PPC Campaign Optimisation – Which Data Should You Use?

Choos­ing the right data to use when opti­mis­ing your PPC cam­paigns is not straight for­ward. Some­times your choice is made eas­i­er by being restrict­ed to only using a cou­ple of met­rics, as that is all that can be tracked, but fre­quent­ly, in this age of Big Data you have more met­rics than you know what to do with. In this blog, we’ll explore some of the things to be con­sid­ered when you go about choos­ing your metrics.

Cam­paign Goals

The first and most obvi­ous ques­tion to ask your­self about your objec­tives is what is the ulti­mate goal of your PPC cam­paigns? For online retail­ers this is fre­quent­ly sim­ply to dri­ve online sales, but this is only the begin­ning. Is the num­ber of sales the most impor­tant thing? Or is it rev­enue from those sales? Should you be look­ing at the mar­gin or prof­it which is made on each sale? Are you inter­est­ed in dri­ving sales from return­ing cus­tomers, or do you want to focus on sales for new cus­tomers? The answers to these ques­tions will be dif­fer­ent for dif­fer­ent adver­tis­ers, but should serve as a good start­ing point for adver­tis­ers from all dif­fer­ent sec­tors, not just retail.

Enrich­ing Data

Anoth­er key area to look at is the amount of data that you have to opti­mise with against your key­words. If your prod­uct or ser­vice that you are sell­ing is a con­sid­ered pur­chase, or of a high val­ue, which involves a lot of research, then you will prob­a­bly find that you don’t have many orders per key­word per day/week. This presents a dif­fer­ent chal­lenge as you can’t rely sole­ly on data from the final trans­ac­tion in order to opti­mise your PPC cam­paigns. You will need to start fac­tor­ing in engage­ment or indi­ca­tor met­rics. These are the actions that a user takes on site dur­ing the research/consideration phase, which show they are more like­ly to com­plete the final trans­ac­tion in the future. These should obvi­ous­ly car­ry a low­er weight­ing than the final trans­ac­tion, but will prove cru­cial in pro­vid­ing data to key­words which dri­ve clicks but no final orders. Is that spend wast­ed, or is it nec­es­sary to secure orders fur­ther down the line?

Brand Aware­ness

The two areas above have focused on direct response adver­tis­ers, where there is a clear end goal as an objec­tive. But what about if the adver­tis­er doesn’t sell any­thing through their web­site, instead run­ning PPC cam­paigns to bring traf­fic to site for brand aware­ness? You should still have objec­tives for your cam­paigns, and clear KPIs, but what should they be? Tra­di­tion­al­ly these KPIs have focused on impres­sions (or impres­sion share) and CPCs, ignor­ing what actu­al­ly hap­pens once the user has clicked on an ad. More advanced strate­gies look at on-site engage­ment and inter­ac­tions, such as page views, time spent on site, bounce rate, brochure requests, video views etc. The focus here is not just bring­ing traf­fic to site, but ensur­ing the vis­i­tors are qual­i­ty users, who show a high brand engage­ment. Addi­tion­al­ly, the focus of dif­fer­ent cam­paigns may be dif­fer­ent, accord­ing to the cam­paign type. For instance, if you are bring­ing some­body to site who has searched for a competitor’s term, do you have a dif­fer­ent user jour­ney and KPI in mind to some­body who has searched for your own brand term?

I’ve posed many ques­tions through­out this post which may be use­ful start­ing points when think­ing about your cam­paign objec­tives. I’ve not pro­vid­ed the answers though, as those will be dif­fer­ent for every adver­tis­er. If you’d like to hear more about how we, at Adobe, are help­ing our cus­tomers tack­le these ques­tions, includ­ing hear­ing first hand expe­ri­ences from a lead­ing Adobe Media Opti­miz­er cus­tomer, then come along to the ses­sion which I am co-pre­sent­ing at Adobe Sum­mit in Lon­don. The ses­sion is enti­tled Big Data vs Smart Data: Using ana­lyt­ics data to dri­ve adver­tis­ing ROI, so I hope to see you there!

Cov­er Image© aaab­bc — Fotolia.com — Foto​lia​.com