Posts Tagged ‘Startups’

Event Recap: The Secret to Scaling a Tech Organization

October 14, 2013

On September 25th, 2013, I  attended the event, The Secret to Scaling a Tech Organization, presenting Gilt CTO, Michael Bryzek. This event was hosted by the New York Technology Council.

25 years ago, innovation happened in large research organizations handled by centralized bureaucracies. As a result, in those days, scaling technology was a slow and arduous process. But with the advent of the 21st century, the rise of the internet has changed this system. The bureaucratic model of managing innovative organizations, especially those that deal with technology, can no longer meet demands of the 21st century global economy. With social and economic paradigms shifting towards decentralization, how can today’s large scale businesses and organizations scale technology to meet the demands of today’s 21 century economy while still being innovative and efficient?

Michael Bryzek, CTO of Gilt, uses his organization as an example of the solution to this question. Gilt is a unique approach to online shopping where customers are given a limited amount of time (36 hours) to access discounted top brand products. This way, brands remain exclusive and customers get discounts on brand name goods.

But how does Gilt scale to accommodate not only the flexibility required for its market but also its overall growth? The innovation and efficiency required to create the systems that facilitate search engine optimization while exposing new brand products to customers on a daily basis under the umbrella of discreetness requires a new type of organizational model.  Since innovation is what drives the growth of start-ups, according to Bryzek, the secret to scaling a technology organization lies simply with the concept of trust. The ability of an organization to create an environment that cultivates innovation is tied to how individuals can interact with each other, as he says, “trust is an important ingredient to scale, recruiting, and innovate within any organization… Behind a company are people and if true trust is created within an organization, then people create great things.” In order to create the environment for innovation, the workplace setting must be merged with an environment where people can interact naturally.

Gilt’s organizational model is decentralized and rather than building bureaucratic hierarchies, they focus on fostering autonomous groups of specialists that are organized into teams. The problem with centralization and organizational hierarchies, according to Bryzek, is its slow speed of innovating ideas. When Gilt attempted to innovate (innovation is defined as not merely having an idea but also executing that idea) using a centralized model (a grand list of ideas), it could not keep up with the speed of the market because it took too long to come to a consensus on what idea they wanted to prioritize. As a result, the people at Gilt wanted to change this and focus on defining a strategy that quickly honed in areas that they want to invest in.

Under this new organization model, the prioritization of ideas must rely on KPI (Key performance indicator (metrics) KPI provides two advantages for an organization, in that, it fosters a debate on what matters to that organization and it provides a metric that justifies the prioritization of an idea. Within Gilt, the metric (such as how many people searched on Google and registered on Gilt in 30 days), will quickly drive the debate to the fundamental things that are important to the organization because an idea is tested empirically, for example: if Gilt had to decide on whether or not they should prioritize on developing a program that helps optimize the discreet searches for brand name Black dresses on Google or, they would check how many people searched on Google and registered.

However, the teams at Gilt aren’t merely a group of professionals divided into specialized positions in the organization. But what Gilt employs are teams of “smart and autonomous people. The teams are organized on the psychology of “people working together” where the members of the team decide what each other does based on what they learn from each other. Furthermore, going along with the paradigm of decentralization and autonomy, each team are afforded ownership of what they produce for Gilt (pieces of code and etc.) This way, if there is a problem with a certain program or platform, the teams that worked on them take notice and quickly fix it.

Therefore, the key to scaling in the 21st century is not to create innovation through centralization but to create the decentralized environment that cultivates innovation. That and a whole lot of trust!


iTechArt Event Recap Data Driven NYC #18

October 9, 2013

On September 17th, I attended Data Driven NYC, an event hosted by Bloomberg Beta and organized by Matt Turck with presentations by Fred Wilson, co-founder of Union Square Ventures, financial data companies: Etimize, Quandl, and Kensho, and featuring spotlights on Paris based data start-ups: Yseop, Datailu, and OpenDataSoft.

The event started with a “fireside” chat with Fred Wilson, the co-founder of Union Square Ventures on the topics of preparatory data assets and how networks have evolved to make such data defensible.

“What are some examples of preparatory data assets and why do they matter?” asked Matt Turck to start the “fireside” chat.

Fred Wilson answered, “We came to realize that the only defensibility that you can get with data is the defensibility around your own data.” The problem of preparatory data assets revolved around the defensibility of that data.

“If you are getting data assets from somebody else, they can always give it to someone else…” Wilson continued, “Even if they are bound by exclusive contracts… you will pay through the nose for that data. So, unless you own your data, you don’t really have defensible data assets.”

Therefore, Wilson believes that the defensibility of data lies within networks. The tremendous amounts of data formed within networks such as Twitter for example, provides advantages since “its users contribute the data.” Data such as search queries, user connections, and content creation leverage the data since it comes from within your own system.

But how do you keep the data in your network?

Wilson first uses a story to highlight the indefensibility of software and the power of the data asset, “Three entrepreneurs start a company and create new enterprise software. They sell their software and start their companies. However, Bloomberg beta hires two programmers to knock off this new software and sell this software for a cheaper price hence cutting into the profits of the major software companies. Then a group of hackers create an open source version of the software which ruins both the software companies and the programmers from Bloomberg Beta. Therefore software alone can’t contribute to the defensibility of data.

For Wilson, networks are the intersection between software and data because what networks provide is connectivity. And connectivity is a major component to the defensibility of data.

Edmodo is an example of how the connectivity of networks attributes to the defensibility of data. According to Wilson, what Edmodo accomplished was, by using facebook’s social networking programs such as the newsfeed, it enabled teachers to communicate with each other for free. Communication “allowed teacher to connect to each other and share lesson plans, homework, tests, and etc. And all of this was free.” By the virtue of using Edmodo to connect with one another, data created by teachers are “living within Edmodo”. Therefore, the network effects will be around “the content created and shared within the platform” And since, now content lives within the system, the most important data (conversational/transactional) stay within the network.

Jean Marc Lazard, CEO of OpenDataSoft, made the first spotlight presentation.

Lazard explained OpenDataSoft as a cloud based platform that condenses big data technologies into a cost effective and easy to use data management system. The platform can collect all kinds of data from any sources and reformat the data into the user’s requirements.

Marc Batty, chief custom officer at Dataiku, made the next spotlight presentation.

Dataiku, according to Batty, is a data science application/platform that collects raw and external data in order to evaluate and predict trends in business processes. The platform uses the data collected to update and re-configure itself via machine learning technologies.

John Rauscher, CEO and co-founder of Yseop, made the final spotlight presentation.

Rauscher explained Yseop as a software centered on artificial intelligence that removes the human element from administrative tasks. Yseop uses data to automatically perform lead generations, up-selling, cross-selling, writing reports, and post meeting summaries. As such, it makes administrative tasks far more efficient since with Yseop, a company’s sales teams can spend more time on customer interactions and less time on administrative tasks in order to close business faster.

Leigh Drogen, CEO of Estimize, made the first financial data presentation.

Estimize, according to Drogen, is an open financial estimates platform which facilitates the aggregation of fundamental estimates through public and private crowdsourcing. By sourcing estimates from a wide range of individuals, Estimize provides both a more accurate and more representative view of expectations compared to only sell side data sets which suffer from biases.

Abraham Thomas, Chief Data officer at Quandl made the second financial data presentation.

Quandl, according to Thomas, is a search engine for quantitative data that makes it easier to use and find all numerical data on the internet. Quandl accomplishes its search inquires through the millions of indexed numerical datasets from around the internet as Thomas explains how Quandl is used, “When you click on a particular dataset listed in our index, Quandl goes to the original source of that data, extracts the most recent version of that data, cleans it up, and gives it to you in whatever format you want.”

Daniel J. Nadler, chief executive officer at Kensho, made the final presentation of the evening.

Nadler explains Kensho as a platform that combines high-speed search algorithms and machine learning to create a new class of predictive analysis tools in finance. By combing high speed search algorithms ad machine learning, Kensho tackles the two biggest challenges surrounding financial analysis on Wall Street today; scale and automation.

About Me:

I have has consulted with dozens of next-generation technology start-ups across the United States in information management, human resources, and custom enterprise and mobile application engineering essentially leading to next stage funding rounds and acquisition.

About iTechArt:

iTechArt is a leading custom software engineering company focused primarily on establishing and operating remote, fully dedicated, teams for product development and customization, proof of concept and software prototyping, application modernization and re-engineering, QA and testing. Since 2002 iTechArt has been successfully acting as a trustworthy and reliable technology partner for dynamically growing technology start-ups.