Metrics and Dimensions and Reports — Oh My!

Recently I started read­ing a new book on social media, and I quickly became con­cerned with one part of the book that pro­vided a laun­dry list of social media “met­rics”. Social media is cer­tainly a hot topic, and many com­pa­nies are inter­ested in how to best mea­sure this new mar­ket­ing chan­nel. Unfor­tu­nately, too many of the “met­rics” in this list were not even met­rics but instead reports (e.g., Buzz (?) by social chan­nel, Sen­ti­ment by vol­ume of posts, Method of con­tent deliv­ery, etc.). I’ve seen this prob­lem not just in this book but across our indus­try — blog posts, arti­cles, whitepa­pers, pre­sen­ta­tions, mar­ket­ing mate­ri­als — and yes, even at my own company.

It’s con­cern­ing that as we open up a flood gate of new met­rics — both mean­ing­ful and use­less — many peo­ple still don’t have a firm under­stand­ing of the dif­fer­ences between met­rics, KPIs, dimen­sions, and reports in the exist­ing web ana­lyt­ics world.

Too often these data terms are being care­lessly inter­changed or mis­used. You may be think­ing: “It’s harm­less if my mar­ket­ing direc­tor wants to refer to all our met­rics as KPIs. At least he’s excited about the data. It’s okay if our agency inad­ver­tently refers to reports as met­rics. No big deal. Brent, you’re putting the ‘anal’ in analytics.”

In 2000, I was under­stand­ing (and I was in the process of fig­ur­ing it out for myself!).

In 2005, I became concerned.

In 2010, I’m just plain upset.

How can we truly take a data-driven mind­set to the next level within our var­i­ous orga­ni­za­tions and the field of web ana­lyt­ics when we don’t use data ter­mi­nol­ogy prop­erly? Sloppy logic waters down the impact that ana­lyt­ics data can have on our orga­ni­za­tions. It cre­ates a weak foun­da­tion for build­ing the data-driven evo­lu­tion we’re try­ing to fos­ter. We can’t have indus­try experts, ana­lysts, con­sul­tants, busi­ness man­agers, and exec­u­tives per­pet­u­at­ing this prob­lem as we enter emerg­ing ana­lyt­ics fron­tiers such as social media, mobile, and apps.

Enough is enough — it’s time to tighten the screws. Are you with me?

Def­i­n­i­tions

What is a met­ric? The dic­tio­nary defines it as “a stan­dard for mea­sur­ing or eval­u­at­ing some­thing; basis for assess­ment.” Avinash Kaushik states “a met­ric is a num­ber.” To be more spe­cific, met­rics are expressed in numer­i­cal val­ues. 4,563 is not a met­ric, but the value of a met­ric (e.g., 4,563 could be the num­ber of leads in a month). The Web Ana­lyt­ics Asso­ci­a­tion (WAA) fur­ther clar­i­fies that there are two types of met­rics: counts (e.g., 125,909 vis­its) and ratios (e.g., 2.1% con­ver­sion rate).

What is a dimen­sion? WAA defines it as “a com­po­nent or cat­e­gory of data. Met­rics (counts and ratios) are mea­sured across dimen­sions.” Dimen­sions can be a vari­ety of attrib­utes such as search engine, coun­try, page, refer­ring domain, date/time, key­word, etc. Dimen­sions are expressed as tex­tual val­ues. For exam­ple, Cal­i­for­nia, Utah, Vir­ginia, and New York would be tex­tual val­ues of the US States dimension.

What is a report? In terms of web ana­lyt­ics tools, a report is a col­lec­tion of data or val­ues for a spe­cific set of dimen­sions and met­rics. Every report has at least one dimen­sion and one met­ric. The data can be pre­sented in graph­i­cal or tab­u­lar for­mat and most likely a com­bi­na­tion of both for­mats. A report is really where every­thing comes together.

You may be won­der­ing about trended met­ric reports such as a Page Views or Vis­its report, and whether or not these reports have a dimen­sion. In these cases, time itself is the data dimen­sion (e.g., June 1, June 2, etc.).

Sim­ple Scenario

Let’s look at a sce­nario where you had 5,500 sub­mit­ted leads in June. 5,500 is the value of a met­ric (Sub­mit­ted Leads) that you’re cap­tur­ing with a cus­tom event. In many cases, met­rics by them­selves may not be that insightful.

The data fre­quently becomes more inter­est­ing when we add dimen­sions. Data dimen­sions break out or allo­cate met­rics across dif­fer­ent tex­tual val­ues or cat­e­gories. For exam­ple, if we added the dimen­sion of “US State” (cap­tured in an eVar) we’d be able to see how many leads we had per state. It may be more infor­ma­tive and use­ful know­ing almost half of your sub­mit­ted leads are com­ing from Vir­ginia than know­ing the total num­ber of leads in a month.

It’s impor­tant to under­stand and be clear on these basic aspects of web ana­lyt­ics data, espe­cially when peo­ple are fre­quently con­fus­ing the terms. A lit­tle more rigor around what we define as met­rics will ensure that not only our web ana­lyt­ics focus is sound, but we’re also bet­ter able to nav­i­gate the still rel­a­tively uncharted waters of social ana­lyt­ics. In my next post, I’ll dis­cuss some spe­cific ways in which we can raise our stan­dards for metrics.