Con­tin­u­ing my series of arti­cles on the “Seven Keys to Cre­at­ing a Data-Driven Orga­ni­za­tion”, I’d now like to focus on the next key area — staffing and train­ing, which I’m going to break into two sep­a­rate arti­cles. Look­ing first at staffing, I like to com­pare the key roles in a web ana­lyt­ics pro­gram to the posi­tions on a sports team. Regard­less of whether your sport is base­ball, hockey, foot­ball, or “fute­bol”, a sports team has sev­eral cru­cial posi­tions which need to be filled.

Most orga­ni­za­tions have the nec­es­sary equip­ment — web ana­lyt­ics tools — but may find it sit­ting idle most of the time while one or two peo­ple run around the field try­ing to cover sev­eral posi­tions. If you don’t have all of the key posi­tions cov­ered then it can be dif­fi­cult to win the game or be suc­cess­ful with web ana­lyt­ics. Imag­ine how suc­cess­ful your favorite sports team would be with­out a pitcher, goalie, quar­ter­back, or striker — let alone a great one. Make sure you have a box of Kleenex on hand when you ask a Chicago Bears fan about their lack of a quar­ter­back (40+ changes to their start­ing quar­ter­back since 1992).

I’ve pre­sented this anal­ogy sev­eral times over the past cou­ple of years and even though I’m a die-hard hockey fan (it’s in my Cana­dian blood), I’ve pri­mar­ily used base­ball for this anal­ogy. At Omniture’s Tokyo Sum­mit in 2008, I used all-star out­fielder, Ichiro Suzuki, in my slides. For the Salt Lake City Sum­mit last Feb­ru­ary, I was orig­i­nally going to use Yan­kees all-star Alex “A-Rod” Rodriguez. How­ever, when he admit­ted to using banned sub­stances, I ended up using another player, Dodgers all-star Manny Ramirez … who was later sus­pended for 50 games for using performance-enhancing drugs.

My point in bring­ing this up is that just like pro­fes­sional sports, there are no short­cuts in build­ing a web ana­lyt­ics pro­gram. That’s the last thing you prob­a­bly want to hear when bud­gets are tighter than ever, and man­agers are being forced to “do more with less”. Other than sup­ple­ment­ing your pro­gram with fully-trained and expe­ri­enced con­sul­tants, there’s no magic pill for adding more staff or trans­form­ing inex­pe­ri­enced resources into web ana­lyt­ics experts. How­ever, even in this tough econ­omy, some deter­mined com­pa­nies view the ben­e­fits from prop­erly staffing a web ana­lyt­ics pro­gram and train­ing up employ­ees on the tools as cru­cial to their short– and long-term suc­cess. Let’s hope one of these com­pa­nies is not your competitor.

Web ana­lyt­ics play­ing field

At one of my pre­vi­ous employ­ers, I was part of a mot­ley group of soft­ball play­ers. The com­pany pro­vided us with branded t-shirts and decent equip­ment, but each week we strug­gled to get suf­fi­cient num­bers out to each game. If some­one had a last-minute sched­ul­ing con­flict, got sick, or pre­ferred what­ever was on TV that night (pre-DVR era), we were unable to field a com­plete team and would have to for­feit the game. It can be equally frus­trat­ing for par­tic­i­pants in a web ana­lyt­ics pro­gram when not all of the posi­tions are being cov­ered, and the com­pany is forced to for­feit many of the ben­e­fits derived from becom­ing more data-driven.

At a high-level, the fol­low­ing posi­tions are crit­i­cal to a suc­cess­ful web ana­lyt­ics pro­gram, although the size of your orga­ni­za­tion may change the need for cer­tain roles as well as the num­ber of peo­ple in each role:

  • Exec­u­tive spon­sor: Senior exec­u­tive who sets pri­or­i­ties, pro­vides high-level sup­port, resolves inter­nal con­flicts, and pro­motes data-driven decision-making through­out the organization.
  • Web steer­ing com­mit­tee: Formed by the exec­u­tive spon­sor and con­tains exec­u­tives from all web stake­hold­ers. It is focused on over­all web per­for­mance, strat­egy, and future initiatives.
  • Omni­ture owner: Man­ages Omni­ture rela­tion­ship and web ana­lyt­ics pro­gram at com­pany. The sin­gle point of con­tact for all web analytics-related issues for the organization.
  • Core team: Cen­tral­ized team of web ana­lysts focused on over­all busi­ness mea­sure­ment, report­ing, and analy­sis. They report to the Omni­ture owner.
  • Busi­ness leads: Busi­ness man­agers and ana­lysts at the business-unit level who are respon­si­ble for mea­sure­ment, report­ing, and analysis.
  • Tech­ni­cal leads: Web devel­op­ers who imple­ment Site­Cat­a­lyst at the business-unit level.

Two key play­ers on the ana­lyt­ics field

In a pre­vi­ous arti­cle, I dis­cussed the impor­tance of hav­ing an exec­u­tive spon­sor, who is essen­tially the gen­eral man­ager for your web ana­lyt­ics team. As you build your web ana­lyt­ics team, two key types of play­ers are needed. Web ana­lysts are one key player on the ana­lyt­ics field, and fre­quently fill the roles of Omni­ture owner, core team mem­ber, or busi­ness lead. They typ­i­cally pos­sess the fol­low­ing characteristics:

  • Business-minded with marketing-related background
  • Ana­lyt­i­cal, inquis­i­tive, and detail-oriented
  • Able to bridge gap between busi­ness and IT
  • Strong com­mu­ni­ca­tion skills

Web ana­lysts trans­late busi­ness require­ments into tech­ni­cal spec­i­fi­ca­tions. They need to be effec­tive com­mu­ni­ca­tors as they work with cross-functional groups, mes­sage the value of dif­fer­ent find­ings to the orga­ni­za­tion, and drive opti­miza­tion efforts through­out the com­pany. A 2006 For­rester research study showed that web ana­lysts could gen­er­ate up to a 3000% ROI on their salary cost.

Sim­i­lar to other famous sports duos such as Montana/Rice, Stockton/Malone, Gretzky/Kurri, Pele/Garrincha, etc., good tech­ni­cal leads can form a pow­er­ful tan­dem with web ana­lysts, allow­ing the ana­lysts to deliver the high ROI iden­ti­fied by For­rester. Tech­ni­cal leads pro­vide the tech­ni­cal know-how to get the right mea­sure­ment in place and work­ing cor­rectly. They pos­sess the fol­low­ing attributes:

  • Solid under­stand­ing of inter­nal web archi­tec­ture and systems
  • Cod­ing / web devel­op­ment exper­tise (JavaScript)
  • Famil­iar­ity with Site­Cat­a­lyst deployments
  • Busi­ness acumen

Fre­quently, large com­pa­nies use a pool of IT staff to ser­vice dif­fer­ent IT projects. When you have to work with dif­fer­ent IT resources for each new web ana­lyt­ics project, you con­stantly have to edu­cate new tech­ni­cal resources on web ana­lyt­ics and you’re also unable to des­ig­nate a sin­gle point of con­tact for all future tech­ni­cal issues. For exam­ple, you may want to adjust the tag­ging for a Flash micro-site a few weeks after launch, but find out that you’re forced to work with a com­pletely dif­fer­ent tech­ni­cal resource because the orig­i­nal web devel­oper has already been assigned to another IT project. It may not make sense to have an IT per­son fully ded­i­cated to only web ana­lyt­ics projects, but it is a best prac­tice to use the same resources for all web ana­lyt­ics initiatives.

In my next arti­cle, I’ll focus on the train­ing con­sid­er­a­tions that are required to become more data-driven as an orga­ni­za­tion. Get ready for train­ing camp.

Brent Dykes
Brent Dykes

John, I agree. Knowledge sharing is very important. I cover that a little more in the next part of this article, where I talk about establishing a web analytics community. You're absolutely right that the core team (if you're in a large organization) should play a key role in promoting user adoption throughout the company via various training opportunities. In terms of IT's role in web analytics, I've interacted lots of different organizations. I find that when the two sides -- business and IT -- form a close partnership on web analytics then you're going to have a better implementation that measures the right things in an efficient manner. When one side forces its will over the other, you either end up with a wasteful or impotent implementation -- both of which are suboptimal situations.

John Hunter
John Hunter

I think a big key is to have this important knowledge shared. It is important for IT staff and business staff understand the importance of web analytics. When they are doing their particular jobs they need to integrate these ideas into their tasks. I would actually prefer to have the knowledge spread out (rather than a dedicated team) though if the organization is large enough some dedicated staff can help - and one of their focuses should be to educate others. I find that IT often understands the importance of analytics better than the rest of the organization - and try to stop foolish decisions from being implimented. But often people see IT as an organization that should just implement what we tell you. And those doing the telling don't have an understanding of the impact of the decisions they make. Some may not like it but in the web space technical expertise is required to understand the options and strengths and weaknesses of various options.


great example showing grossman dropping the ball on something so easy, it really illustrates the point you're making. it is better to have one person dedicated to one project, it may cost more money but in the long run with all the time you save, by getting it done correctly will be worth it.

Brent Dykes
Brent Dykes

Sorry about poking at a still tender wound for Bears fans. I hope Jay Cutler works out for the Bears at the QB position. I've worked with a few great analyst / technical tandems over the years as well as many dysfunctional teams. I found the "assigned / dedicated" technical person to be really critical to organizational success with web analytics as they worked closely with web analysts. The implementations were tighter. Changes could be made more quickly. More people at these companies ended up relying on the data and reports. They could also push the limits of our tools and use them in innovative ways. All of these benefits fueled the success of their analytics programs.

Adam Ware
Adam Ware

Brent, why did you have to take a great post like this and ruin it with that picture of Rex Grossman? Seriously though, great breakdown of the stakeholders, I love the playing field model. I think your last paragraph about having to educate IT staff is spot on also. Having technical resources dedicated to analytics has a dramatic effect on efficiency and growth of the analytics program. Oh, and by the way, our quarterback situation has been resolved in Chicago... for now. - Adam (@wheresitworking)