Page 117 - MYM 2015

Big Data:
What It Can Do and What It Can Not Do
Glen L. Urban PhD
OHistory of Big Data
ne of the hottest topics in marketing
today is “big data”, but in reality data has been getting bigger for 40 years.17 In the 1960’s for example, rms relied on Nielsen warehouse withdrawal data to measure
market share. This was based on manual inventory changes recorded at warehouses and reports were available on a bi-monthly basis. In the early 1970’s this data became computerized and became available weekly, but it was not until UPC codes (Uniform Product Codes) were invented by IBM in 1974 that data increased exponentially.
Information Resources Inc pioneered in 1979 by placing UPC scanners in a panel of retail stores. That made available daily data on 10,000 SKUs (Stock Keeping Unit bar code) at 1,000 stores. Within ve years such data was available for most stores. This purchase data made possible “logit” analysis (advanced regression analysis) and allowed accurate estimates of response to price promotion (Little and Guidoni, 1983). The 1980’s saw major advances in data and analytic modeling. By 1990 most rms had their own computerized data stored by product, consumer, and purchase date
along with the price and point of purchase stimuli by
Abstract: Big Data has exciting potential to contribute to Marketing effectiveness, but it cannot solve all problems. A balanced view of Big Data supplemented by eld experimentation and panel based behavioural research is recommended. Mine the data with marketing analytics, but also design speci c research to support high priority marketing problems.
17 Big data refers to the exponential growth in data. In marketing today it includes sales data at the micro level (product code SKU, store location, scanner records, customer relationship data bases), clicks on internet, word content on search and email and social media sites like Twitter, location data from mobile devices, monitoring of targeted ads served to individuals, and video records. The term big data typically includes the marketing analysis tools and methodologies used to extract value from the data.
your
mind I117
store where the purchase was made. Competitive information came from IRI and Nielsen store scanner panels. This enabled computerized interactive decision support systems to be built to implement and price, distribution, and promotion decisions.
In the 1990’s the internet dramatically impacted the amount of data. Now click streams could be tracked and targeting of individual consumers on the internet was possible. In the 2000’s Google pioneered sponsor links on search and targeted banners. Patterns
of clicks, words, and billing by click/conversion established a major increase in the amount of data available to advertisers. Emails could be monitored and ads delivered based on the words used in the email messages. Virtually all rms began to use targeted banner ads and search links and data bases recorded massive click streams histories. This data allowed more ef cient targeting of new media messages.
The advent of wide scale social media around 2005 triggered the next expansion of data. By 2010 we could then trend Twitter words, track clicks on sponsored brand posts on Facebook, and record views of YouTube videos. This built a huge archive of consumer level data that can be used to target marketing efforts.
marketing