Trident Capital Blog

Cloudnotes

I’ve been meaning to post these notes a couple of weeks ago but extensive travel and due diligence on two companies kept me otherwise occupied.  Let’s start with a positive statement: The future for SaaS applications is getting brighter.  Discussions I had with the CIOs and business unit executives of the companies I recently visited support that the majority of applications used today by mid- and large-size companies can be delivered via the cloud.  They use the following rationale to arrive to this conclusion: Of the applications used in a corporation, 15% are either company-specific, reflecting proprietary business processes, or need to adhere to particular regulatory requirements.  As a result, these must be on-premise applications.  The remaining 85% of business processes can be automated using off-the-shelf, packaged applications.  Of these packaged applications, 80% can be delivered through SaaS, i.e., they are based on single instance multi-tenant architectures, or, at the very least, they can be cloud-based applications delivered over private, public or hybrid clouds.  The type of cloud used depends on the customer’s customization and on-boarding requirements.  The remaining 20% must be delivered using off-the-shelf on-premise applications.  Getting this message across will require IT vendors to continue educating the market and offering their SaaS applications in pieces that are easily “consumable” by corporate users.  Issues such as security, and application and data integration, about which I wrote in the past, will still need to be addressed before cloud computing in general and SaaS applications in particular can be broadly adopted.

Let me now switch to last week’s Google announcement on the launch of its Application Marketplace and why I thought it is worth commenting about. 

1. The launch of the Marketplace demonstrates Google’s continued interest in business applications (the Marketplace launched with applications from the 50 vendors, such as Concur’s and Appirio’s).  With the App Marketplace Google has the opportunity to innovate and clearly differentiate itself from Salesforce, whose AppExchange is a Marketplace competitor, if it can incorporate lessons from its Android App Store.  For example, Google can provide additional services to the marketplace participants such as billing, analytics and potentially even an alternative business model by trading annual license fees for ads displayed inside the applications.  I have been writing about the upward trend in the Customer Acquisition Cost (CAC) of SaaS applications.  These marketplaces have the opportunity to become low-friction sales channels and better rein in such costs.
2. Participation in the Marketplace will provide an additional forcing function for SaaS application and data integration, since, at the very least, the applications must be integrated with Google Apps.
3. Now that Microsoft is becoming serious about cloud computing, I expect that it will start thinking seriously about establishing a similar cloud-based sales channel for its partners. Competition among these marketplaces, including AppExchange, will be good for the dissemination of SaaS applications and in general and will improve the sales economics for such applications.

The final note stemmed from the recent acquisition of 3tera, a company I considered investing over a year ago, by CA for a rumored $100M price (approximately 30x trailing revenues).  The acquisition made me think whether Trident should re-consider its position and try to invest in PaaS/IaaS companies.  We have stayed away from this part of the cloud computing stack because we believe that

1. Private PaaS and IaaS companies are not capital efficient as will require significant investment to remain competitive with large IT vendors introducing cloud-based solutions.  For example, it is reported that Amazon has invested hundreds of millions around its EC2 IaaS service so that it can offer processing at a fraction of ¢1 per hour per server).
2. Enterprises considering moving any part of their IT infrastructure to the cloud would use a large IT vendor rather than a small private company.  Two recent presentations delivered by Microsoft and IBM executives reinforced these thoughts that PaaS and IaaS may difficult areas for venture investments.  For example, IBM recently surveyed 1100 of its customers regarding their cloud computing plans.  The survey showed that most of these customers still don’t understand where and how to apply cloud computing.  Furthermore, 60% of the surveyed companies stated that they will use large IT vendors to move workloads to private clouds but are more hesitant about using public clouds.
3. While CA bought 3tera for an eye-popping multiple, I doubt other IT vendors will follow suit and be willing to pay similar multiples for cloud computing companies.  CA bought 3tera in order to acquire its technology and, most importantly, in order to access 3tera’s cloud computing expertise.  CA needs to catch up quickly in cloud computing.  Telcos, particularly European ones, and Indian IT outsourcing companies may also need to acquire cloud computing expertise but these types of companies are not known for being aggressive IT acquirers.

Criteria for Determining a Company’s SaaSyness

On a couple of separate occasions last week I was asked to write about my list of characteristics that define true SaaS application companies.  Establishing these characteristics is becoming important now that there is so much interest in cloud computing and is getting easier to confuse SaaS applications with cloud-based applications.  Coincidentally, there are a couple of recent posts and active discussions on this topic taking place in different forums (here, here, here and here).

As you can imagine, I spend a considerable amount of time every week with our SaaS portfolio companies, as well as talking to the management teams of SaaS companies seeking investment from our firm.  The continued interest of venture investors in SaaS, and more recently in cloud computing, is resulting in management teams trying to position every company that uses the cloud as a SaaS company.

Let me start with a definition of a cloud-based application so that we may be able to better distinguish it from a SaaS application.  Any application that can be deployed in a cloud (public, private, or hybrid) is a cloud-based application.  Let’s consider a couple of cloud-based application examples.  A company like Goldman Sachs using its own application developers can create a proprietary application and deploy it in its private cloud (that resides in its own data centers). Alternatively, Goldman may contract an IT services provider, like IBM, to develop a proprietary application and deploy it in the private cloud that the provider, in this case IBM, runs for Goldman. According to the above definition, these two are examples of single tenant cloud-based applications.  But they are not SaaS applications.

A true SaaS application company has the following characteristics:

1. Product:
1. Offers an application using a single instance multi-tenant architecture
2. Simultaneously runs at most three versions of the application (the current version and the two previous versions) in its data center.
3. The application supports cloud-specific functionality such as social networking among users, rich collaboration among users, presence, and activity streams.
4. Uses processes and the associated software tools for data and application integration so that data from other cloud-based and/or on-premise applications can be moved to the SaaS application, and the SaaS application can communicate with other relevant applications.

1 (a) and (b) are important because they impact the company’s costs of doing business and ultimately its margins.  While not necessary, 1 (c) is very important because it allows true SaaS companies to offer value and further differentiate themselves from their on-premise counterparts (some of which are aggressively moving their applications to the cloud).  Finally, 1 (d) is important since no application is an island but needs to be integratable into each customer’s application fabric.

2. Business Model: The software license model is subscription-based.  Furthermore, no more than 30% of the company’s revenue is coming from professional services.
3. Services: Application customization only through configuration changes rather than programming.  Example configuration changes include the simple activation (or deactivation) of certain application modules or menu-driven changes to the application's look and feel.  Application onboarding should be accomplishable through the web and not require extensive on-site presence.
4. Sales: Extensive use of self-service processes.  I have written more extensively about this topic here and here
5. Support: Extensive use of self-service processes.  I have written more extensively about this topic here.
6. Marketing: Broad use of the web for lead generation, and market reach.
7. Operations: The performance of every aspect of the business such as customer acquisition costs, lifetime value, cost per lead, uptime, etc. must be measured constantly. Relevant posts on metrics can be found here and here.

During the first meeting or call with the management team of a prospective SaaS investment we want to make sure that the company possesses the above characteristics before we can move into further due diligence.  Finally, as you can already figure out, while all SaaS applications are cloud-based applications, the reverse is not true.

Trident Capital Internet Portfolio Company Summit

On Friday my partners and I hosted in our office in Palo Alto a meeting with the executives of our online advertising and ecommerce companies.  Executives from Syndero, HomeAway, Zeo, AccountNow, Advanced Payment Solutions, Turn, Sojern and Tellapal participated in the meeting.  The purpose of the daylong meeting was to enable the participants to exchange information and best practices for customer acquisition and improving lifetime value (LTV).  Below are some of the major points that were made during the day:

1. Being best in class on customer acquisition implies combining the right people, an efficient process, and best in class tools.  The process typically being used involves: creating a control group, hypothesizing a particular behavior of this group that will be capitalized during the marketing campaign, testing the hypothesis on the desired customer touch points (email, search, display, affiliates, etc.), analyzing the test results, learning from the analysis and optimizing, forming a new hypothesis, repeating.  As one of the participating CEOs mentioned “every minute, some aspect of a customer acquisition campaign is being optimized.”
2. Social media is becoming an increasingly important channel for generating interest in a category but once a company starts promoting their particular brand through the social channel, user engagement appears to drop.  Word of mouth is important but it has not yet been integrated effectively with social media.  This is an area that holds great promise and where the participating companies will be investing over the next 12 months.
3. Test and measure every parameter that goes into customer acquisition.  Best in class companies can manage very complex test matrices.  It is important to be able to quickly hypothesize and test offers, creative, etc.
4. The right analytics around the data being collected define best in class companies with the lowest customer acquisition and retention numbers.  Our portfolio companies, by constantly sharing best practices, have moved beyond measuring the effectiveness of customer acquisition campaigns in terms of clicks, and impressions, i.e., website analytics.  Instead they are now measuring the impact of each customer interaction on a particular outcome. 
5. While companies are investing heavily on improving the quality and the uses of the data they capture through their customer interactions, they were less definitive about the impact of data they purchase from data exchanges.
6. A good off-the-shelf software platform for direct response/branded response doesn’t exist (investment opportunity?).  The portfolio companies have created such platforms by manually integrating stand-alone applications.  As a result, the operational complexity of working with multiple applications rather than a single platform presents a big challenge for marketers.  Good direct response software platforms can be used not only to improve the effectiveness of customer acquisition and customer retention of existing products, but they can also be used to effectively test the likely success of new products before introducing them to the market. 
7. Success with online channels is leading our ecommerce companies to start experimenting with offline channels such as Direct Response TV.  Every test associated with an online customer acquisition campaign can become a leading indicating for the effectiveness of a corresponding offline campaign.
8. The companies are being proactive about data privacy issues.  They are using services, including legal services, on how to best safeguard and use the data they collect and what to include in the creative they test for customer acquisition.
9. Metrics on how to calculate lifetime value per customer are still evolving.

Not only were the executives able to share best practices and share ideas about how to improve the operations of their respective companies, but after the meeting ended several stayed behind working on business deals between their companies.  A great way to end the week!

Conferences … Conferences

Another busy week attending conferences.  Goldman Sachs held its annual 3-day Technology and Internet conference.  That was following by Pacific Crest’s Emerging Technology Summit that had a SaaS, Mobility and Cleantech tracks. 

Highlights from the Goldman conference:

1. More large software companies are embracing SaaS. I sat through a few of the public company presentations and noted that Adobe, and Intuit are starting to talk more about their SaaS initiatives.  In addition SAP talked about its SaaS initiatives, i.e., the next version of Business by Design that is due next summery and the recent introduction of on-demand BI through Business Objects.
2. Increasing interest for acquisitions.  Several companies mentioned their intent to be active acquirers particularly around software.  There is appetite for both large acquisitions as well as smaller technology-driven acquisitions.
3. Improving IT spending environment.  Overall IT spending in the Global 2000 is stabilizing and there is more confident talk about the PC and server refresh cycle (noted by both Intel and Microsoft).  There is also increased IT spending from China and India whose companies are trying to upgrade their infrastructures to be compete internationally.  Companies are also starting to use services more extensively to catch up with projects they had put on hold during the past 12 months (research reported by Goldman Sachs).  New software initiatives (with the exception of server virtualization) not a top priority during 2010. 
4. Social networking panel.  No significant increase in budgets around social initiatives yet.  Investments are campaign based.  Companies are trying to figure out how to get hold of the collected data and use it for marketing.  The panelists expect significant market consolidation to take place during the next 5 years.  Data privacy is a big concern and the panelists all recommended that the industry starts self-regulating on this issue.  Email continues to be a big channel for driving new users to social networks. 
5. CIO panel.  IT budgets will be moderately up in 2010.  The consensus is around 5% but this will not be uniform.  In some industries the IT budgets may remain flat while others may have higher than 5% increases.  The panelists also indicated that while the larger companies will increase their budgets, the SMB will lag behind.  In terms of spending priorities
1. Infrastructure optimization around areas such as high availability and redundancy.
2. Migration of more business processes online.
3. ERP implementations.
4. Server virtualization.

The CIOs agreed that it will be easier adopting SaaS applications than cloud-based infrastructure.  They are aggressively moving low risk applications to the cloud (both by re-implementing on-premise applications to their on-demand equivalent, and by acquiring packaged SaaS applications).  Use of the cloud to address demand spikes (to address processing needs or storage needs) is of great interest to the CIOs.  Cloud portability of workloads is an issue.

Highlights from the Pacific Crest conference:

1. SaaS survey results.  Pacific Crest regularly conducts a survey on the usage of SaaS applications.
1. Over 50 of the survey participants expect that SaaS usage within their organizations will increase over the next 12 months. 
2. Sentiment on SaaS is improving with close to 40% of those surveyed indicating that they are anywhere from satisfied to very satisfied with the SaaS solutions they are using and are therefore likely to recommend the use of SaaS applications to other organizations.
3. Over 25% of the survey participants indicated that they will spend around 30% of their IT budgets on SaaS applications and that number is projected to grow to over 30% in 2011.
4. Close to 30% of those surveyed indicated that they expect to realize 10-20% cost savings from the use of a SaaS application compared to its on-premise equivalent with the savings being almost equally distributed across hardware, software installation, software maintenance, upgrades and services.
5. Top benefits of SaaS: faster implementation, lower TCO, better system reliability, lower internal IT costs, and better scalability.
2. SaaS BI panel. All companies represented in the panel (Pivotlink, Cloud9, Jaspersoft, and Adaptive Planning) are growing well indicating that SaaS BI is catching on.  Pivotlink and Jaspersoft spoke about how their customers are coming from across the corporate spectrum (Global 2000 and SMBs).  Pervasive use, viral adoption, and collaboration are three areas that clearly distinguish SaaS BI from on-premise BI.  Vendors are starting to focus on creating new paradigms for extracting data from operational systems (cloud-based or behind the firewall) and moving it to the cloud.  They agreed that the traditional ETL paradigm cannot work in SaaS BI.  All panelists agreed that the way to successful adoption of SaaS BI is to offer more solutions and applications rather than just platforms.  Offering platforms will limit them to selling to IT organizations (something which nobody appears to want) with cost and personnel reduction as their primary value propositions (which everybody admitted are tough sells).
3. Benioff’s presentation.  Marc Benioff, CEO of Salesforce.com, gave one of the keynotes. Using Chatter as an example, his company’s latest product, Benioff talked about the next evolution of SaaS applications that will be driven by real-time collaboration and social networking functionality which in turn is expected to result in better communication among employees and higher productivity.  The demo, that was part of his presentation, showed the integration of social applications and social data around Salesforce’s platform.

This was the first SaaS conference I attended where I heard application vendors talking about, or in the case of Salesforce demonstrating, functionality that can only be delivered because these applications are cloud-based.  In other conferences, SaaS was justified as a superior model for delivering software.  Now we are starting to hear about applications whose “cloudness” enables them to have unique functionality making SaaS giving more reasons for customers to be attracted to SaaS.

This week I’ll be attending the Morgan Stanley technology conference and the SMX West conference.