Get to know our team at UpStream. This week we talk about Big Data with our CTO, Brandon Mason.
How do you see your customers using Big Data in the future?
The next wave of big data is all about applied uses. The first wave was the exclusive domain of engineers. The solutions that evolved from that first wave were what engineers were comfortable with – and tended to focus on – what I would consider plumbing. The next wave is all about enabling other groups to put big data to use. I am sure many people would point to Hive or Pig, but let’s face it – that is just not easy enough for the average person to use. When big data reaches the penetration of Microsoft Excel or Powerpoint, then things will really change. Empowering people to put big data to use in all sorts of ways an engineer couldn’t even imagine, that is when we know big data has really made it and we can drop the quotes around the term “big data.”
What success are you most proud of?
“Big Data” has been around for a quite some time, but only recently has it started to drive real innovation. The first wave of innovation was all about infrastructure, which made a huge impact on businesses that had to deal with massive scale. Think Twitter, Facebook, Google, and ad networks. All of them have a massive amount of daily event data to handle, both in terms of transactional systems as well as reporting.
A second wave of innovation is just starting to form and its focus is all about applied big data and real business intelligence. The problem with the first generation of big data tools and applications is that it is only engineer driven. I count myself in that engineer bucket, but unfortunately we tend to think about transactions per second, how many rows can I sum, and interesting ways of getting objects from point A to point B. The second wave is all about putting big data into the hands of practitioners. We recognized that many people in marketing or analytics groups just don’t have access to those same tools, the vast majority of big data is still in the hands of engineers and IT. UpStream is working hard to bring intelligence and access to big data to marketing departments. Building a better attribution mouse trap and putting marketing in the driver’s seat is our vision of the real potential of big data. Participating in this sort of revolutionary change to marketing really makes me proud.
What do you see as UpStream’s advantage over its competitors?
Unique ideas don’t come along every day. Sometimes it happens by accident and sometimes it comes from a lot of hard work and sleepless nights. I would like to think for Upstream it is a little bit of both. We were not originally trying to solve the attribution problem – we were actually trying to help marketing departments with their marketing finance problems (how much to spend on campaign X in order to get Y in returns).
We stumbled into attribution because we are a very analytical and cross functional team. This isn’t an engineering only organization. If it was we would only have come up with the same solutions everyone else did such as last click or choose your own adventure attribution methodology. After reading a recent Forrester report on attribution, I thought back to a college professor who had a quote from Ralph Waldo Emerson on his door, “A foolish consistency is the hobgoblin of little minds.” Companies continue to rely on (and pay significant sums) for solutions they know are wrong or come up with their own formula. A few brave souls have broken with the pack and come up with their own proprietary solutions, but when pressed they still say, “If I told you how we did it I would have to shoot you.” UpStream’s advantage over all these is that we are both transparent and have a proven statistical methodology. We encourage our customers to understand the methodology and get involved.
What are companies changing now that they have deployed the UpStream attribution solution?
It has been really interesting watching dollars switch from some buckets and flow to others. Last click heavily weights a search engine’s (such as Google) influence for many multi-channel brands and underrepresents other channels such as display advertising. Once companies really understand how their spend is driving sales, it is really interesting to see how fast change happens. We also think channels like social, which are neither direct response nor brand, will get a much fairer shake using this type of attribution.
What’s next?
Attribution for us is just the tip of the iceberg. It is the foundation upon which we are building a variety of intelligent solutions. Once you have the statistical basis to better understand what is driving sales, the sky’s the limit. Using our attribution as a base, we are able to drive a variety of additional insights for a marketing department. We can figure out who is more likely to be influenced by advertising and what channel or type of marketing (or combinations) will have the most impact. We can help put a price on the value of a user, which is immensely important to real time bidding (RTB) systems, or identify the real profile of valuable customers, which makes data buying on data management platforms (DMPs) much more efficient. We are convinced that accurate attribution unlocks a variety of doors that have been previously closed to marketing.
