Bessemer’s Top 10 Laws of Cloud Computing

Bessemer’s Top 10 Laws of Cloud Computing The new principles of “Cloudonomics” Co-authors: Adam Fisher, Bob Goodman, Byron Deeter, David Cowan, Ethan Kurzweil, Felda Hardymon, Jeff Epstein, Jeremy Levine, Kent Bennett, Peter Lee, Siddarth Nautiyal, “Subu” S.V., Trevor Oelschig, & Umesh Padval 2012

Bessemer Venture Partners (BVP) is a global venture capital firm with offices in Silicon Valley, Cambridge, Mass., New York, Mumbai and Herzliya, Israel. BVP delivers the broadest plat- form in venture capital spanning across industries, geographies, generations and stages of company growth. From Staples to Skype, VeriSign to Yelp, LinkedIn to Pinterest, BVP has helped incubate and support companies that have anchored tidal shifts in the economy. More than 100 BVP-funded companies have gone public on exchanges in North America, Europe, and Asia. Bessemer Venture Partners has been a leading investor in Cloud Computing for more than 12 years, supporting early pioneers in this high-growth market including Postini, Cyota, Trigo, and VeriSign. BVP’s current cloud portfolio is one of the largest in the venture capital industry and includes leading public and pri- vate companies such as Box, Broadsoft (BSFT), Cornerstone On- Demand (CSOD), DocuSign, Eloqua (ELOQ), LifeLock (LOCK), LinkedIn (LNKD), Shopify, Twilio, and Wix. For more informa- tion, please visit www.bvp.com/cloud. Bessemer Venture Partners 2

At Bessemer Venture Partners, we have had the privilege of working closely with over 50 of the leading Cloud Computing companies over more than a dozen years. We continue to fundamentally believe that the emergence of Cloud Computing – and its three core components of Software-as-a-Service (SaaS), Platform-as-a-Service (PaaS), and Infrastructure-as-a-Service (IaaS) – is completely changing the landscape of the multi-billion dollar software industry. We first presented Bessemer’s Top 10 Laws for Being “SaaS-y” for internal discussion at our annual Cloud/SaaS CEO Summit almost five years ago, and were overwhelmed by requests to share the content more broadly. We decided to not only share the insights and the discussion openly, but to treat this as an active dialogue with ongoing updates to reflect the collec- tive learning of the cloud community. This update is the most extensive we have ever undertaken, and includes many changes and several entirely new concepts for this publication. This work is literally the result of thousands of conversations that the partners of Bessemer have participated in with cloud executives, including our past and current portfolio companies, as well as other leading public and private cloud companies. We have been fortunate to be investors in many of the early cloud winners - such as LinkedIn (LNKD), Cornerstone OnDemand (CSOD), Broadsoft (BSFT), Eloqua (ELOQ), Lifelock (LOCK), Postini (Acq: Google), Netli (Acq: Akamai), Trigo (Acq: IBM), Keynote (KEYN), Cyota (Acq: RSA), and Verisign (Acq:SYMC) - and continue to ac- tively invest behind one of the largest cloud portfolios in the venture capital industry. We don’t claim to be brilliant on these topics ourselves, but we do believe this full community – and thus the content - represents the leading thinking from true cloud experts. Periods of tremendous transformation create tremendous opportunity, and we consider ourselves privileged to be working with many of the great entrepreneurs who are currently creating the next giants of the “software” industry. We welcome your thoughts and feedback at [email protected] and also invite you to visit the Cloud Computing section of our website at www.bvp.com/cloud. Bessemer Venture Partners 3

#1Drink Your Own Champagne Drink it, live it, love the cloud. Use your own product, and that of your customers, partners, and peers. To understand the cloud revolution, you have to be a part of it. You need to understand the issues, challenges, and opportunities with cloud deployments – starting with your own. If you have a product that touches end users, then end users within your company should be power users of the product. You should all become experts in the strengths and weaknesses of your product, and be able to discuss customer issues and roadmap priorities in some detail. Most of our cloud portfolio companies are power users of their own products and even those of their competitors. Similarly, you should leverage the cloud for your internal systems. This will not only give you a direct understand- ing of the customer experience and best-of-breed strategies of cloud businesses, but it will free up your technical resources and balance sheet to focus on your core product and customers. Although your technical team may be great Bessemer Venture Partners 4

at setting up email servers and voice systems, they are far too valuable to your organization to waste their time on such rote tasks. Your cost of capital is also likely very high, making upfront hardware and license costs unnec- essarily expensive for a young company. By pushing as much as possible into the cloud, you avoid management headaches and make these expenses variable. We highlight many of the best-of-breed offerings by category in Bessemer’s Cloudscape, and there are dozens more emerging monthly. Cloud businesses should also leverage PaaS and IaaS whenever possible for core product development. There has been a massive change in the consumer internet world over the last few years, with the vast majority of Michael Tessler new internet websites and applications now using cloud environments such CEO, Broadsoft (NASDAQ:BSFT) as Amazon Web Services as the foundation of their development from the first day. We believe that quality of service and portability issues are rapidly “We enable telephony services delivered resolving themselves at a level that will be suitable for enterprise deploy- from ‘the cloud,’ and every employee ments, and a similar migration will occur within the software and Software- is a power user of our communication as-a-Service worlds over the coming years. There has also been an explo- services.This way everyone in our com- pany is an expert on the product and can sion of cloud based PaaS services that are part of the “developer citizenry” easily understand a user’s needs, from the revolution, giving application developers access to a long list of powerful veteran sales executive APIs that can be quickly integrated into product offerings, and purchased to a new customer on a consumption basis with little or no upfront commitment. account manager.” #2 Build for the Doer, Build Employee Software Employees are now powerful customers themselves, and not just through their managers. We’re witnessing the “Con- sumerization of Software,” so focus on ease of use for SaaS, and “developer citizenry” for PaaS and IaaS. The gig is up. Pandora’s box is open. Your customers all now know that software doesn’t have to suck anymore. They use rich internet applications including Facebook and Skype to communicate with their friends; they use LinkedIn to manage their business networks, Google or Wikipedia to find accurate online content, Yelp to find restaurants, and Travelocity to book flights. Your potential customers are now looking for similar “cheap and cheerful” products in an open revolt against the years of oppression by the likes of SAP and Oracle. You should therefore beg, borrow, and follow: take inspiration from the best online products you can find and leverage the fact that you’re naturally smaller and more nimble than the incumbents to provide the best user experience imaginable. Whether it’s a “Freemium” model, a hybrid sales model with a heavy inside corporate sales element, or even an enter- Bessemer Venture Partners 5

prise sales focus with products that delight the user, building for the end user in SaaS will drive adoption and thus monetization. Products will now see rapid adoption by virtue of being intuitive and dynamic as opposed to being confusing and complex. Customers no longer require you to capture every use case or business need in your product, and they’re willing to forgo con- siderable flexibility in return for rapid on-boarding, progressive discovery, and context-sensitive help. David Patrick Individual employees and mid-level managers can now take out their corpo- CEO, Apperian rate credit card and expense products, and are becoming direct consumers in the process. The best possible way to land a large enterprise customer is to “Our goal is to provide a platform for call up the CIO and say “we’re excited by how much you like our product and enterprise mobile teams to be we’re happy to note that we now have several hundred users of our product successful. We try to hide all of the within your corporation. We wondered if you were interested in rolling these complexity around issues such as security, role based provisioning, and scalability, into an enterprise license with the administrative dashboard, integration to and just let them provide great mobile your other systems, coordinated billing, provisioning and security?” Many applications to their end users.” cloud companies are doing this with great success. Although this is finally becoming more widely accepted as a best prac- tice, we must still emphasize the importance of building a single instance, multi-tenant product, with a single version of code in production. “Just say no!” to on-premises deployments. Multi-instance, single tenant offer- ings should only apply to legacy software companies moving to a dedi- cated hosting model because they don’t have the luxury of an architectural redesign. Of course it is possible to use virtualization to provide multiple in- stances, but this hybrid strategy will make your engineering team much more expensive and much less nimble. You also want to leverage your core in- Jeff Lawson frastructure as much as possible, even when expanding internationally. This CEO, Twilio generally means investing early in backup and disaster recovery, but if you’re managing your own datacenter, avoid a second production facility as long as “We now enjoy an active community possible (at least past $2M CMRR). of over 100,000 developers behind a pretty straightforward philosophy: Make a Hero out of your Doers. We’ve also started to see unprecedented levels of developer empowerment Empower them to get stuff done by within organizations, meaning that PaaS and IaaS vendors that focus on killer building great products leveraging offerings for the end developer are being discovered and embraced at fantas- our communication platform. Make it tic rates. Let the developers make your decisions, and understand that they easy to use, transparently priced, and above all, no shenanigans!” will be making the decisions for your customers. Build and use clean APIs, and promote your offerings through user conferences, hackathons and devel- oper evangelists to spread the message. Bessemer Venture Partners 6

#3 Death of the suite; long live best-of-breed and even best-of-feature For most of the last two decades, major software vendors such as SAP, Oracle, People- Soft, Microsoft, JDA, and others have pushed the concept of an “integrated” software suite on the market. With Cloud Computing, the pendulum is swinging back forcefully in favor of best-of-breed applications. The high level message from suite vendors to prospective customers was the idea that purchasing all of your primary business software from a single vendor had significant benefits to the end customer in the form of system interoperability, consistent archi- tecture, common look and feel of applications for end users, and deeper vendor com- Rob Reid mitment. There was some real appeal in this positioning, and in the 1990’s the client- CEO, Intacct server applications and infrastructure stacks from these vendors were state-of-the-art, so customers embraced these “integrated” solutions en masse. “With a suite, if the vendor falls behind you Unfortunately however, there were some real downsides associated with selecting a are locked into subpar business performance. suite strategy, because no single vendor was a leader in every application category. As With best-in-class, if a a result, the customer was often forced to accept second tier applications for many of solution falls behind you their business needs. In companies where the finance team drove the evaluation, Oracle can always replace it, was thrust upon other departments which then suffered through weak HR, operations, and achieve optimal performance. The Cloud and CRM. If Human Resources drove the decision, they would buy PeopleSoft (which allows open and seam- apparently wasn’t as often, because Oracle won), and if the company was in Germany less integration so our or deep in manufacturing, then SAP was often the preferred choice. Of course superior customers often work with best-of-breed options were often available in many of the functional areas, but the suite Salesforce.com for CRM, Cornerstone OnDemand vendors did their best to spread fear, uncertainty, and doubt (F.U.D.) into the market for HR, and Eloqua for around these offerings and integrations between systems were often quite painful. Marketing and pick Intacct as the very best Finance Therefore, for the better part of the last two decades the job of the CTO in major cor- and Accounting solution available.” porations has centered around this complex decision of where and when to use a suite versus best-of-breed solutions. Should they buy a suite, with easier integration but limited functionality, or best-of- breed, with optimal performance but higher integration costs? This decision was made even more complex by the horror stories from many large corporations attempting to integrate large suite offerings across their companies with staggering costs and implementation times. Software licenses frequently ran into the millions of dollars (or tens of millions!), professional services would ultimately be another ~3x the software license costs, and the more you cus- tomized and configured the product to fit your needs, the more expensive it would be and the harder it would be to implement the next version. Many companies publicly disclosed spending upwards of $100M and 5+ years attempt- ing to deploy systems, often cancelling the entire project midway through and throwing it away or trying to unwind the initiative, leaving a trail of fired IT executives along the way. A large part of the momentum around Cloud Computing today is because IT departments now realize they can avoid Bessemer Venture Partners 7

many of these implementation headaches and functionality shortcomings, and instead get the best of both worlds by working with best-of-breed vendors. Cloud Computing provides the opportunity to leverage best-of-breed applica- tion offerings, with the standardization and pre-integration of many of the applications and APIs. You can pick the world’s best application for every need, every user, and every business case. You can deploy exactly the number of seats you need, where and when you need them. Since the internet is the common underlying infrastructure, deployments can now be done in days or weeks, and service ratios are a small fraction of the software subscription costs. This means that with Cloud Computing, the pen- dulum is swinging back to best-of-breed, and away from integrated suites. Better APIs are now allowing users and developers to literally leverage only the best product services, including the ultimate consumerization of software down to the “best-of-feature” level. This fragmentation means more choices and more pricing transparency for end users and application developers alike. Of course, a handful of vendors are working hard to also create a generation of SaaS suites as well, and some will likely emerge to have success. But for the foreseeable future, these companies won’t claim to address all (or even most) of the application needs of an enterprise, but will instead carve out multiple vertical slices for an enterprise and then highlight preferred ecosystem partners to fill the gaps. Workday is having considerable success running a focused version of the “PeopleSoft for SaaS” strategy for very large enterprises, NetSuite has carved out market share in the mid market, and even the narrowly-named Salesforce.com is now responding aggressively by building out several new functional clouds. However it is now easier than ever to compete as a best-of-breed business. What does this mean for cloud entrepreneurs? The entire software landscape is now open. It’s the great land rush. It’s noon on April 22, 1889 in Oklahoma and the gun just went off. Now’s your chance to plant your flag and claim a valuable plot of land in the software landscape, while the incumbents are giving it away. Dr. Jonathan Golovin CEO, Retail Solutions “The CPG industry is undergoing a massive change as social, mobile, location services and personalization change their ability to interact with shoppers and consumers. The cloud offers our clients the ability to immediately leverage the RSi offerings, to gain business value with focused applications that leverage massive data sets and gain access to the most relevant and up to date information on mobile devices to drive both savings and increased sales.” Bessemer Venture Partners 8

# Grow or Die 4 In technology, very few things remain constant. As a result, you either grow up to be- come a dominant company in your category, or get passed by and killed off by some- one who does accomplish this goal. Not surprisingly, growth rate is often the biggest driver of valuation multiples in both private and public markets. Investors, employees, and partners aren’t buying into your current company as much as they are investing into some future version of your business, and growth rate determines the size of the business, at that future period. The power of compounding numbers is straightforward but still surprising, when you Aaron Levie consider that a $1M business that grows 100% each year will be a $1B revenue busi- CEO, Box ness within 10 years, whereas the same business growing at 10% per year will reach less than $2.6M in revenue a decade later. Both growth rates may sound interesting “At Box, we discovered that the against GDP growth, but in technology, that’s the difference between life-changing market for cloud-based content wealth for a core management team and a wasted decade. management and collaboration was far larger than any of the analysts or existing As a senior executive of a cloud business, you will likely want investors to pay you a players had recognized; because huge valuation based on current financial metrics, and you will need to convince pro- cloud solutions are instantly avail- spective employees to walk away from rich cash compensation packages from others able to businesses of all sizes and all geographies, and new mobile in exchange for the potential upside of options for your stock. How do you do this? devices and platforms enable Present a credible plan for capital efficient hyper-growth. every worker to bring technology into the workplace, opportunities For a business growing 300+% a year – as many of our today are 10x larger than tradi- tional enterprise software startups. early investments often do – the discussions are much The corollary to this re-sizing, of easier because time is an easy variable. Valuation gap? course, is that you must grow No problem, just wait a few months and the business will rapidly for a sustained period to grow into it. Even if you make an investment and later win in your market.” believe you are off on “fair” valuation by 50%, all other things being equal, you will still grow through the gap in just over a quarter. Similarly for the entrepreneur, there is massive Al Lieb value in deploying capital and resources of an investor against the plan to get leverage from CEO, ClearSlide the investments earlier. If you can bring capital into the business earlier – even if it’s at a “Growth really is the lifeblood of lower valuation – and steepen the slope of the growth line to get the compounding effects of any startup. If your product can deploying the capital, then the dilution will quickly pay for itself over time. drive increased revenues for your customers, it becomes highly viral. Related to high level growth are many more detailed metrics that you’ll want to un- The “Land and Expand” sales derstand including areas of acceleration and deceleration in the business and cohort model works well for us, because performance over time. In technology, you’re either getting bigger or you’re getting once a few sales reps start closing smaller. Growth, and ways to efficiently accelerate the rate of growth, should always deals faster with ClearSlide, the rest of their team quickly follows.” be top of mind for your entire team. Bessemer Venture Partners 9

# Play moneyball in the cloud, and check 5 the scoreboard with the 5 Cs of Cloud Finance Consumer internet companies like Google and Amazon were built on deep consumer data and analytics. Complex manufacturing businesses like Intel measure production inefficiencies down to the particle level. Elite athletes and sports teams measure per- formance in excruciating detail to try to highlight areas for improvement and competi- tiv w from other industries and fields, you have to have a set of metrics in place before you can track and then predictably improve upon them. Although enterprise software companies have long organized themselves around a clear set of business metrics – including bookings, maintenance & support fees, and revenue – the new software economy of Cloud Computing is just starting to converge Tobias Lütke on its own set of metrics. After surveying hundreds of leading public and private Cloud CEO, Shopify Computing companies, 5 key “C” metrics now rise above the others as essential top level performance indicators: CMRR, Cash Flow, CAC, CLTV, and Churn. We rec- “The beautiful thing about run- ommend EVERY cloud business track and report on these as a starting point, plus ad- ning an online business is that ditional metrics that are relevant to your teams and functions as appropriate. you have almost unlimited data about your customers and your business. We have a primary 1. CMRR, ARR, & ARRR – Committed Monthly Recurring Revenue, set of metrics, but are then Annual Recurring Revenue, and Annual Run Rate Revenue. constantly running analyses in Many 1st generation cloud businesses turned to TCV (Total Contract Value) or ACV areas like churn, CAC, CLTV, and cohort performance to find (Annual Contract Value) as their top level metric as a carry-over from the legacy soft- new insights into the business. ware world of tracking “bookings.” In the Cloud Computing world, these metrics can In our offices we have big be easily manipulated and are often misleading, and therefore we recommend much screens hanging everywhere more focused metrics around the recurring revenue in a normalized time period. which share these numbers with the entire company.” TCV and ACV are flawed for many reasons, most notably with regard to duration and services. For very late stage companies (like Salesforce.com) with standard contract terms and service ratios, these metrics can still be workable, but for highly dynamic private companies they can be highly misleading. If your renewal rates are strong, then contract duration isn’t a major variable, whereas cash collection and the size of the monthly subscription will massively impact your business (see points on Cash Flow and Churn in the following section). Therefore, a focus on TCV has a tendency to encourage sales professionals to focus on longer term (often multi-year) deals to push up TCV, instead of pushing on the more important elements of monthly subscription value and cash pre-payments. ACV does help to reduce this over-emphasis on duration by just focusing on the first year of the deal, but shares the second major flaw that TCV is also burdened with, which is an over-emphasis on services revenue as part of the “contract value.” To be very blunt with our perspective: professional services revenue is bad for cloud businesses in most cases. It’s low gross margin revenue that slows down your implementations and can only scale in proportion to your services Bessemer Venture Partners 10

headcount. For these and many other reasons, Wall Street investors and your customers hate to see a large mix of services revenue in cloud businesses. You should focus your product development, sales, and client success teams on reducing the implementation friction, time, and cost as much as possible. In most cases, you shouldn’t reward your sales team on “contract value” by giving them quota relief and commissions against services revenue. To address these concerns, many cloud businesses now focus on Monthly Recurring Revenue (MRR), which represent the combined value of all of the recognized recurring subscription revenue on a monthly basis. We recommend compa- nies actually take this a step further and track the forward view of Committed Monthly Recurring Revenue (CMRR) as the primary internal business metric. MRR is a great base metric, but CMRR is more insightful because it includes all MRR, plus signed contracts currently committed and going into production, and minus “churn,” which is the MRR that is no longer committed from customers that have turned off the service, or are anticipated to do so in the future. To prove the point - here are two deal options, which would you pick? Deal A: 6 month prepaid contract; renews monthly; $10k monthly subscription; $10k services. (TCV: $70k, ACV: $130k, CMRR: $10k) Deal B: 3 year contract; 3 months prepaid; $5k monthly subscription; $80k services. (TCV: $195k, ACV: $140k, CMRR: $5k) Despite lower TCV and ACV, Cloudonomics says you should pick Deal A every time. Deal A will gross ~$370k of revenue over 3 years, whereas Deal B will only gross ~$260k. Deal A will also likely be much higher gross margin given the lower services ratio. In fact, there are only two reasons to even consider Deal B and they are related to Churn Risk and Cash Flow, but we also attempt to correct those misconceptions later in this paper. In almost every case, CMRR is the single most effective metric. For businesses that don’t operate with long term contracts – such as PaaS businesses with monthly or consumption based contract terms – “Committed” is the calculated MRR at that instant (as committed to by the CFO, VPS, and CEO combined) based on the current customers in production with their existing consumption levels adjusted for sea- sonality or known usage trends. This is meant to be the best true baseline number of expected product/subscription/ usage revenue for the month, without adding any new customers, upselling any new products, or expanding usage beyond the current footprint, but subtracting out all known churn. This single metric gives you the purest forward view of the “steady state” revenue of the business based on all the known information today. The monthly focus also tends to drive many positive behavioral changes within a team in- cluding a monthly sales and development cadence, better sales compensation plan and cash flow alignment, reduced customer price sensitivity, and heightened awareness around small MRR changes. Many leading cloud companies therefore use CMRR as the basis for everything from the financial model to the sales commission plan. This is the single most important metric for a cloud business to monitor, as the change in CMRR provides the clearest vis- ibility into the health of any cloud business. For external purposes, you will likely want to highlight slightly different versions of these metrics: the Annual Recur- ring Revenue (ARR) and Annual Run Rate Revenue (ARRR). ARR is simply the currently recognized portion of this Bessemer Venture Partners 11

monthly revenue, multiplied by twelve. ARRR is the ARR, plus any non-recurring revenue related to items such as professional services, transactions, and implementations. These external “vanity” metrics can help drive home the run rate scale of your business, especially when used to describe the forward business model. Your current CMRR may be $1.75M and projected to grow to $2.17M at year end, so for external audiences you may get maximum impact by summarizing the business plan by saying: “As we exit this year our Annual Run Rate Revenue (ARRR) should cross $30M, which includes $26M of Annual Recurring Revenue (ARR).” 2. Cash Flow - Start with Gross Burn Rate and Net Burn Rate, then hopefully turn to Free Cash Flow over time. CMRR gives you a great sense for the revenue health of the business, but can very often be disconnected from the “cash health” of the business. As any scrappy entrepreneur will tell you, a business will live or die based on its cash management in the early days, and therefore detailed cash metrics are also needed. Gross and Net Burn Rate (cash flow) metrics are critical for cloud businesses because the working capital require- ments are higher and the payment terms are often back end weighted. Gross Burn Rate is all of the expenses paid for in the month including debt and finance charges. Net Burn Rate is simply all cash received during the month minus all the expenses, which nets out to the cash burned in the month. These numbers are obviously lumpy based on the timing of collections and payables, so many companies further refine this by adding a “rolling 3 month average” Burn Rate set of metrics. Cloud businesses typically show significant positive Free Cash Flow (FCF) long before they turn GAAP EBIT positive, so hopefully you will be able to flip your Burn Rate (negative cash flow) metric to a positive one as you grow, and start tracking FCF instead. By tracking your CMRR and Burn Rate you have a very good sense for the steady-state health of the business. At any point you can divide your monthly Net Burn Rate into your cash balance to understand your “months of runway,” which many of our CEOs monitor routinely. You should always have a variety of insurance plans – which we strongly recommend you discuss as a Board and keep updated – that could include cuts in the business to quickly reach cash flow breakeven. One of the great benefits of these cloud businesses is the recurring nature of the revenue streams, which means that you can very easily model and predict the steps you would need to take to instantly bring the Gross Burn Rate of the business in line with the CMRR, so that you glide into a Net Burn Rate of zero as your working capi- tal catches up. Many of our companies with extremely aggressive growth plans and high burn rates talk openly about their “insurance” or “rip cord” plans, where they can always freeze hiring, slow marketing spend or make targeted cuts if their cost of capital spikes or their sales metrics deteriorate, and still work back to breakeven. 3. CAC – Customer Acquisition Cost Payback Period. The CAC Payback is a statement in months, of the time to fully pay back your sales and marketing investment. This is worthy of much more detail and therefore broken out further in Law #6 later in this paper. 4. CLTV – Customer Lifetime Value. Understand your Customer Lifetime Value (CLTV), because a profitable business rests on the shoulders of profitable customers. CLTV is the net present value of the recurring profit streams of a given customer less the acquisition cost. Part of the attraction of Cloud Computing business models is that once you have repaid the initial Customer Acquisition Bessemer Venture Partners 12

Costs (CAC), the cash flow and profit streams from customers can be quite attractive. However, whereas the CAC ratio can at least ensure that you recover your incremental sales and marketing costs on each customer, it still doesn’t tell you if these customers are highly profitable over time. To measure this, many customers have modified the consumer internet concept of lifetime value, into a similar cloud CLTV metric. To simplify the calculation, let’s assume that a customer generates $10,000 of annual re- curring revenue for a company with a CAC Payback ratio of 12 months, a 70% Gross Mar- gin and 10% each of R&D and G&A costs. The $10,000 of revenue will generate $7,000 of gross margin and $5,000 of profit each year ($7,000 less $1,000 of R&D and $1,000 of Michael Fertik G&A costs). Over 5 years, this customer will generate $25,000 of profit (5 years x $5,000/ CEO, Reputation.com year). A CAC Payback ratio of 12 months means a $7,000 upfront acquisition cost, making the CLTV equal to $25,000-$7,000= $18,000 Obviously if the retention period is longer, “We are constantly looking at and/or you benefit from net positive CMRR renewal rates that actually grow your average our dashboard of metrics to tune the performance of our business. customer relationship over time, these numbers can be much larger. These metrics give us a com- mon framework as we steadily For those who may choose to get more analytical, this is equivalent to ($18,000/5) = find ways to reduce CAC and $3,600 of annualized profit or 36% profit margin. The calculation can be refined with a Churn, while increasing CMRR and CLTV.” better allocation of the S&M costs (the parts used to support current customers) and by discounting the profit streams (in this example, a 15% discount rate would reduce the CLTV to $12,300 or 25% annual- ized profit margin). It’s also worth noting that for young companies, it may be more of an art than a science to estimate the lifetime of the customer as your churn data is still limited, but we’d conservatively take 3-4 years for SMB custom- ers, and 5-7 years for enterprise customers. 5. Churn & Renewal Rates – Logo Churn, CMRR Churn, and CMRR Renewed. It’s very difficult and expensive to grow subscription businesses if you have moderate customer churn-- and prohibi- tive if your churn is high. As detailed financial models of CLTV and Free Cash Flow demonstrate, the single biggest driver of long term profitability for your cloud business (and thus valuation) is the renewal rate of your customers. Whereas the largest legacy enterprise software companies literally made tens of billions of dollars over the last decade with “shelf-ware” projects that never got fully implemented, project failure is not an option for cloud businesses or the customer will simply turn you off, regardless of the contract terms. This is another reason not to overly focus your team on long term contracts, because it creates a false sense of customer lock-in regarding your unhappy customers, and may leave money on the table from lost upsell opportunities with your happiest customers. Over time when you have successfully realized most of the upsell potential of your accounts and have built in deep ties to the account, you may want to push for longer term contracts for predictability (which Wall Street does still value), but this shouldn’t come at the expense of CMRR. Cloud executives need to track renewal rates in detail to capture “logos lost” (lost customers) as well as the percentages of CMRR renewed and lost. The standard approach is three key sub-metrics, all related to this concept of renewal rate: Logo Churn %: This is a percentage calculation of all your customer names (“logos”) that have churned over the measured time period. If you started the year with 500 customers and 460 of them were still paying customers at some level at the end of the year, then you have churn of 40 customers and your annual Logo Churn is 8% (40/500). Bessemer Venture Partners 13

(Note: As with all renewal metrics, you exclude all new customers signed during the time period. They’ll be captured in your next renewal report.) CMRR Churn %: This is a percentage calculation of all your customer CMRR that has been lost over the measured time period. If you started the year with $500k of CMRR for your same 500 customers, and the 40 customers that churned represented $30k of CMRR at the start of the year, then your Base CMRR Churn Rate is 6% annually ($30k of starting CMRR churned/$500k of starting CMRR). CMRR Renewal %: This is a percentage calculation of the total CMRR of your renewed customers at the end of the year, divided by the total CMRR of your existing customers at the beginning of the year. Of your 460 renewed custom- ers, if they have been upsold on new products and grown in their usage of the product during the year to the point where their CMRR equals $550k in total, then your Total CMRR Renewal Rate is 110% ($550k end of year CMRR just from customers who were on board at the start of the year/$500k CMRR of all customers at start of year). The top performing cloud companies can enjoy annual Logo Churn rates below 7% and CMRR Churn rates below 5% - with most of the churn due to death (bankruptcies) or marriage (acquisitions) - and CMRR Renewal rates well above 110% due to upsells into this installed base. Your specific business is likely to have additional “key” metrics that are worthy of show- casing on the top level executive dashboard, but we have found these five to be pretty uni- versal across the vast majority of cloud businesses. You will find yourself reviewing them at different frequency levels: CMRR, Cash Flow, and Churn tend to be highly dynamic and thus daily or weekly metrics, whereas CAC and CLTV are more strategic and thus longer term in their nature. Many of our top performing cloud CEOs have modeled their executive team objectives Sarah Friar and bonus plans around a subset of these metrics exclusively (typically CMRR growth, CFO, Square Churn, and Cash Flow) and we would encourage you to consider doing the same. As “I’ve enjoyed working with the you approach being a public company you will likely chose to keep these metrics in- Bessemer team for over four ternal only and will instead add in “Street Metrics” including GAAP Revenue, Gross years on the 5 C’s of cloud Margin, and EBITDA. finance, back from my days at Goldman Sachs when we Finally, with these metrics in place, use them to drive your rolling financial budgets and utilized some of these CLTV forecasts. Although you will likely want to keep your annual “Board Plan” of record models in our valuation work locked down for the year, you should get in the habit of revising and sharing your busi- and at Salesforce.com when we used some of these metrics ness forecasts monthly or quarterly. Some executives may come to a Q3 meeting and for our own performance and simply publish their second half Board Plan as their forecast, but that suggests the com- to evaluate the performance pany hasn’t learned anything in the 6-9 months since that budget was originally drafted. of our acquisitions. I’m a big There may only be slight variations in expenses, CMRR, or cash flows, but get in the believer that if it’s worth doing, it’s worth measuring with a habit early of revising forecasts and using that as another way of sharing the positive and clear set of metrics, and we’re negative trends of the business. In return, a collaborative Board of Directors should view embracing many of these the forecasts as rough projections and not overreact to information shared. You don’t same CAC and CLTV concepts want your first attempts at forecasting to be as you get ready for your IPO roadshow. at Square as well.” Bessemer Venture Partners 14

# Build the Revenue Engine, and only invest 6 aggresively if you have a short CAC Payback Period Hyper-growth is the goal of most aggressive cloud CEO’s, but how do you know if your sales and marketing investments are ultimately “profitable?” The answer to this question can be found through measuring your Customer Acquisition Costs (CAC) and the CAC Payback Period. When we introduced a similar concept of a CAC Ratio five years ago the response was extremely positive, so we have included it again, but with a significant simplification based on real world feedback from dozens of companies. The CAC Payback Period is now a statement in months, of the time to fully pay back your sales and marketing investment. This single number is the key to deter- mining your level of sales and marketing investment. Total Sales & Marketing Costs of Prior Quarter CAC Payback Period = (Months) New CMRR Added in Prior Quarter x Gross Margin% of the Business It can be calculated simply by dividing the sales and marketing costs of the previous time period (typically a month or a quarter) excluding any account management costs attributed to your “farmer” organization, divided by the new CMRR gross margin added during the same time period (forget the effect of churn and upsells for now). $770,000 11 Months = $100,000 x 70% As an example, if your company added $100k of CMRR in a quarter with 70% gross margins, the denominator would be $70k. If your fully burdened quarterly sales and marketing costs were $770k then that would be the numerator. The CAC Payback Period would equal an encouraging 11 months. For SMB customers with higher churn rates and thus shorter monetization windows, CAC Payback Periods of 6-18 months are typically needed, whereas enterprise businesses with high upsells and long retention periods may be able to subsidize payback periods of 24-36 months in some cases. A CAC Payback Period of 36+ months is typically a cause for concern and suggests you may want to slam on the brakes until you can improve sales efficiency, whereas a Pay- back Period of under 6 months means you should invest more money immediately and step on the gas (and please call Bessemer immediately because we want to fund you!) as your customers are likely very profitable within the first year. It’s also worth noting that many companies also chose to track a dollar based average CAC per customer, yielding con- clusions such as “last quarter our Customer Acquisition Cost was $12,500 in the SMB space and $45,000 in the mid market.” The CAC Payback Period and customer averages typically work very well for growth stage businesses, but for earlier stage businesses you likely also need to use some rules of thumb around sales rep performance because it can be very difficult to isolate the financial impact when including executive team involvement and early friendly customers. Bessemer Venture Partners 15

Freemium or heavy “land and expand” businesses may also find that just measuring the CAC Payback of their initial deals understates the leverage in their model. In those cases you may choose to do a blended CAC and include upsell MRR as well as account management costs in the formula, or focus on the medium-term value of the customer contracts rather than just the initial subscription amount, by including all CMRR growth and all sales and marketing costs including account management. Years ago, Bessemer was fortunate to invest behind Mark Leslie at Veritas, and as a result our firm became big believers in the Sales Learning Curve (SLC), a concept Mark helped pioneer. The core concept is that software or- ganizations often fail because they staff up their sales efforts too quickly, before the sales model has been refined. This concept is even more critical for cloud businesses, given the large upfront investment required to acquire customers. Ramping up too quickly will burn precious cash reserve and may fool the product team into missing some critical elements of the real product market fit that will be important to go big over time. This typically means you should hire sales reps slowly up front, only focus on your core geography until your business starts to scale considerably, and separate your “hunters” and “farmers” as you start to ramp. For an enterprise-oriented direct sales business, it typically takes at least $300,000 CMRR to climb the Sales Learn- ing Curve. You should tune your model before you scale, which typically means stopping at three field sales reps until you hit at least $300,000 CMRR or at least two of your reps are making their ~$100,000 CMRR quotas. For companies with hybrid inside sales models or freemium offerings, the quotas can be lower but the attainment hurdles are similar. You know you can profit- ably scale sales when a couple of sales reps are at an annualized run rate to sign annual contract values (CMRR x 12 months) equal to twice their fully-burdened cost of sales. In this case, fully-burdened is not just the salary, bonus, and benefits of the sales rep, but also allocations for sales engineering support, executive support, marketing expense, and professional service expenses associated with securing the customer. For a direct, enterprise sales business model, these thresholds are likely to be around $80,000-100,000 CMRR (approx. $1-1.2M annualized), and for tele-sales models, this Russell Glass may scale down to $60,000-75,000 MRR ($720,000-900,000 annualized). It is usually CEO, Bizo time to accelerate sales hiring when at least two out of three sales reps are hitting quotas at these numbers, and the business has achieved some scale to suggest that the processes “As an online B2B marketing are repeatable- at least $300,000 of CMRR. The “repeatable” aspect is critical: too often business, sales and marketing companies scale their sales forces aggressively after their first senior rep is getting traction efficiency is our lifeblood. in the market and then quickly realize that the new hires struggle to sign their first deal We have to be able to justify the ROI on each incremental because they don’t have three VP’s and the CEO alongside them. marketing dollar our custom- ers spend with us, so the You should also separate your “hunters” and “farmers” and pay them all on CMRR performance metrics are very growth. As soon as you have climbed the Sales Learning Curve and have a sizeable important. We have also managed to run the business customer base, you should supplement your sales force with renewal-oriented account near cashflow breakeven managers. When a cloud company starts to hit the sales inflection point, it is important since the very early days, to keep the new business reps (the “hunters”) busy with finding new deals, while a team so every new sales hire is of account managers (the “farmers”) tends to the established customers. The new ac- expected to return a multiple back to the business.” count team should be paid on new CMRR with a standard deal structure (such as a one Bessemer Venture Partners 16

year deal, with quarterly pre-payments), and incentives for more favorable cash flow terms (such as multi-year pre- payments). You should think of account management as a sales function, and that group should be compensated in a similar fashion – but modeled on your CMRR and churn assumptions instead of a new CMRR sales quota. You’ll typically find that the compensation plans are weighted more heavily to base (~75-80%) with less bonus upside, whereas standard “hunter” compensation plans are typically 50% base and 50% commission at quota. To create a repeatable sales process, focus is also critical, which typically means that you should restrict yourself to your core market. If you’re based in North America with a direct sales model (in person and/or tele-sales), prove your business in North America first. Channels are hard in Cloud Computing in general, and if you can’t sell it yourself then it’s unlikely others will be able to sell it more effectively for you. Only after reaching $1M in CMRR should you invest heavily in channels or international expansion. If you’re a freemium business or enjoy viral adoption more broadly, you may want to consider regional marketing and/or support resources in other regions earlier in the life of your business, but avoid becoming over extended before nailing your home markets. You can think of this as a “bowl- ing pin” strategy on a geographic basis, or good military strategy by avoiding concurrent wars on multiple fronts; it’s also good business strategy for Cloud Computing. Finally, do not confuse any of this with a message to build your business slowly or to under-invest in sales or market- ing. In fact, it’s quite the opposite. The 5 C’s of Cloud Finance and the Sales Learning Curve are tools to help you know when and how to invest aggressively to maximize the business’ long term value creation with the least dilution for you as the executive team. If the metrics are strong, you will be able to finance the business at very attractive terms. You will actually be destroying value if you don’t invest behind success when the ROI is strong. #7 Make online sales and marketing a core competency You’re a cloud business, so by definition, your sales prospects are all online. Savvy online sales and marketing is a core competency (sometimes the only one) of every successful cloud business. Amir Ashkenazi Numerous studies show that your customers are now doing most of their primary re- CEO, Adap.tv search online, and this should not surprise you. As a consumer, you wouldn’t imagine “We’re still in the very early days buying a car, making an offer on a home, planning a vacation, or completing other large of the online advertising revolution purchases without doing some research online. The same is now true for executives at and the flood of dollars continue your target customers. You should therefore be aggressive in marketing to them online. to move to online display, video, mobile, and social at rapid rates. Your prospects are online, so your This is a clear example where business-to-business (B2B) marketers need to learn from marketing and sales efforts should their business-to-consumer (B2C) counterparts. The most innovative B2C companies be as well.” Bessemer Venture Partners 17

are lead generation machines, leveraging social media marketing, search engine opti- mization (SEO), viral marketing, search engine marketing (SEM), email marketing, and other technically-advanced methods. Yet many B2B companies don’t have a clue. The incumbent technology leaders like IBM, Oracle, and SAP have done very little in online marketing, and thus have given their smaller challengers a huge opportunity. Whether they use an automated product like Eloqua or a team of marketing analysts and spreadsheets, online marketing and demand generation is a “must have” for cloud companies. In this new era, the creative elements of marketing are becoming secondary and Joe Payne quant jocks and analytical wizards are starting to take over the CMO and VP Mar- CEO, Eloqua keting positions. At the marketing executive’s fingertips should be detailed reports (NASDAQ:ELOQ) showing pipeline sources, costs per lead, funnel conversion rates by stage, costs per “More and more of the mod- acquisition by source and campaign, effectiveness by channel, and so on. If you are ern buying process the CEO or a Board member, you should review these reports closely and make them is happening before an ac- the basis for assessing marketing effectiveness and performance. tual conversation with a sales rep ever takes place. Modern marketing is increasingly The most advanced marketing executives are also starting to embrace social media about Revenue Performance and to do multimodal attribution analyses. Customer and prospect conversations are Management, and reading no longer defined by website text, email messages, and sales discussions. Twitter, Fa- the digital body language of cebook, LinkedIn, Pinterest, blogs, and dozens of other social media tools constantly your prospects, often before they even reveal themselves.” facilitate discussions about your market, your competitors, and likely your company and product. You need to get into position to monitor and help define these discus- sions, and the savviest marketing teams will use this to their considerable advantage. A strong head of online marketing will also be able to give you detailed demand generation forecasts based on different budget levels. It’s important to understand the natural limits of organic traffic as well as the slope of the supply curve for each of your various paid lead sources. Your blended cost per lead may be very attractive, but if a large portion is organic/free traffic and your marginal cost of an incremental paid lead is quite high, then you may not be able to scale marketing spend in an efficient manner. If SEM is a lead source, you should study the quantity and pricing of your main keywords and do burst Jeff Zwelling testing to validate the assumptions before dramatically CEO, Convertro increasing budgets. If these data are good, highlight them to your prospective investors. We love to find businesses “The Cloud Computing world is just starting to apply the with 6 month CAC paybacks and the capacity to absorb lessons learned from the consumer internet market and really 10x more marketing spending. If it’s bad, you don’t need change the game through aggressive online lead genera- to highlight it – but you better understand those limits tion. Proper channel attribution is critical to optimizing the right mix for your online marketing spend.” and how you can address them over time. Bessemer Venture Partners 18

On the softer side of marketing, it’s grade school all over again and you want your company to run with the cool kids. Ecosystem matters now more than ever, and the right partnerships can really elevate your company and your brand through positive associations. Try to align yourself with the best-of-breed leader in each of the adjacent segments to you and partner with them, host hackathons together, use each others’ products, and cross promote as much as possible. However, don’t overthink the relationships and try to force channel deals too early in the life of your company. Informal co-selling and referral deals are much less effort and typically result in equal effective- ness in the early stages. Therefore, it is still the case that most cloud businesses have to be comfortable with the fact that they will live or die by their ability to sell directly. #8 The most important part of Software-as-a-Service isn’t “Software” it’s “Service” The only acceptable reason to lose a customer is death (bankruptcy) or marriage (acquisition). Every cloud com- pany is in the service business, and therefore your customer service can be the difference between failure (churn) and huge success via high retention and upsells. Reliable delivery of your core ser- vice is a necessary but not sufficient condition for delighted customers. How your handle yourself in challenging or unique situations will truly define how they view you as a partner and a vendor. On the positive side, when they experience a massive business spike are you able to handle it flawlessly and without interruption or hiccups? When your customers launch new products or enjoy huge spikes in demand, are you able to stay at least one step ahead of them at all points? Unfortunately, there are also disaster scenarios to watch out for, such as system bottlenecks that weren’t properly load tested. There are painful examples where a customer received national TV coverage and experienced a flood of Todd Davis customer interest, only to have one key component of the customer facing system col- CEO, LifeLock lapse under the load. Even business processes can break down, as we have seen with (NYSE:LOCK) fatal examples of cloud companies actually intentionally capping the usage of their customers during these periods of hyper growth for fear of fraud, payables exposure, “Our business is all about or system abuse. Simple communication can usually quickly solve these problems, but trust and protecting your account management and support processes without proper escalation protocols can personal identity. We often amplify these breakdowns, and result in immediate account churn and vocal ex- proactively monitor our member’s identities. If you customer detractors in the market. are a LifeLock member and your identity is ever stolen, There are also service events that are significant negative events for you and your we are there to help and customers, but can be turned into positives if handled correctly. Outages are the most provide the support you need for your recovery.” notable, but bumpy product upgrades, faulty product configurations, and weak Bessemer Venture Partners 19

integrations can all create similar issues. Amazon Web Services (AWS) is a fantastic IaaS platform, but has ex- perienced several very public outages in its availability zones with very direct impact on customers. When some of the AWS availability zones failed for four days in April 2011, hundreds of very visible web com- panies ranging from Reddit to The New York Times were taken down entirely. Yet many AWS clients on these same availability zones – including Netflix, Twilio, Bizo, and SmugMug – were relatively unaffected because they had Jim Franklin designed around the AWS architectural shortcomings CEO, SendGrid with added controls in their own products. “One of the things we do right is offer 24/7 support Hundreds of articles have now been written on this single -- even if you’re not a SendGrid customer. We just feel it’s outage and the different ways companies handled the is- the best way to help companies in a bind. Whether their email infrastructure is falling over or they are struggling with sue (just google “high scalability aws outage big list”), and deliverabiity, we are here to help without obligation. The the important thing is to realize that each crisis is also an level of trust this creates is incredible and has been a key opportunity. Several sites that were taken down by the ingredient of our success.” outage appeared frozen and confused, and provided almost no communication or transparency back to their users. As a result they endured thousands of negative emails, tweets, and comments during the painful down days, and lost significant customer loyalty in the process. In contrast, the most advanced of the sites that were down during this period became very transparent with their customers/users about the issues and timelines, and in many cases, actually created workarounds that allowed them to bring their sites and systems back online in other AWS zones or on other platforms, long before AWS themselves had fixed the issues. Even more impressive, however, were the sites like Twilio that had anticipated the issues and stayed up, and then went on to publish a blog post and take press calls to share their learnings and explain the best practices to others. This not only helped to build significant credibility with their enterprise customers, but also helped attract even more smart de- velopers who were like- minded. Salesforce.com responded to their own outage issues with an entire customer portal on www.trust.salesforce.com that is a model for late stage companies to follow, and simple, transparent and honest com- munication with your customers in the early stages is enough to work magic. It’s common knowledge in the cloud world, but still important to mention that these “silver lining” opportunities seldom exist in the case of data loss or leakage. It is essential that cloud companies treat backup, disaster recovery, and security as a priority. System or security failures that result in a loss of customer data, will very often be fatal to the relationship and in extreme cases, to your company. How you handle these business challenges can either further endear you to your customers or send them running into the arms of your competitors. In terms of business and team management, as the “service” functions of your company continue to grow, you will likely want to (need to) separate client success, account management, and professional services. Bessemer Venture Partners 20

Client success is the front line of customer support, and should be focused on resolving customer issues and driv- ing broad adoption and happy users. Account management is tasked with managing the ongoing sales part of the relationship, including renewals, upsells, and expansion. These are the “farmers” in your organization who are sales executives at their core, but longer term relationship thinkers who will invest in the long term goals and objectives of your customers over time and help them use your products to get there. Professional services groups are getting smaller in modern cloud companies – and even non-existent in some – but can still play a vital role in customer onboarding, configurations, and integrations when needed. The goals and objectives of these three func- tions are very different, and the skill sets of the top performers in these functions are typically very different, so it is likely a cloud company will want to split these functions out into 2 or 3 different organizational groups as you scale from 50 to 100+ people and can afford to hire specialists. Each of these functions should work with the technical team to enable proactive monitoring of the product for likely churn or upsell opportunities. You can easily check who logs into your product, how often, what they do inside the product, and what results they achieved. So now you need to track the key usage metrics and measures, and create internal dashboards to know which customers are getting the most value (potential upsell candidates) and which are likely to churn (time to intervene proactively). As you get more insight into the issues, you may want to consider techniques like expanded online training or even unlimited subscription-based training (as many leading SaaS com- panies are now doing) to drive adoption and awareness. Cloud businesses can also learn from their consumer internet peers, by taking advantage of their web application architecture to analyze detailed customer usage data, use a/b test variations, iterate on small details of a page or a feature, and evolve the product each and every day. Given the growth of Employee Software and the fact that many customer orders are starting very small and growing very large over time, you may find that upsells from account management become more critical to your long term busi- ness model than the initial sale itself. If, after all your efforts to satisfy the customer, you find that you still face a churn situation and they want to transition off your product, recognize that it doesn’t need to be binary. Some of our most creative SaaS portfolio companies have started offering archiving services for their customers where they offer a very limited use (typically read only), single user license at ~20% of the prior MRR. This offering has many positive results often including ongoing high margin revenue for the company, a slower migration plan due to the lack of a hard cutover date, an easier migration path back if the customer ultimately regrets their decision to churn, and higher satisfaction among former customers as well as current customers. Bessemer Venture Partners 21

#9 Culture is key as you build your dream team Bessemer has enjoyed the privilege of backing hundreds of companies throughout our rich history, and the one single determinant of success above all else is the quality of the team and how well they work together. This not only applies at the early stages when we may first get to work with a founding team, but also to the extended team that they hire as the company grows. Success as a private company is all about navigating evolving markets and complex partner, customer, and competitive dynamics. Therefore we cannot emphasize enough how important it is for you to hire superstar talent at every level and for every position. With this in mind, it is also encouraged to reject candidates for the reason of “not a cultural fit.” Oracle, Google, and Apple are all great companies, but with very different cultures. They all enjoyed success due to their ability to hire specific employees that could be suc- cessful in those organizations. Not surprisingly, we often hear our companies talk specifically about the compa- nies they want to target for candidates based on having similar corporate cultures, and ones they intend to avoid. In terms of the other elements of talent management, three personality traits that we see our best CEO’s repeatedly screen for are: 1) a clear pattern of success 2) a will to win and 3) self critical and accepting of failure. In terms of success, it often doesn’t matter as much where the candidate is coming from previously as that they have been a part of success. This often includes the obvious (Salesforce.com, VMWare, etc.) but increasingly includes success in other industries or fields such as Google, NASA, Cal, sports, or the military. The second element is all about personality and drive. Startups require a special breed of talent that wants to tackle the hardest problems with very limited resources, and have fun in the process. It’s a “run through walls,” “take the hills,” and “will to win” drive that is tough to teach and tough to fake, but is magic when present in talented employees. The third element is humility to admit mistakes, accept input, and change direction, because startup executives will face failure many times along Keith Krach the path to success. CEO, DocuSign Talent management isn’t just at the point of hire, but is ongoing. Most people agree “Building a world class that “A players” hire “A players,” and 5 excellent engineers can beat a really good high performance team is team of 50. Yet many still compromise. We let that marginal person stick around, my favorite part of being a CEO – and my most because we figure having someone in the job is better than having no one. Most of important role as a leader. us have been tired of the long search so we decide to settle for the best candidate The company with the we’ve met, or let an exec hide an under-performer on his or her team because he’s best people wins. Creat- liked by everyone and tries hard. Read Jack Welch’s book, look at the data from ing a lasting, high-growth company requires that we Cornerstone OnDemand, or talk to a CEO who has ever had to do a layoff, and surround ourselves with they’ll all tell you that firing the bottom 10% of the company will actually make people better than us, align you stronger and increase your output. Tech markets are very efficient with talent our team to an inspiring and small companies don’t allow really weak performers to hide for long, but the vision, and constantly raise the standard so we keep same principles apply. getting better and better.” Bessemer Venture Partners 22

Similarly, everyone involved with the company is part of the extended team. You should approach adding an advisor, a BOD member, or an investor with the same rigor that you do other members of your core executive staff. Ask questions and check references to make sure the person is going to bring real expertise to the dis- cussion, relationships and perspective that are additive, and a style that is a strong fit with the team. Of course economics are always a part of these decisions as well, but focus on the person first and then making sure it’s a “fair” deal. You aren’t going to build a dominant company by selecting second tier executives and advisors in order Anne Berkowitch to save money on their compensation packages, or by selecting second tier investors CEO, Select Minds for a little less dilution. “Team building is a critical part of For young and/or first time CEO’s, one of the most powerful things you can do to add our company, as well as our prod- experience to the management team is to mix in experienced external hires with young uct. Culture is central to recruiting and aggressive startup operators. The extreme examples of this are 20-something - particularly in driving referrals founders like Mark Zuckerberg at Facebook and Aaron Levie at Box who have both - and can even be powerful when looking to rehire past employees done a fantastic job of hiring extremely experienced executives while still preserving from your alumni network.” the energy and culture of a hot startup. Team is the single most important factor in determining whether or not we want to invest in a company, because our history shows that it’s the single most important determinant of company success. Bessemer Venture Partners 23

# Cash is (still) king – Cloudonomics requires that you 10 focus on cash flow above operating profits, and plan your fuel stops very carefully The cash flow characteristics of a cloud business are wonderful in the long term, but can be lousy in the short term. Cloud companies require you to fund research, development, sales and marketing up front in return for a multi-year stream of revenue. This typically demands enough invest- ment capital (over stages) to fund 4+ years of runway before a company can achieve positive cash flow (GAAP profit is even longer). Imagine you are flying a private plane from Silicon Valley to Wall Street (which sometimes is the figurative or literal goal), and you need to stop a couple of times for fuel (investment capital) for the trip. It is critically important Greg Becker that you plan your equity and debt financing events in advance to maxi- CEO, Silicon Valley Bank mize value and minimize dilution. “We believe we’re still in the early Understanding the cash flows of your business – including Gross and Net days of the Cloud Computing revolu- Burn Rate – is critical to survival in the early days and critical to your tion, and are eager to bank the future dominance in the long term. There have been many promising cloud leaders of this next wave. We are big startups that stepped on the gas too early and were wiped out as a result. supporters of Bessemer’s cloud metrics and laws, and have actually developed Always model the business with a comfortable cash cushion and recog- specialized lending products designed nize that most cloud businesses paradoxically consume more short-term around them. We have already loaned cash as growth accelerates. As a business, it is critical to weigh forward tens of millions of dollars to Bessemer investments carefully. Cloud businesses typically require multiple rounds companies behind these metrics, often in cases where other banks couldn’t be of investment and a good amount of capital. For example, it took $126M competitive because they still focused for NetSuite to go public, $61M for Salesforce.com, $41M for Eloqua, and on legacy software metrics.” $45M for Cornerstone OnDemand. We are now seeing a second generation of cloud businesses that have the potential to be more efficient than many of their predecessors by leveraging Platform-as-a-Service (PaaS) and Infrastructure-as-a-Service (IaaS) to out- source more and shift many startup costs to variable models. However, even in most of these cases, significant Steven Kokinos CEO, ThinkingPhones “We were able to grow ThinkingPhones into a profitable business off of some very modest friends and family capital years ago. However, our CAC and CLTV metrics were good and our market is massive, so we decided to raise money from Bessemer to re- ally lean into the business to seize the full opportunity.” Bessemer Venture Partners 24

capital will still be required to build a dominant cloud business. If you plan these stages thoughtfully, you will be able to minimize dilution by progressively decreas- ing your cost of capital, mixing seed capital with venture capital and possibly debt before attracting public market investors. With the increased sophistication of debt and equity providers in the cloud space, we are now seeing multiples of CMRR (and similar subscription measures) as the primary valuation metric. It is worth noting that many businesses are not natural candidates for venture capital and should look to other sources of capital for early funding. Friends, family mem- bers, former co-workers, and strategic partners are all time-tested pools of friendly startup capital. Venture capital firms can be extremely valuable as partners in your Adam Miller business, but are really only a fit for opportunities where the team is “swinging big” CEO, Cornerstone and have the potential to be billion dollar businesses at some point. OnDemand (NASDAQ:CSOD) Although we place a great emphasis on the risk of underestimating the cash cushion needed to build a large cloud business, another great risk is that of underinvestment “In the early days, our in the business. Trust your metrics and your dashboard and invest behind success. company was literally financed on our personal If your cloud metrics show strength and your unit economics are meaningfully credit cards so I tracked positive, then it would actually be irresponsible to not invest aggressively in growth. our cash position on a The Cloud Computing market is no longer a secret and the competitive dynamics daily basis. Bringing on continue to heat up quickly. If customers love your product and the financials make Bessemer as an investor allowed us to then ac- sense – it’s time to run after the opportunity aggressively to grab the full market po- celerate our business and tential, or someone else will! take it to the next level. But even to this day as a If this sounds like your business, please reach out to us at Bessemer as we’d love to be $1B+ public company, old habits die hard; cash your partner and join you for the path ahead. is still king.” Bessemer Venture Partners 25

BONUS You can ignore one or two of these rules, but no more. Great companies innovate, but pick your battles! You might be reading this and saying to yourself, “But wait – we’ve got a great channel partner that is going to take us to the moon!” or “I took an early chance in Europe and it’s now driving the majority of growth for my entire business.” To which we would say “good for you.” Nothing is ab- solute, and we certainly believe it is possible to ignore one or two of these core tenets and still succeed. In fact, several of the companies we have worked with have also chosen to break one of these laws at some point in their lifecycles with success. Larry Ellison* If you find yourself questioning several of these 10 Laws however, it’s CEO, Oracle (NASDAQ:ORCL) probably time to step back and take a hard look at your business. As for- mer Cloud CEOs and investors ourselves, we have learned the hard way “I thought the cloud was just water that much of the battle is just learning from the mistakes of those who vapor until I read Bessemer’s 10 Laws went before us. In our analysis of hundreds of cloud businesses, we en- of Cloud Computing. I now realize that I may have been insensitive and a bit countered several successful companies that were on the borderline with closed-minded in my criticisms. Armed one or two of these laws, but none that successfully challenged several with this secret playbook, I can now of them. destroy SAP’s entire legacy software business with a single press of a button from my island fortress in Hawaii, while We hope that you can benefit from some of these best practices we’ve still surfing Pinterest with my left hand.” learned through the years and these “laws” can help you run your cloud *[Ok, we made that one up. But he business more effectively. If you have thoughts, edits, or additions, please might be thinking it!] send them over to [email protected] as we always welcome new input! Bessemer Venture Partners 26

The Ask If you found the content in this document valuable, then please support the companies that helped contribute the insights behind this analysis: Adap.tv - www.adap.tv Online video advertising network and exchange Apperian - www.apperian.com Helping enterprises build, deploy and manage internal mobile applications BillGuard – www.billguard.com A personal finance security service powered by collective consumer vigilance Bizo – www.bizo.com Precise online ad targeting based on 100 million business professionals globally Box – www.box.net Access and share content from anywhere Clearslide – www.clearside.com Sales communications platform for presentations, emails and in-person meetings Convertro – www.convetro.com Advanced tracking and attribution to optimize display, optimize yield, optimize profit Cornerstone OnDemand – www.cornerstoneondemand.com Comprehensive solution for learning, performance, compensation, succession & compliance management Criteo – www.criteo.com World leader in personalized retargeting, making display advertising perform like search Docusign – www.docusign.com World’s largest electronic signature platform Eloqua – www.eloqua.com Complete revenue performance management platform GFI Teamviewer – www.gfi.com Web security and productivity for the SMB including remote connection, support, and security products Bessemer Venture Partners 27

InsightSquared – www.insightsquared.com Data intelligence for small and medium sized businesses Intacct – www.intacct.com Professional-grade accounting, business and financial management applications delivered in the cloud Intego – www.intego.com Don’t let viruses mess with your Mac. Intego makes software products that protect Macintosh computers Liazon – www.liazon.com An online private benefit exchange platform for small and medium sized businesses LifeLock –www.lifelock.com An industry leader in identity theft protection LinkedIn – www.linkedin.com World’s largest professional network on the internet Metalogix – www.metalogix.com The leading provider of content lifecycle management solutions for Microsoft SharePoint and Exchange Mindbody – www.mindbody.com Business management software for the wellness industry including fitness, yoga, salons and spas Onestop – www.onestop.com Full-service, e-commerce technology and distribution partner Parallels – www.parallels.com Worldwide leader in virtualization and automation software Perimeter eSecurity– www.perimeterusa.com Comprehensive, on-demand messaging and security services to thousands of businesses Pinterest – www.pinterest.com A virtual pinboard that lets people organize and share all the things that inspire them Retail Solutions - www.retailsolutions.com Turn downstream data like point-of-sale and customer loyalty into actionable visibility SelectMinds – www.selectminds.com Social recruiting automation Bessemer Venture Partners 28

SendGrid – www.sendgrid.com Reliable email delivery, scalability and analytics; relieves businesses of maintaining email infrastructure Shopify – www.shopify.com Easy to use online storebuilder for SMBs Soluto – www.soluto.com A revolutionary approach to improving PC performance to help consumers get the most out of their PC Twilio – www.twilio.com Communications Platform-as-a-Service for scalable Voice, VoIP and SMS apps in the cloud Wix – www.wix.com Web based service for building stunning websites Bessemer Venture Partners 29

Herzliya, Israel Larchmont, NY Menlo Park,CA Mumbai, India Cambridge, MA