Marketing attribution – Taking the guesswork out of marketing

Analytics

Today, the typ­i­cal cus­tomer jour­ney spans mul­ti­ple chan­nels, from brows­ing online and buy­ing in-store, engag­ing on social media net­works to more tra­di­tion­al offline media. With so many touch points now in the mix and the cus­tomer jour­ney con­stant­ly evolv­ing, gain­ing true cus­tomer-based insight on the impact of your mar­ket­ing spend has become a com­plex process.

A blue­print for under­stand­ing the cus­tomer con­ver­sion jour­ney

Mar­ket­ing teams are recog­nis­ing the need to get a 360° view of their cus­tomer, to be able to mea­sure their jour­neys and to eval­u­ate the impact and influ­ence of the actions they make at each phase.

How­ev­er, there remains wide­spread uncer­tain­ty amongst mar­keters about exact­ly what should be mea­sured, what mod­els should be used and how to take action based on the data-dri­ven find­ings – such as how to allo­cate mar­ket­ing spend more effec­tive­ly based on the true val­ue of each chan­nel.

Accord­ing to the 2014 State of Cus­tomer Ana­lyt­ics Sur­vey by Forrester/Burtchworks, only 54% of Mar­keters cur­rent­ly use or are pilot­ing Attri­bu­tion Mod­els. Anoth­er sur­vey by CMO.com found that few­er than 20% of respon­dents felt ful­ly con­fi­dent in their abil­i­ty to mea­sure over­all cam­paign effec­tive­ness, allo­cate bud­get with ROI in mind and com­mu­ni­cate per­for­mance to oth­er C‑Level exec­u­tives. Too many chan­nels and too much data can make it dif­fi­cult to cap­ture, mea­sure and attribute con­ver­sions appro­pri­ate­ly…

Move beyond the lim­its of ‘last click’

Tra­di­tion­al­ly, mar­keters have reg­is­tered the impact of each chan­nel accord­ing to the max­im ‘last click wins’ to answer these ques­tions.

This method is still very much applic­a­ble, but a new and more advanced way of mea­sur­ing the true influ­ence pro­vides a far more accu­rate and holis­tic pic­ture of the cus­tomer jour­ney and what impact your mar­ket­ing bud­get is real­ly hav­ing.

Attri­bu­tion mod­el­ing allows you, the mar­keter, to quan­ti­fy the influ­ence that over­lap­ping touch points and chan­nels have in the over­all con­ver­sion process, as well as reveal­ing the non-con­vert­ing paths. With this view, the abil­i­ty to ful­ly opti­mise your media mix and spend becomes a real­i­ty.

Build a clear, com­pre­hen­sive pic­ture

The typ­i­cal objec­tives that mar­keters have in their sights include:

  • Mov­ing away from focus­ing on one area of mar­ket­ing that is eas­i­est to mea­sure and ignor­ing oth­ers by pro­vid­ing a full pic­ture, with all offline and online mar­ket­ing data in one place
  • Gain­ing fresh insights into which aspects of the mar­ket­ing mix are starters, play­ers or closers in the con­ver­sion process and which chan­nels influ­ence oth­ers
  • Hav­ing the knowl­edge to stream­line busi­ness deci­sion-mak­ing and mit­i­gate risk in often chaot­ic mar­kets
  • Under­stand­ing the cus­tomer jour­ney and cre­at­ing a smoother, ‘stick­i­er’ expe­ri­ence by push­ing more suc­cess­ful chan­nels where appro­pri­ate and pos­si­ble

Pre­dict­ing the future with accu­ra­cy

Armed with this under­stand­ing, mar­keters aim to gain bet­ter return on their mar­ket­ing spend with improved lev­els of engage­ment and rev­enues or to contain/reduce their mar­ket­ing spend while retain­ing the same lev­els of engage­ment and rev­enues.

Algo­rith­mic attri­bu­tion mod­el­ing replaces a leap in the dark with the tools to run ‘what if’ sce­nar­ios.

Mar­keters can instant­ly see the impact of turn­ing off or reduc­ing the spend with­in a sin­gle chan­nel, divert­ing cus­tomers to a dif­fer­ent chan­nel or oth­er­wise adjust­ing mar­ket­ing activ­i­ty.

In an oper­at­ing cli­mate where bud­gets are often pared to the min­i­mum, pre­dic­tive attri­bu­tion mod­el­ing is a great resource for organ­i­sa­tions look­ing to opti­mise their cor­po­rate mar­ket­ing strat­e­gy and, at a more gran­u­lar lev­el, strength­en the impact of indi­vid­ual mar­ket­ing ini­tia­tives.

Get­ting start­ed

Embark­ing on intel­li­gent attri­bu­tion mod­el­ling for the first time can seem daunt­ing, but there are tools and sup­port read­i­ly avail­able to help you.

It is essen­tial to have the fol­low­ing in place:

  • The right tech­nol­o­gy that is flex­i­ble enough to cap­ture data from all the mul­ti­ple chan­nels by which your cus­tomer inter­acts with you
  • The ana­lyt­ics to explore the end-to-end jour­ney from the customer’s per­spec­tive, with data stored cen­tral­ly, and the capa­bil­i­ty to apply weight­ings and rel­e­vant rules based on your busi­ness needs
  • A flex­i­ble algo­rith­mic attri­bu­tion tool that offers a port­fo­lio of dif­fer­ent mod­els to manip­u­late the data and take the guess­work out of rethink­ing your mar­ket­ing strat­e­gy and pro­grammes

Every jour­ney starts with a first step…

Start with a real­is­tic phased strat­e­gy and select one or two chan­nels to col­lect data from for attri­bu­tion mod­el­ling. Many organ­i­sa­tions scru­ti­nise their online chan­nels first as they are more than like­ly already tracked and more eas­i­ly avail­able.

  • Start mea­sur­ing the con­tri­bu­tion of these chan­nels and explore for the unknowns. Don’t under­es­ti­mate the pow­er of attri­bu­tion mod­el­ing to sur­prise. The tried and test­ed chan­nel you have been using for years may be revealed as less influ­en­tial than you expect­ed in dri­ving engage­ment and sales.
  • Start tak­ing action and look to make incre­men­tal changes and keep test­ing until you find the right mix to dri­ve sales and max­imise bud­get.

Attri­bu­tion mod­el­ing is also not a one-off event. Rather it is a con­tin­u­ous process that will be influ­enced by your organisation’s own mar­ket­ing plan­ning cycle and the con­stant evo­lu­tion of the ways your cus­tomers engage with you. Along the way, you’ll make mis­takes, but will learn from them, too.

To find out more about how Attri­bu­tion can help your busi­ness, come along to “Under­stand how well your mar­ket­ing is actu­al­ly per­form­ing with Attri­bu­tion” at Adobe Sum­mit on Thurs­day morn­ing. Be sure to reg­is­ter on the web­site or with the Sum­mit app.

Cov­er image © apops — Fotolia.com


Analytics
Michael Plimsoll

Posted on 04-22-2015


Join the discussion