How we got to now in enterprise software, and where we are headed — Part 3/6
The recent past — Systems of Record and productivity with Client Server and distributed systems
The unbundling of the Mainframe in the 70s paved the way for the rise of an independent software industry. With the emergence of the Client-Server architecture model by the 80s, the next era of enterprise computing started to find traction, and thrived for the next 30 years. Enterprise software products of this era could largely be classified into three broad categories:
a) Functional — ERP, CRM, HCM, SCM, etc
b) Horizontal — Spreadsheets, databases, web servers, app servers, systems integration engines, etc
c) Vertical — utilities meter data management systems, retail management systems, electronic health records systems, etc.
Business Model Dynamics:
The dominant business model of this era is the ‘perpetual licenses’. Licenses were mapped to CPUs (and/or cores). Since the license is perpetual, software vendors also mandated 15–20% of the license fee as annual maintenance fee to provide the customer with upgrades and patches. However, most vendors’ terms honor the perpetual only for the existing features. If there was a major new feature or bolt-on feature, it was a separate line item with its own incremental price. This model of licensing was ingenious and it rained money on software companies. Profit margins in the range of 85+% became the norm. Valuations rose. A company with decent cadence and customer base attracted 5x-8x revenue multiples in valuation. Since in software business, the cost of making the product is essentially front-loaded (due to negligible production and distribution cost), once the product is ready for the market, companies invested heavily in sales team to sell as much as they could, and sell as fast as they could. Companies like Microsoft invested aggressively in the partner ecosystem and OEM arrangements as well, to maximize distribution. The steady maintenance revenue acted like a reliable annuity that funded future development and R&D. Once companies found their initial footing, the maintenance revenue annuity helped them in gradually moving up the value chain and in building moat. Life was good! Since selling into an existing customer is easier than finding a new customer, companies invested in helping the customers ‘burn down the licenses’. That is, making the customer consume the licenses rapidly by nudging them with ‘solution selling’ and ‘value selling’. There was also a strong correlation between customers burning down licenses to them coming back and buying more. A healthy license burn down would help the vendors greatly with customer stickiness and loyalty. Customers that has high volume purchases were offered Enterprise License Agreements (ELAs) and Unlimited License Agreements (ULAs). These deals usually came with steep discounts and incentives for customers that worked as a win-won proposition for both parties.
Never the less, a parallel and important phenomenon of this era is the tenacious rise of Open Source Software (OSS) and the ensuing tug-of-war between free and commercial software worlds. Red Hat was one of the early successful pioneers that innovated on business model for open source software, and soon many other companies adopted this strategy. These companies offered commercial add-ons (predominantly OA&M tools), along with support and consulting services for the open source software.
Investment and M&A Dynamics:
The Venture Capital mechanics were fine-tuned and perfected with the semiconductor boom in 60s and 70s. By early 80s, VCs and entrepreneurs formed a healthy symbiotic relationship. In Silicon Valley, the cottage industry of VCs ensured a steady supply of capital available for entrepreneurs to get started. VCs have also been quite proactive in cross-pollination of knowledge and improving the quality of tribal knowledge for building software companies. Once a company found its first set of customers to sell to, the combination of maintenance revenue, plus additional rounds of VC money helped with the fly-wheel effect to kick-in. Once companies got to the range of $80M — $100M in annual revenue, going for IPO was the favorite exit option.
In this era, companies also innovated in various models to attract and retain talent. The concept of RSUs, Stock options, and ESPP have become a mainstay in most companies’ benefits package during this era. Sharing the wealth has been a win-win business strategy.
Most of the M&A activities happened around two kinds of synergies: (a) technology (b) business. For technology synergy, companies acquired other companies for the technology that they could weave into their stack. For example, Oracle bought BEA in 2008 and assimilated WebLogic server (and other products) all across its other product platforms. For business synergy, companies made acquisitions to get a foot hold in new markets, revenue streams and to accumulate new customers. For example, SAP bought Business Objects to augment BI capabilities to its ERP.
A steady supply of capital, healthy exits (M&A and IPO), and an upbeat stock market of 80s and 90s helped to propel the enterprise software market — until the infamous Dot-com Bubble happened in 2000. The dawn of the next era in enterprise software happened in the aftermath of the dot-com bubble.
Another interesting dynamic of this era is that large companies that are flush with cash reserves started venture capital arm. Google was one of the early pioneers of this type of corporate startup incubation. Other companies quickly followed suit — including Intel, Microsoft, Qualcomm, Salesforce, to name a few.
The rise of x86 (Intel) and SPARC (SUN) architectures, operating systems like Unix/Linux — gave rise to this era of client server and distributed software architectures. Strong interplay and partnerships developed between the infrastructure companies (hardware, networking, storage, etc.), and the software companies. Vendors invested in developing industry standards so that interoperability between the stacks increased, making their product more desirable in the market. Standards like TCP/IP, XML, SOAP, WS-*, Kerberos, HL7, and dozens of other key standards paved the way for strong network effects and interplay of various software stacks.
In much of this era, backend applications ran on Unix/Linux/Windows servers and the frontend applications ran on windows terminals (fat clients) or in the web browser (thin clients). Oracle EBS, SAP, Siebel, PeopleSoft, Hyperion, Retek, Infor, and others largely fit into this architecture. Customers leveraged these applications as the core Systems of Record for their businesses. They essentially performed the tasks of information capture, workflow orchestration, and analytics on a database.
The second breed of enterprise software came from vendors that produced databases, web servers, app servers, systems integration engines, security management, etc. These products augmented and helped to maximize the cumulative value of IT systems for the business — along with the systems of record. Service Oriented Architecture became the popular architecture of choice for building distributed systems and their integration. However, SOA is also succumbed to its own success and many poorly executed projects led to many unnecessary side-effects.
The third key breed was the industry vertical products that targeted the industry vertical specific use cases.
In the world of Open Source Software, The Apache Foundation has led the incubation and development of numerous popular open source software products (such as the Apache Web Server). Enthusiastic developers and companies with Open Source based business models continued to contribute to the traction and success of Open Source software phenomenon.
Elsewhere, since the mid-1990s, B2C companies like Amazon, Google, Facebook, Apple innovated rapidly in spoiling the customer with friendly, intuitive products and instant gratification. Knowledge workers that used crummy ERP UI of the day at work were also using (and loving) the mobile apps UX (thank you, Steve Jobs!), blazing fast Google search, Amazon’s suggestions, among other things. This led to the bottom-up pressure of bring your own device (BYOD) culture and the subsequent consumerization of IT by 2011.
The software of this era provided rapid productivity increase and allowed the knowledge workers to be much more effective. Software of this era is said to have improved the macroeconomics metrics as well. The workflows and analytics that the mega software stacks provided empowered companies in substantial ways to optimize business operations, with agility to go after new markets, and with business model innovation.
Companies procured ‘best of breed’ software and hardware of their choice and owned the onus of making it all work together. This substantial undertaking led to the buildup of IT teams and business lines increasingly depending on IT for their day-to-day activity. While this story ran with reasonable success for 20+ years — a few key critical challenges began to rear their head.
a) Companies needed large upfront capital investment in software, hardware, data center, teams, and consulting services. Companies also invested heavily in customizing the ERP software during 80s and 90s — which made future upgrades and changes much harder.
b) As the proliferation of various on-prem software platforms increased, the effort, the complexity, and the cost of maintenance was non-linear. Delivery cycles got longer. Line of Business knowledge workers began to feel underserved by their IT teams and a chasm developed.
c) As the pace of the business increased especially over the last dozen years, complexity of on-prem software model started becoming the Achilles heel for the customers. To keep up, line of business knowledge workers started finding their own means and ways of getting this done — without relying on their IT. This led to the rise of the phenomenon of shadow IT.
During the client server era, companies largely built their initial moat in terms of getting the 1.0 product out of the door quickly (usually with the help VC funds) and then rapidly building effective distribution and sales strength. There was a mad rush for the ‘land grab’ of the market share and to prevent the fast followers from succeeding. As such, marketing teams’ key focus became (a) sales lead generation, (b) product messaging and positioning and (c) brand awareness — usually in that order. Sales teams dovetailed into marketing to make product sales happen, and to sustain the growth curve. Terms like Go-to-market, route-to-market, solution selling, sales plays came the common vernacular in these companies, and rightfully so. As discussed earlier, the maintenance revenues provided a stable recurring stream. The spooling and flywheel effects would kick in, increasing the moat for the company over time. Oracle and SAP are classic examples of this model.