At Adobe’s annual Sum­mit event in Salt Lake City, I hosted an inter­est­ing ses­sion titled “Play­ing to Win — Ensur­ing Your Orga­ni­za­tion Sup­ports Ana­lyt­ics Suc­cess”, which essen­tially focused on the crit­i­cal topic of web gov­er­nance. Dur­ing this ses­sion, I intro­duced a new gov­er­nance frame­work and matu­rity model, which you’ll be hear­ing more about in the com­ing months. I also had the great oppor­tu­nity to present with Andrew Carr from Ora­cle, who shared some valu­able insights into some of the chal­lenges his orga­ni­za­tion faced with its web ana­lyt­ics pro­gram and how his team was able to over­come var­i­ous issues.

If that was all that hap­pened in this ses­sion, I would have been sat­is­fied with what we were able to share and accom­plish. How­ever, another key com­po­nent of this web governance-focused ses­sion was the round­table dis­cus­sions that hap­pened after our pre­sen­ta­tions between Sum­mit atten­dees. In groups of 4–8 peo­ple, vol­un­teer mod­er­a­tors led dis­cus­sions around var­i­ous web gov­er­nance top­ics, includ­ing lead­er­ship, strat­egy, peo­ple, process, and product/technology. Just like last year when we facil­i­tated a sim­i­lar round­table dis­cus­sion for­mat, sev­eral par­tic­i­pants approached me after­wards express­ing their appre­ci­a­tion for such an insight­ful and use­ful session.

Now before you start curs­ing that you weren’t able to par­tic­i­pate in this valu­able Sum­mit ses­sion, I’d like to share some of what these dif­fer­ent groups dis­cussed like I did last year. Ten dif­fer­ent mod­er­a­tors pro­vided me with their groups’ key take­aways. I’ve taken on the role of editor-in-chief and have sum­ma­rized their key dis­cus­sion points as follows:

“Win­ning“

  • If you want to take your web ana­lyt­ics to the next level, sev­eral groups men­tioned that you need to demon­strate value (wins, results, etc.) to the organization.
    • If you’re dri­ving value for your com­pany, you will need to move beyond just focus­ing on reporting.
    • A few groups rec­om­mended out­sourc­ing all reg­u­lar report­ing or automat­ing reports to free up ana­lysts’ time for more mean­ing­ful projects.
    • Look to part­ner with a will­ing inter­nal group on a project that could help drive some momen­tum internally.
  • Mul­ti­ple groups men­tioned how dif­fi­cult it can be to work with mul­ti­ple busi­ness units. In order to be suc­cess­ful, you need to man­age your stake­holder groups effectively.
    • Be proac­tive and don’t wait for their requests to come to your team.
    • Sched­ule a reg­u­lar meet­ing with them to align pri­or­i­ties and review progress.
    • Assign dif­fer­ent team mem­bers to man­age dif­fer­ent stake­holder groups.
    • Ensure you have solid, well-documented busi­ness processes in place.
    • Add project man­age­ment staff (PMO) to your program.

Strat­egy

  • A cou­ple of state­ments stood out: “In gen­eral, peo­ple do not know what the goals are” and “Hard to get [the strat­egy] because no one agrees on what the website’s focus should be.
  • Mul­ti­ple groups rec­og­nized the impor­tance of clar­i­fy­ing the organization’s goals as it directly impacts the qual­ity and rel­e­vance of the data pro­vided by the implementation.
    • One group men­tioned that the lack of strat­egy on the ini­tial imple­men­ta­tion has con­tin­ued to plague and impede their progress.

Lead­er­ship

  • Many peo­ple expressed frus­tra­tion in not hav­ing a “true” exec­u­tive spon­sor or cham­pion.
    • You need some­one who has author­ity and influ­ence within the orga­ni­za­tion but is also both involved and com­mit­ted to help­ing the com­pany become more data-driven.
  • A lack of top-down account­abil­ity was also men­tioned as a key stum­bling block.

Struc­ture

  • Sev­eral groups debated age old ques­tion of where the ana­lyt­ics team should sit.
  • Most peo­ple felt IT was the least favor­able option.
    • Aside from the own­er­ship of the ana­lyt­ics team, IT’s involve­ment was referred to as a bar­rier and peo­ple cited col­lab­o­ra­tion issues in work­ing with IT.
  • Some peo­ple felt that hav­ing the ana­lyt­ics team within the mar­ket­ing group was ideal because com­mu­ni­ca­tion and coor­di­na­tion was much eas­ier between all of the stake­hold­ers when they were mostly within the same mar­ket­ing organization.
  • Other par­tic­i­pants favored an inde­pen­dent ana­lyt­ics or cus­tomer insights team, which was out­side of mar­ket­ing, IT, or any other group to pro­vide an unbi­ased view of the data.
  • Wher­ever the group sits, be care­ful about the nam­ing of the group because it can shape the per­cep­tion of its importance.

Train­ing

  • Most groups saw the impor­tance of for­mal train­ing but didn’t have the band­width to pro­vide it.
  • One group rec­om­mended requir­ing train­ing to be com­pleted before grant­ing access to the tool or data.
  • Par­tic­i­pants found users don’t log into the tool for weeks and then become frus­trated when they are not able to find what they need.
    • Set up a weekly or bi-weekly meet­ing for users to answer spe­cific ques­tions they have about the tool or data.
    • Offer one-on-one or small group infor­mal train­ing ses­sions as needed.

Data

  • Orga­ni­za­tions need to trust their data or else it will not be used, the ana­lyt­ics team will lose sup­port for main­te­nance or upgrades, and even­tu­ally the tool will be viewed as “broken”.
    • Imple­men­ta­tion is an ongo­ing process (not “once and done”) as new online ini­tia­tives are intro­duced and cur­rent ones evolve.
    • Peri­od­i­cally audit the reports to deter­mine if they are still rel­e­vant or need to be enhanced.
    • Ensure all imple­men­ta­tion projects include a data val­i­da­tion stage.
    • Secure ade­quate lead times so that all new pages/sites/apps/campaigns can be thor­oughly tested before being launched.

Com­mu­ni­ca­tion

  • Mul­ti­ple groups iden­ti­fied com­mu­ni­ca­tion as being a key suc­cess factor.
    • There needs to be good com­mu­ni­ca­tion between all stake­hold­ers as changes imple­mented by one group can impact other teams’ met­rics and reporting.
    • One par­tic­i­pant shared a painful expe­ri­ence where his team’s KPIs dis­ap­peared with­out warn­ing one day when another team decided to change their imple­men­ta­tion and metrics.
  • Doc­u­men­ta­tion can be chal­leng­ing to keep updated, but with­out it your orga­ni­za­tion may be vul­ner­a­ble if key peo­ple leave.
    • Each report suite should have its own doc­u­men­ta­tion and each vari­able should have at last one sen­tence detail­ing why the vari­able was set and what it is used for.
    • Cir­cu­late an inter­nal white paper on what all of the met­rics mean to pro­vide greater transparency.
    • Pub­lish a reg­u­lar newslet­ter for your inter­nal ana­lyt­ics com­mu­nity and key stake­holder groups.
  • Dif­fer­ent groups shared var­i­ous frus­tra­tions with the busi­ness, indi­cat­ing that there are still some com­mu­ni­ca­tion oppor­tu­ni­ties to clear up inter­nal misconceptions.
    • Busi­ness doesn’t under­stand why the web data doesn’t match data within other sys­tems and why direc­tional data is useful.
    • Busi­ness lacks appre­ci­a­tion for the imple­men­ta­tion and update process for data collection.
    • Busi­ness fails to under­stand how much time it takes to answer analy­sis requests.

As you can see there were a lot of good insights stirred up dur­ing this ses­sion, and this sum­mary only rep­re­sents a por­tion of what was shared and dis­cussed. For many of the ses­sion par­tic­i­pants, it was great to sit down with indus­try peers who are expe­ri­enc­ing sim­i­lar chal­lenges in their orga­ni­za­tions (you’re not alone!) as well as inter­act­ing with more sea­soned prac­ti­tion­ers who have already over­come some of those chal­lenges (there’s hope!). If you par­tic­i­pated in the round­table dis­cus­sions and one of your key take­aways wasn’t cap­tured in the above sum­mary, please feel free to add a com­ment. A big thanks to all of the mod­er­a­tors and par­tic­i­pants who joined the conversations!

  • http://wegmans.com Cleve Young

    Another issue which we ran into dur­ing the our last imple­men­ta­tion was the data struc­ture was com­pleted by Mar­ket­ing and IT, with help from a con­sul­tant, and was never reviewed by the ana­lysts in a timely man­ner. By the time they got around to involv­ing us the process was so far along it was too late to make many mean­ing­ful adjust­ments. I had no prob­lem with the con­sul­tant com­pany as they can only go by what they are told. The issue is too often Mar­ket­ing gives broad sim­plis­tic busi­ness require­ments which IT will struc­ture with­out really under­stand­ing what is required to effi­ciently ana­lyze the data to answer those ques­tions. Now, we as the ana­lysts spend far too much time mak­ing adjust­ments and workarounds to com­pen­sate for the poor data which leaves us less time to do mean­ing­ful and rel­e­vant analysis.

    Bot­tom line, make sure the analy­sis team is involved through­out the entire process, not just at the end where it is often too late.

    Regards,
    Cleve

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

    Great insights, Cleve. I sec­ond the motion that ana­lysts need to be involved dur­ing the entire process, espe­cially upfront dur­ing the devel­op­ment of the mea­sure­ment strat­egy and also dur­ing the data val­i­da­tion phase.

    Brent.

  • http://www.target.com Wendy Krautkramer

    Great arti­cle and great tips! Use­ful as well for new forms of ana­lyt­ics being pur­sued with exist­ing tools (ex. adding per­son­al­iza­tion to a site, etc..).

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

    Wendy,

    I’m glad you found it use­ful. Thanks goes to all who shared their wisdom.

    Cheers,
    Brent.