Fundamentals of Data Analytics
“Big data” analytics of the computing age has a very recent history. Brought about in the 1990s largely, when the internet became a tool beyond the U.S. military and selected companies, data analytics actually became useful to the consumer and companies (who tracked consumers, while vice-versa also happened). There were initially 2 main types – paid and owned media, then “earned media” was added. While paid media was strong (and remains so for some advertising), click-through-rates (CTR) for banner advertising has been declining due to “banner blindness” resulting from excessive ad use of this type. Despite the rate of 0.1-0.2% (CTR), advertising may still serve use as a “cookie placer” to track progress via third-party websites (beyond the primary website of landing) – these collect and share data after tracking the user. Owned websites allowed custom dashboards, content analytics, and mobile analytics became able to be gleaned; in addition, an important parameter, conversion rate, was determined. Finally, social media “engagement” resulting in such conversion played an important role, as did real-time analytics at hyper-speed.
Fundamentals of Analytics: Present Status
Social media is covered in a separate series (see RavishOnSocialMedia.com). However, metrics on data analytics such as Facebook’s total likes, reach (organic, paid, viral), engaged users (clicking posts), PTAT (people talking about this… total comments, shares, likes, stories, etc. over a period from 1, 2, or 4 weeks); Twitter keeps track of followers, retweets, replies, CTRs, as well as impressions (views); YouTube notes views, subscribers (signed up), likes/dislikes, comments, favorites, sharing; SlideShare notes followers, views, comments, and shares; Pinterest notes followers, number of separate pinboards created for a given account (e.g. companies based upon categories create separate boards); pin numbers, likes, repins, and comments; Google + uses similar metrics in its own terminology to Facebook; Flickr shares photos, using views, favorites, and comments as metrics.
Social media listening tools require a balance between data capture (some which capture certain channels, in part or whole, and categorize them), spam prevention, integration with other data sources, cost, mobile compatibility, application programming interface (API, allowing data capture and re-purposing), and other variables that aid in customer relationship management (CRM) interface. Companies like Marketwire acquired Sysomos (with eventually products like MAP and Heartbeat), with consolidation in the rather fragmented industry with over 230 “listening” companies. Radian6 with its visibility, languages, media types, regions, source filters, keywords, and other features, offers ability to monitor social sites, as do several other programs.
Data analytic methods include analysis of social media, email marketing, mobile marketing, display advertising, and search marketing, via paid or natural (organic) methods. Also, programs like Google Trends show use of specified category/words for groups (locally or globally, by region, for example).
More specifically, audience analysis tools, such as location-based results, personal history, browser type, device type, and other inputs can help tailor content suitable for users in the future – search insights, user surveys, website profiling, web analytics, social listening, influence analysis, sharing analysis, social profile/analytics, and event triggers, all have been used by programs.
Content audit items can include content type, ownership, topics, and keywords – facilitated by programs such as PageTrawler and others – with metrics being developed regarding linkage among pages, visits, time spent, exit rate, and bounce rate (single-page visits). Along the same vein, but faster, real-time analytics such as Woopra or Chartbeat allow approximately 15-second updates of ongoing user interaction to site. In addition, analysis of content – e.g. performance, competition, price, etc. can be set up through a “learning agenda” which focuses upon answering specific questions raised, so that actionable steps can be taken by a company.
Social Media Engagement Software “SMES” which includes Argyle Social and HootSuite for small-medium businesses, also is aided by analytics (from programs such as SpredFast) of virality (number of users creating a story, impressions, engagement. HootSuite allows integration of data and analysis across various major social platforms with all features including Mobile except geolocation; Argyle Social allows this integration (though to limited major social media and a designable blog), with added feature of e-commerce integration as well. Enterprise SMES tools include Spredfast (allowing also Social Media Management System or SMMS) for campaign management or similarly, Wildfire (which lacks historical data and is primarily Facebook-focused, while having a robust content creating tool also), Sprinklr (which automatically tags social media profiles so targeting can occur more effectively), Vitrue (acquired by Oracle, which allows publishing in advance also, with analytics), and Buddy Media (acquired by SalesForce).
Finally, a metric of digital influence, as defined by a “tipping point” or score of popularity, can be identified by Klout scores across platforms such as the usual social media networks (Facebook, etc.) but now Instagram, YouTube, LinkedIn, etc. With some 400 unknown metrics reportedly feeding into Klout, there is some question as to what is being measured; also, competitors such as PeerIndex, Kred, Tweetlevel, Bloglevel, and others (including potential for customization to value certain parameters such as SMES and other engagement, for example) exist. By a combination of paid, owned, shared, and earned media, categories of VIPs, exclusive access individuals, pich list individuals (e.g. for press releases), listening only (not usually with special access, but can influence adversely) individuals, an effective compendium of “digital influence” can be tabulated.
Fundamentals of Analytics: Future
One-stop shop: Each of the categories described above, involving “fundamentals,” have significant future implications with regards to their potential evolution. The future of analytics is a field that will likely grow by both organic and acquisition means – use of technology such as the cloud (to create a one-stop shop) may assist most users and hence build greater customer base, and this in turn could result from acquisition of companies already doing what may take longer and be more expensive in the longer run to do organically. A one-stop
Real-life integration (further): Off-line media tools still do not exist, which if possible, could integrate into these powerful online tools (e.g. if a recommendation were made over dinner or on a call) – while this would involve some aspects of privacy being agreed upon, in the strictly analytical world, more accurate “conversations” would be depicted.
Privacy Issues: A discussion of analytics contains almost inseparably a discussion of privacy-related issues. While it is conceivable that users “allow” for things that could conceivably be helpful (e.g. directing a website ad based upon prior website visits being tracked), the potential for the same data to result in something adverse (e.g. only being shown high-price items when the user may have been likely to buy a lower-price item if this were shown) exist. Certainly, the ethical implications of social media conversations being used commercially may be either welcome or raise concern. While blanket statements to this topic such as “people don’t care so much about data analytics as long as it helps them, though they should not intrude privacy substantially – by aggregating these in groups or de-identifying the specific person’s name” may be floated, not all individuals will ascribe to these principles. Hence, the debate, as with all technological advancement, remains one of responsible use. Data analytics offers, to paraphrase economist.com, a “bigger crystal ball” which in turn (according to forbes.com) makes a statistician in the running for the most attractive person of the future.
An interesting compilation, from Harvard Business Review (www.hbr.org) noted that in 2012, 2.5 exabites (one billion gigabytes) are created each day, and Walmart collects more than 2.5 petabytes every HOUR from customer transactions – with 1 petabyte being 1 quadrillion bytes. This trend is likely already significantly higher, as of 2015 and beyond. The key question is, “how will that information be used in the future?”
(Information above has been acquired from numerous sources, including www.economist.com; www.forbes.com; Hemann C, Burbary K: Digital Marketing Analytics, Que, 2014; and/or www.hbr.org)