded in the database space, and to our CEO who was previously a top leader at Microsoft, made the companies feel like we were truly invested in their success. The fact that no one at Snowflake was too important or too busy to hop on a customer call was truly a differentiator.
On the technical side, because we were a young company with a product that was still maturing, things would break. However, when things broke, the engineering team was quick to respond and make the necessary fixes. We fully understand that for our early champions, deciding to purchase an unproven technology like Snowflake was a risky career move. We never wanted to feel like we were letting the customer down, which is why we placed such a high premium on ensuring that Snowflake worked as advertised given the product maturity at the time.
One principle I’ve heard more and more recently is having engineers join sales calls to give product and engineering empathy for what customers want. Can you give us some insights into how Snowflake approached collaboration across its sales, product and engineering orgs?
As much as engineers would like to live behind their keyboards (and some for good reason), we knew the importance of having our early technical individuals engage with early prospects and customers. Snowflake’s early engineers were highly exceptional talents who see the world differently from us sellers. They would ask questions and infer customers' painpoints which we otherwise were not capable of articulating, let alone deciphering.
As a seller, I was always very careful not to sell or promise features that didn’t exist. In the early days, there were a lot of features that Snowflake lacked. If a customer mentioned a particular feature, I didn’t know whether this capability was easy to implement or extremely difficult. There were multiple times on calls where engineers would take the time to hear customer suggestions about product improvements, and sometimes these requests were actually relatively easy to implement. The engineers would then go away, implement the feature, and the customer was truly impressed that we took their suggestion. It’s these potentially small acts which actually had a significant impact in developing trust between Snowflake and its customers.
You were the first Snowflake employee to help the company go international and successfully penetrate UK/Europe. Can you share some of your learnings from launching in Europe? What were the nuances you had to cater to in Europe? We know that different European markets require a very localised sales team typically, for example.
From a people perspective, there was a book that our first VP EMEA recommended to me called the Culture Map. This book was incredibly helpful in learning and recognizing distinct culture differences, particularly related to high-context and low-context cultures. Even understanding how different cultures would use the word “interesting” to describe Snowflake at the end of a meeting was a small, but important detail to understand.
When it came to breaking into individual markets, reflecting back, I believe we launched the most successfully in markets where we were able to establish a partnership with a very trusted local Solution Integrator (SI), most often a small boutique firm. These SIs helped provide us with local insights such as what verticals/companies were the most cloud friendly, what customers should we initially not bother with, and who were the most influential people in the data space that we needed to speak with. Oftentimes, the first SI we spoke with was not the one we ended up developing the best partnership with. It was critical that we found an SI who was ultimately going to decide that Snowflake would be their preferred vendor to recommend and bet their next 10 years of revenue on this conviction.
From a localisation standpoint, the first sales hires were for France and Germany. Aside from the UK, these were the two most important markets to have a local presence. There were some markets like Israel, which is used to working with sellers who are not based in the region. Although, if your goal is to target public and financial services companies, then you need someone in-region. I successfully sold into Israel for three and a half years while sitting behind my desk in London before we put a full-scale team in place. That being said, I would travel to Israel about four times a year in order to establish that personal relationship, which was a game changer in terms of improving trust with the customer moving forward. Also, if you haven’t been to Tel Aviv, it is an amazing place to visit.
Validating LLM Outputs from Cohere
Braid Is Dead, Long Live Braid from Amanda Peyton (must read)
Thematic vs Thesis Driven Investing from Fred Wilson
Quote of the week
‘But both thematic investing and thesis driven investing are better than a generalist approach because they both promote domain expertise which is critical to building a sustainable investment advantage. I think "generalist" or "opportunistic" investing is likely to underperform domain expert driven investing in all but the most turbulent markets.’ Fred Wilson