The Netflix-Comcast truce has demonstrated once more how crucial video has become for today’s internet. YouTube alone streams enough footage each month to theoretically entertain every single human alive for four hours. Facebook users spend an average of 84 minutes a month watching clips on the social network, topping five billion views in January.
The data inside each clip and metadata about each and every viewer’s interaction with a video can make or break marketing campaigns. But are companies making use of the vast treasure trove of data that all those streamed videos give them?
So far, the answer is no. Using big data to boost one’s sales and marketing activities may sound like old news, but most companies today don’t use the full suite of modern business intelligence (BI) tools at their disposal.
Some have embarked on implementing the open source Hadoop framework for data warehousing, including newer iterations such as Impala that make up for the lack in speed of the initial Hadoop versions. Some companies are trying new approaches to turn the entire web into a data repository, connecting sources across the cloud to each other and to their various on-premise datasets to run complex queries in a browser. And some are betting on new appliances that supposedly make mining your data as easy as a search query.
But most companies still struggle to make sense of the basic requirements for all the different big data technologies out there — from budget to necessary staff skills. They also need internal buy-in to connect entirely new data sources to their sales, marketing and other activities to get that 360-degree view of their value chain and operations that software guys have been promising.
That’s a pity because mining video data is a particularly valuable asset. The foray into the rich data sets of social media and video lets companies large and small literally see more and sell more.
Take one European enterprise my firm works with. This company noticed that its sales of one product had shot up and almost drained inventory in a few days. But why? When the sales team talked to the social media guys, they found out that a video about the new product had been viewed more than 100,000 times the day before the spike occurred.
The firm used a team of two to pull together data across the web and inside its firewall: online orders and conversion rates, data from its YouTube and Vimeo accounts, plus Google Analytics and Facebook Insights. Turns out the bestseller story was a bit more complicated.
YouTube views had indeed shot up, but they had only led to a 15 percent increase in orders. What really drove the unusually high sales was something else: the moment when die-hard fans started spreading the word. They shared the clip everywhere from Tumblr to Facebook and got their friends to watch it on mobile devices. Viral plus handheld generated a 40 percent sales increase, but that tidbit only showed up once all the dots were connected.
The company went a step further and pushed this analysis to its salesforce users. It was less a pep talk than advanced prep work for the next launch. “This intel convinced us to syndicate the same content on different channels, but to properly engage each type of audience, whether we’re talking to impulsive, Twitter-happy buyers, careful researchers on Quora, or collector types on Pinterest,” the marketing head told me. “For the next launch, we decided to focus on a mobile promotion that generated similar sales.”
Mining video data is the next big thing in harnessing big data. It simply is too big a data pool to ignore it. YouTube alone has more than a billion unique viewers each month, 80 percent of them from outside the U.S. The number of subscriptions has tripled since last year, and 40 percent of all content is viewed on mobile devices. This is why the POV should meet the POS.
Only when you mash up all these pieces of information, and do so as quickly as possible, do you stand a chance to establish cause and effect. It might not sound as sexy as “big data,” but mining video clips brings enterprises one step closer to understanding marketing success — and how to repeat it. Even better, there are tools out there that do not require nerds.
It would be wrong to declare one data source is suddenly more important than all the others, but companies need to put the spotlight on video and marry those insights with bone-dry sales and marketing numbers.