Читать книгу Winning with Data - Bien Frank - Страница 7

Slow data is caused by an inefficient supply chain. Today's data supply chains suffer from a fundamental flaw in their architecture: The number of people seeking data dwarfs the number of people supplying data. The taxi dispatcher relaying passenger pickups by phone serves scores of drivers, each seeking their next fare. In many companies, this ratio may be much greater than 100:1. Is it any surprise that the data analyst team is seen as an enormous bottleneck, a chokepoint for the organization?

In the past, this flawed architecture functioned because most companies had a relatively small amount of data, most of it created by humans, and the competition wasn't using data for a competitive advantage. Without a substantial corpus of data to interrogate, only a handful of executives asked questions of their company's data, limiting the total number of requests. Most of the time, these requests were financial in nature and managed by the CFO and his organization.

But the amount of data that companies store today has exploded. According to IDC, from 2013 to 2020, the digital universe will grow by a factor of 10, from 4.4 trillion to 44 trillion gigabytes. It more than doubles every two years. This supernova of data contains insights relevant for every person within an organization.

Today, computers generate data at rates that far outstrip humans. Facebook records more than 600 petabytes of data daily on its users, almost all of it generated by computers. This trend isn't constrained to social networks. For example, Marketo, Eloqua, Pardot, and Hubspot pioneered the marketing automation software category not more than 10 years ago. These tools help B2B (business-to-business) marketers optimize demand-generation programs and prioritize leads. Market automation software snares data on website visitors to answer typical marketing questions: What content are they reading? How frequently are they visiting the site? What are the best messages to generate more leads?

Now that we're collecting all this data, we expect instant answers from it. In larger companies, the burden for answers rests on the shoulders of the data team. These scarce data analysts must process an ever-lengthening queue of work. Each request carries with it a unique set of intricacies. Perhaps the query involves a new data set, or a new type of data analysis, or a new visualization. And maybe the requesters of the data weren't quite sure what they were asking at the outset, so they revise the requested analysis, adding again to the workload and slowing processing time for everyone else. Rarely do most employees understand the complexity of their requests: the number of steps, the turnaround time, or the number of players required to answer their questions.