Déjà new’ all over again

“It’s crazy that you guys haven’t met each other already,” my friend Bryan wrote in the email that he had sent to us, “because you have an awful lot in common.” A couple of weeks later, with a day of meetings in Silicon Valley already on my calendar, I scheduled a “get to know you” session with Bryan’s friend Sanjay Mathur at his analytics consulting firm’s offices in Sunnyvale, Calif. Walking into a tree-filled open courtyard that was ringed by two floors of office suites, I was suddenly stopped short by a powerful sensation, a visceral recollection of a nearby place and a long-ago time.

This building, I was startled to realize, was a nearly exact replica of the one where my analytics consulting firm had had its first offices, just a few miles up the road. Indeed, the two were separated only by two train stations – and a little more than 20 years.

This sense of déjà vu continued when I arrived at the offices of Silicon Valley Data Science (http://svds.com/). I was greeted by Sanjay, who was unflaggingly gracious despite my tardy arrival and leisurely conversational style (I had long ago downshifted from start-up speed to professorial plodding). Like me, Sanjay was of 100 percent Indian descent and had grown up mainly in the United States. Like me, he had brought together a small group of partners to form a consulting firm with the idea of using data, math and software to tackle business problems [2]. Like me, he was the company CEO.

Once I stopped marveling at the similarities, however, some major differences quickly came into focus. When we started Onward, I had slightly more than a year’s worth of full-time work experience, about the same as my partners Rob Luenberger (who was just finishing graduate school) and Stergios Marinopoulos (who had been doing contract software development for a few months since leaving Stanford). We had begun in a very ad hoc manner, adding clients and staff slowly over several years and funding the business solely through revenues. Our approach to software development for our clients was basically “get Sterg to build something.” We had almost no experience with demonstrating the economic value associated with most of our projects. Our first several hires were all friends with whom we had worked previously, and our contracts with our clients were often very uniquely structured.

By contrast, Sanjay had already spent 16 years at Accenture, rising from newly hired associate to partner and building several analytics-centric practices in the process. He had also followed this up with a stint as senior vice president for product management at an early cloud-based enterprise services company. Similarly, co-founder and Chief Technology Officer John Akred came into this venture with deep experience solving business problems using statistics, machine learning and large-scale application development on platforms ranging from SPSS and SAS to Hadoop and Cassandra. To fund the company’s launch, the SVDS founders had no trouble raising $3 million in venture capital based on a healthy valuation.

Since then, the company has grown rapidly, with 45 full-time employees on board shortly after their second anniversary, many of them already proven leaders in their respective areas of specialization. This hiring has in turn strengthened the company’s deep expertise in all of the pieces of the analytics value chain. Client projects are all based on the agile project methodology, and the company’s ability to create value through a series of iterative sprints is a key part of its customer value proposition [2] and contracting structure.

By the time I left the meeting with Sanjay, I felt a profound sense of envy. So many smart things that he and his partners have done from the outset, and such a skilled and rapidly growing team. In less than two years, such an impressive roster of clients, ranging from global brands to large-scale retailers to computer game makers to financial services firms to startups such as Edmunds.com. Such clarity about what they are doing and how they are doing it, and such great upside! Walking out the door that night, I found myself haunted by a simple nagging question: Why had we never managed to achieve all of this in our eight-year history?

Pondering all of this on the hour-long drive back to Oakland, however, I managed to regain some perspective. The world has changed profoundly since the time that we had started our company back in 1994. “Math” and “data” were four-letter words that we couldn’t put in our company name or on our business cards for fear of scaring off prospective clients. To bring in business, we did a lot of what my friends in sales call “missionary work,” evangelizing to a world that was far less technologically advanced than today’s; for example, I clearly recall one of our clients, a C-Level executive at a consumer software company, confidently assuring us that “our customers are not going to be getting to the Internet.”

Standardized enterprise software systems (e.g., CRM, ERP) were not nearly as ubiquitous as they are today, and as such getting our clients to deliver the type of data that these systems routinely capture was always an adventure. Data transfers via FTP were often an exercise in frustration while data extracts on tapes (and eventually CDs) moved slowly through both the client’s IT department and the U.S. mail. Our PCs were so much slower, our hard disks had so much less capacity, and nobody was talking about mobile or social or cloud computing.

These are different days. The global buzz around “predictive analytics” and “big data,” and “data science” is almost deafening. The positive name recognition that SVDS gets from both “Silicon Valley” and “data science” is significant, both in the CEO’s staff meetings and the CIO’s office. Two decades of expanding IT infrastructure (and decreased storage and computing costs) have made huge mountains of data readily available (and far cheaper and easier to analyze). Connectivity is now something that all of us take for granted, and today’s SVDS projects often have the benefit of corporate collaboration tools that make working remotely with clients so much easier. Today’s challenge is figuring out how to use all of this data to drive positive business outcomes, rather than simply convincing business leaders that it is worth trying.

In the end, this visit to SVDS left me slightly wistful and somewhat proud. “Wistful” about the many things that we had not known about when we started Onward and the things that we might have done differently if we had a chance to start again today. “Proud” that we had had the courage to launch our venture and the tenacity to persevere (and ultimately succeed) despite being woefully inexperienced and somewhat ahead of our time.

Eventually, I realized that what I was really feeling was not envy but simply loss. Visiting with Sanjay at SVDS had been a clear reminder that I no longer possessed the same enthusiasm, vitality and sense of infinite possibility that were such hallmarks of my early professional career. This evolution is natural, of course, and yet it can still be a challenge for old, retired people like me to come to terms with the times in which we had come of age and the routes that we had taken to get from there to here. Perfect hindsight, unfortunately, causes us to see many of the choices that we made along the way as mistakes, even when we rationally understand that this labeling is often both unfair and inaccurate.

However, even after I sorted all of this out, “acceptance” (the final stage of grieving [3]) came a little bit more slowly. This sense of loss is especially acute in our business, for the early 21st century is a time filled with a huge volume and variety of opportunities for analytics professionals, as the early success of SVDS so clearly illustrates. As the philosopher and economist John Stuart Mill said long ago, “That which seems the height of absurdity in one generation often becomes the height of wisdom in another [4].” Call it “déjà new.”

REFERENCES

1. http://svds.com/post/welcome-silicon-valley-data-science
2. For more details on this, see http://svds.com/post/successful-data-teams-are-agile-and-cross-functional
3. http://psychcentral.com/lib/the-5-stages-of-loss-and-grief/
4. http://www.bartleby.com/73/2001.html