In this “Seven Keys to Cre­at­ing a Data-Driven Orga­ni­za­tion” series, I’ve cov­ered dif­fer­ent aspects such as secur­ing an exec­u­tive spon­sor, ensur­ing you have a mea­sure­ment strat­egy, and hav­ing ade­quate staffing and train­ing in place. Now we’re going to focus on the next key prin­ci­ple: estab­lish­ing and main­tain­ing cor­po­rate stan­dards.

In the past, I’ve com­pared hav­ing cor­po­rate stan­dards to the rules of the school play­ground. Whether you were play­ing soc­cer, foot­ball, or some other game at recess or lunch break, if every­one adheres to the agreed upon rules of the game, all of the stu­dents play­ing have fun. How­ever, when one or more of the stu­dents decide to do their own thing and ignore the rules or cheat, the game can be quickly ruined for everyone.

As web ana­lyt­ics was begin­ning to take off, early-adopter teams often oper­ated inde­pen­dently from other groups at their com­pany. They took full advan­tage of the flex­i­bil­ity and fea­tures of the tools for their spe­cific needs, and they tracked what­ever they wanted in what­ever man­ner they felt was appro­pri­ate. How­ever, as the usage of web ana­lyt­ics has spread within these com­pa­nies, they have come to real­ize it makes sense to coor­di­nate or stan­dard­ize what is mea­sured for the greater good of the organization.

Are we play­ing the same game?

Stan­dards are eas­ier to intro­duce and main­tain within your part of the orga­ni­za­tion. How­ever, if you need to report on busi­ness per­for­mance across dif­fer­ent busi­ness units or groups, how con­fi­dent are you in the num­bers? When you say “pass me the ball”, will you receive the antic­i­pated ball in your base­ball glove or might they just as eas­ily pass you a soc­cer ball, foot­ball, or golf ball instead? When you say “what’s the score?” will you hear vastly dif­fer­ent scores from var­i­ous parts of the busi­ness (e.g., “298”, “30-Love”, “4–3″). In other words, if you ask for a spe­cific report or met­ric, will you get what you expected? Can you trust the data across your business?

What level of cor­po­rate over­sight is required?

In gen­eral, larger orga­ni­za­tions (e.g., large multi­na­tion­als = mul­ti­ple web­sites + mul­ti­ple teams) require more coor­di­na­tion and stan­dard­iza­tion than smaller com­pa­nies (e.g., small online retailer = one web­site + one team). The orga­ni­za­tional struc­ture of your com­pany will be a key fac­tor in deter­min­ing how much cor­po­rate over­sight is required.

Typ­i­cally, com­pa­nies have geo­graphic– (e.g., coun­try, region, etc.) and product/function-based sites (e.g., brand sites, part­ner sites, online store, sup­port, etc.). Over­sight can get fairly com­plex when var­i­ous geo­graphic and product/function struc­tures con­verge. For exam­ple, Elec­tronic Arts’ web ana­lyt­ics team has to nav­i­gate and man­age sev­eral dif­fer­ent structures:

  • 80+ game sites
  • 15 dif­fer­ent game studios
  • 3 pub­lish­ing labels
  • 35 coun­tries
  • 13 online stores
  • 3 regions

Another fac­tor in deter­min­ing the right level of over­sight is cor­po­rate cul­ture. If your com­pany has a more decen­tral­ized struc­ture where indi­vid­ual busi­ness units are more autonomous, a rigid set of stan­dards may not fly. How­ever, even decen­tral­ized com­pa­nies like to com­pare and bench­mark dif­fer­ent parts of the busi­ness, and this can only hap­pen with some level of imple­men­ta­tion and report­ing stan­dard­iza­tion. As a result, it can become a del­i­cate bal­ance between the needs of indi­vid­ual busi­ness units and those of the over­all orga­ni­za­tion. The more stan­dard­iza­tion there is, the more the over­all enter­prise ben­e­fits. How­ever, too much stan­dard­iza­tion can limit the value of web ana­lyt­ics to indi­vid­ual parts of the busi­ness, which are often the key stake­hold­ers in cre­at­ing a data-driven organization.

Going back to the school­yard anal­ogy, just like ignor­ing the rules ruins the fun, hav­ing too many rules can also ruin the fun and over­all par­tic­i­pa­tion. When peo­ple aren’t enjoy­ing the game (i.e., not see­ing enough ben­e­fits for their team), they’ll walk away from the game and do their own thing. On the other hand, the indi­vid­ual play­ers may need to rec­og­nize who orga­nized the game and under­stand that a few extra rules won’t nec­es­sar­ily ruin all their fun.

A more cen­tral­ized orga­ni­za­tion might be able to more eas­ily stan­dard­ize its imple­men­ta­tions and report­ing, but it is still impor­tant to bal­ance the needs of the over­all orga­ni­za­tion with the needs of the indi­vid­ual busi­ness units. In some cases, your com­pany may set­tle on being coor­di­nated rather than com­pletely stan­dard­ized. The appro­pri­ate level of stan­dard­iza­tion will depend on the company’s orga­ni­za­tional struc­ture, cor­po­rate cul­ture, web ana­lyt­ics matu­rity level, and level of effort the com­pany is will­ing to invest.

Five rea­sons why cor­po­rate stan­dards are important

Intro­duc­ing cor­po­rate stan­dards can face some inter­nal resis­tance and can be dif­fi­cult to main­tain. Why is it worth all the effort? I’ve iden­ti­fied five reasons:

  1. Cre­ates a shared set of KPIs and reports: Orga­ni­za­tions need a com­mon cur­rency of met­rics and reports in order to bench­mark and com­pare dif­fer­ent parts of the busi­ness. With­out shared KPIs, no use­ful com­par­isons are possible.
  2. Enhances data integrity: When data is mea­sured and col­lected in the same man­ner, man­age­ment and users are more con­fi­dent that met­rics can be accu­rately com­pared across coun­tries, divi­sions, or products.
  3. Reduces sup­port time and costs: When the report­ing and met­rics are more stan­dard­ized, then sup­port mate­ri­als and resources can be shared across more busi­ness units. In addi­tion, the web ana­lyt­ics com­mu­nity within the com­pany can sup­port indi­vid­u­als more read­ily because they have more in common.
  4. Facil­i­tates best prac­tices shar­ing: When the inter­nal web ana­lyt­ics com­mu­nity have more in com­mon, it can facil­i­tate more data dis­cus­sions and best prac­tices shar­ing between dif­fer­ent teams, which raises the bar through­out the business.
  5. Accel­er­ates user adop­tion:  Cor­po­rate stan­dards will help to ensure the reports are clean and more user friendly, which will accel­er­ate user adop­tion of the tools and reporting.

“Build-it-and-they-will-come” fal­lacy

One final thought on cor­po­rate stan­dards is the mis­con­cep­tion that once they have been intro­duced, every­thing will take care of itself. In terms of the famous quote “Build it and they will come” from the 1989 Kevin Cost­ner film “The Field of Dreams”, it doesn’t work that way with cor­po­rate stan­dards. Orga­ni­za­tions are no bet­ter off if cor­po­rate stan­dards are com­mu­ni­cated and ini­tially adopted, but then later on not followed.

Ongo­ing mon­i­tor­ing and com­pli­ance is just as impor­tant as estab­lish­ing work­able cor­po­rate stan­dards. In order to make sure that stan­dards are being fol­lowed, a num­ber of the pre­vi­ously dis­cussed key areas come into play:

  1. An exec­u­tive spon­sor can empha­size and rein­force the impor­tance of adher­ing to the standards.
  2. A mea­sure­ment strat­egy serves as an ongo­ing guide­line and ref­er­ence for what needs to be mea­sured and how.
  3. Ade­quate staffing is impor­tant when you need a cen­tral­ized core team of ana­lysts to mon­i­tor and man­age ongo­ing stan­dards com­pli­ance through­out the organization.
  4. Train­ing rein­forces the cor­po­rate stan­dards and ensures they are more read­ily under­stood and followed.

All of these fac­tors are inter­twined in sup­port­ing the main­te­nance of cor­po­rate stan­dards. Build it and they will come. Main­tain it and they will come again.

In my next blog post, I’ll cover the impor­tance of deliv­er­ing quick wins to the organization.

  • http://www.webanalyticsdemystified.com Eric T. Peterson

    Brent,

    Again I have to say I am very pleas­antly sur­prised to read your work. As I men­tioned to Matt Langie last we talked it is refresh­ing to see Omni­ture talk about the whole pack­age required to be suc­cess­ful with an invest­ment in dig­i­tal mea­sure­ment. Some in your orga­ni­za­tion have his­tor­i­cally dis­agreed with a lot of what you’re say­ing so good for you for stand­ing up for what is right and true.

    I’d be inter­ested in your thoughts/position on the value of “mod­els” (e.g., the Web Ana­lyt­ics Matu­rity Model I cre­ated back in 2004 at Jupiter­Re­search and those that have emerged since.) I looked at your Gather, Refine, Align approach which appears to be more of a con­sult­ing frame­work than a struc­ture to score ana­lyt­ics usage within the business.

    Maybe we can dis­cuss in Salt Lake City in March?

    Any­way, nice work. I’m a fan.

    Eric T. Peter­son
    Web Ana­lyt­ics Demys­ti­fied, Inc.
    http://​www​.web​an​a​lyt​ics​de​mys​ti​fied​.com

  • http://blogs.omniture.com/author/bdykes Brent Dykes

    Eric, I’m glad you’ve enjoyed my blog posts. I want to help orga­ni­za­tions to become more data-driven and ensure they get the most from their web ana­lyt­ics investment.

    You’re right about my “Gather, Refine, Align” approach — it’s mainly a con­sult­ing frame­work used to estab­lish a well-defined web mea­sure­ment strat­egy. It’s not intended to be a scor­ing model for ana­lyt­ics usage within organizations.

    I’d love to dis­cuss my thoughts on the value of “mod­els” at Sum­mit in March. That would be great, and I look for­ward to see­ing you in SLC.

  • http://www.the-omni-man.com Adam Greco

    Great post!! This is def­i­nitely an under­rated area of focus. It is also pos­si­ble to use alerts and sched­uled reports to find instances where stan­dards are not being fol­lowed (i.e. Pages begin­ning with http:). Keep these posts coming!

  • http://blogs.omniture.com/author/bdykes Brent Dykes

    Adam, thanks for the encour­age­ment on the blog and the great sug­ges­tion to set up alerts and sched­uled reports to help with stan­dards monitoring.

  • http://www.rudishumpert.com Rudi Shumpert

    Brent,

    Great post here! I espe­cially like the “Five rea­sons why cor­po­rate stan­dards are impor­tant” Com­ing from a more tech­ni­cal back­ground I am all about the standards.

    It would be inter­est­ing to put those 5 items out on a poll and see how peo­ple at dif­fer­ent lev­els in a com­pany ranked those 5 items.

    –Rudi

  • http://blogs.omniture.com/author/bdykes Brent Dykes

    Rudi, that would be an inter­est­ing poll. Obvi­ously shared KPIs (#1) is one of the main goals of intro­duc­ing cor­po­rate stan­dards. If those shared KPIs are not consistent/accurate across the busi­ness, then #2 can under­mine #1. #3–5 are about the inter­nal web ana­lyt­ics com­mu­nity you’re try­ing to build. I feel they’re all impor­tant, but you’re right that they may be per­ceived dif­fer­ently depend­ing on your role and level.

  • http://immeria.net Stephane Hamel

    Very inter­est­ing posts Brent, Omni­ture offers great tools and from my expe­ri­ence, chang­ing the orga­ni­za­tional cul­ture to adopt a con­tin­u­ous improve­ment mind­set is often an issue. By pro­vid­ing a frame­work that encom­pass train­ing, staffing, strat­egy and exec­u­tive spon­sor­ship you demon­strate the crit­i­cal role of ana­lyt­ics for busi­ness improve­ment that goes far beyond the tool.

    Since Eric raised the ques­tion of web ana­lyt­ics matu­rity mod­els and said in the past this is not some­thing he believes in, I would like to share the results of an 18 months study of why some com­pa­nies fail and some oth­ers suc­ceed, a syn­the­sis of sev­eral other “mod­els” from var­i­ous fields (includ­ing the old one from Eric, which still has some good foun­da­tions but isn’t at par with the cur­rent state of the indus­try) , and over 20 years of expe­ri­ence. The result is a propo­si­tion for a “Web Ana­lyt­ics Matu­rity Model” (WAMM) which is gen­er­ally very well accepted, pend­ing, of course, one accept the idea that a framework/model might be use­ful. I would also love to hear your thoughts about mod­els in gen­eral, and the pro­posed WAMM in par­tic­u­lar. More info is avail­able at http://​imme​ria​.net/​w​amm

    Stéphane Hamel
    Imme­ria con­sult­ing ser­vices
    WAA Direc­tor & trea­surer
    UBC Award of achive­ment tutor, ULaval teacher
    http://​imme​ria​.net

  • http://blogs.omniture.com/author/bdykes Brent Dykes

    Stéphane, thanks for your com­ments. I agree with you that cre­at­ing a cul­ture or mind­set of con­tin­u­ous improve­ment (I coin it as ‘data-driven’) is crit­i­cal to the suc­cess of web ana­lyt­ics at any company.

    In terms of my thoughts on web ana­lyt­ics matu­rity mod­els, I’m famil­iar with some of the exist­ing mod­els such as Eric’s from his Jupiter­Re­search days. I’m look­ing for­ward to review­ing your new WAMM model in more detail (you prob­a­bly noticed that I recently down­loaded it). I see there has been a fair amount of debate about the value of mod­els. I will look for­ward to a lively dis­cus­sion with you and Eric on this topic at this year’s Summit.

  • http://www.idpro.co.uk/ Simon Iddings

    As the arti­cle implied, “…Ongo­ing mon­i­tor­ing and com­pli­ance is just as impor­tant as estab­lish­ing work­able cor­po­rate stan­dards…” and espe­cially so for cor­po­ra­tions. The need to adapt to changes in a rapidly chang­ing world and demands but be taken into account first, instead of forc­ing con­sumers and the orga­ni­za­tion itself to stag­nancy. Once a goal is stream­lined the types of cor­po­rate stan­dards to adhere to will be in place effortlessly.