PROPHET-izing the World of Data. Interview with Ashley Swartz, CEO of Furious

Ashley Swartz, CEO of Furious, started her career in manufacturing finance in enterprise and supply chain optimization software. But in this world of Iot, she was able to move into the manufacturing or products and mobile phones as a first introduction to mobile content and advertising. “I started my own agency in 2003 and then moved over to consult for the sell side of MVPDS, publishers and technology providers,” she explained, before founding Furious Corp in 2013 and building our platform which we call PROPHET. PROPHET connects and normalizes data from over 40 disparate advertising systems to report, forecast and optimize inventory and revenue across all sales channels and ad formats.

Charlene Weisler: How did you choose the name Furious?

Ashley Swartz: My nickname given to me by Matt Seiler when I worked for him years ago is Red Fury so I decided to name my company Furious. Matt Seiler, was CEO of PHD US when I worked for him and met him and most recently he was CEO and Chairman at Mediabrands.

Charlene Weisler: What MVPDs did you work for and what data did they use at that time on the sell side?

Ashley Swartz: I previously worked for Intel's On Cue platform, and we are currently working with multiple MVPDs today at Furious Corp. The datasets vary based on whether it is linear, linear addressable or digital. It included various 1st party data, set top box data, Nielsen data, Rentrak, 3rd party data including meta data, targeting data, etc. And, IF they had a DMP they might have ACR data (if needed) and or video player data.

Charlene Weisler: Tell me more about how the Furious platform works.

Ashley Swartz:  Furious' platform, PROPHET, is a horizontal platform that connects all the disparate advertising systems for a seller, automating the ingestion and cleansing of data to enable [near] real time reporting across all inventory, revenue (price) and audience for all platforms.  Its primary value as an enterprise system is in using this data to optimize yield at a portfolio level to maximize revenue and efficiency. It is powered by the most advanced state of the art optimization and forecasting algorithms known in the Data Science industry and we enhance them using domain specific knowledge of the media space and client specific data, thus creating best of breed adaptive and robust algorithms which present unprecedented accuracy and reliability. This is what powers PROPHET's forecasting, planning, pricing, allocation and optimization.

Charlene Weisler: Who is your competitive set?

Ashley Swartz: SAP, Freewheel, IBM and Accenture

Charlene Weisler: What other industries do you incorporate and what are their best practices that work in media?

Ashley Swartz: PROPHET leverages lessons learned from other large industries who have used Enterprise Resource Planning (ERP) software like SAP to connect their business, automate workflow, increase yield and productivity, as well as finance portfolio management tools to manage a mix of assets, optimize and rebalance portfolios over time. PROPHET utilizes methodologies from manufacturing, like Kaizen (continuous improvement) in its business logic using machine learning to power its self-correcting predictive and optimization algorithms.

Charlene Weisler: What is your definition of programmatic?

Ashley Swartz: The use of data or technology to improve the ROI on media.

Charlene Weisler: What do you think the impact of smart TVs will have on TV sales, content and measurement.

Ashley Swartz: In the short term, very little. As the owners of smart TVs begin to be more representative of a population, more valuable. We still have immense privacy issues to overcome to make the data actionable, and OEMS have a long way to go in ensuring the integrity of the data and establishing a realistic perspective on fair market value of the data.

This article first appeared in www.Mediapost.com

Carpe Datum at Gigaom

Have you heard of Gigaom? Me neither. But after attending their truly fascinating Data Structure Conference in NYC I think that a solution to true seamless cross platform media measurement may be at hand.

At Gigaom there were many innovative, ground breaking tech companies that seemed to have the capability to improve on the current measurement of media and consumer behavior through the use of artificial intelligence, machine learning, data blending and data discovery tools. And many of these companies have the ability to merge on one user interface all platform, device, app, software and site data and run it all in or close to real time. Today’s tech environment facilitates storytelling and data visualization to better leverage business intelligence which is exactly what we in the media space are seeking. Needless to say I was impressed by just about everyone I spoke to. Some of the innovative companies represented at Gigaom Data Structure are in this video:

Data Blending

Let’s begin in the area of data blending. According to a recent Gigaom Research study (Sector Roadmap: Data Discovery in 2014) “Data blending is the term used to describe the performance of analytics on a collection of data sets, each emanating from a different data source.” They say that data discovery products currently on the market are now capable of real time or close to real time data blending allowing users to pull in data, quickly mash it up and analyze it. Some companies like cloudera offer the ability to serve many different types of user workloads with access to the same data set within a single interface. This interface, according to CEO Tom Reily, “can manage all types of analytics and identify trends in customer experience.”

So I am thinking, can we take television data (Nielsen, STB or other), online, mobile and tablet (name your sources) and maybe even print, retail or transactional data and place it all on one interface and data blend? I know there are some companies offering this capability now in our industry but with limited datasets and outputs that still seem to be fairly silo’ed.

Artificial Intelligence

AI still seems very futuristic to me but in fact is fairly common in applications today. Some firms, such as alchemy api are in the business to “make computers more human” according to CEO Elliott Turner and others like Watson Solutions’ Stephen Gold offer “cognition as a service.” But what does this really mean for media? I think the possibilities are endless. We may be able in finally pinpoint how a viewer fully interacts with a piece of content, including the soft measurements of sentiment and engagement. We may even be able to use AI to predict which pieces of content are most effective at driving human behavior whether for tune-in, reaction, affection or to create a call-to-action.

Machine Learning

Combined with a level of AI, machines can be programmed to learn from experience. In this way data can be mined more efficiently and with greater precision to create software applications. Tim Tuttle of Expect Labs explained, “We built mind meld to listen to conversations and find information for you. Now we want to take that technology and apply it to any data you have." Think of Siri who seems to retain knowledge with each interaction. Siri is just the tip of the iceberg in machine learning capabilities. At some point we may rely solely on the machine to map out data insights. I hope I am retired by that time.

Data Discovery Tools

Donald Farmer, VP Product Management, Qlik believes that “When it comes to big data understanding we are in the Dark Ages.” This is exacerbated by unnatural interfaces that once had a purpose under old media but no longer apply. Farmer gave the example of the keyboard QWERTY system that was originally set up to avoid the typewriter keys from jamming. We still use this unnatural keypad interface even though our devices today do not have the problem of typewriter keys jamming up. How to we advance from old legacy systems and processes?

This and many other questions are a part of a huge data surge that impacts many businesses, including ours in media. Attending the Gigaom lets me know that we in media are not the only ones grappling with processing, measurement, analysis and data wrangling issues.