Does the suc­cess or fail­ure of your paid and nat­ural search ini­tia­tives cre­ate more ques­tions?  Do you want to under­stand bet­ter how vis­i­tors react to your site enter­ing from exter­nal search?  Do you want to know how to improve your site to take bet­ter advan­tage of the traf­fic you are dri­ving to your site from paid search?  There is no magic bul­let, but I hope to give you the skills to bet­ter under­stand how paid search is affect­ing your vis­i­tor behav­ior, so that you can test dif­fer­ent cre­atives and land­ing pages.

If you are new to Dis­cover, you may won­der, “Where are my Search Key­words – Paid and Search Key­words – Nat­ural reports?”  To get these reports you need to first load the Paid Searches report in Dis­cover, which is found under the Traf­fic Sources menu.  From here we can break down paid or nat­ural by key­word to see our paid or nat­ural keywords.

What is the impact of all my paid search ini­tia­tives on my business?

Please note: If you have set up your paid search to be recorded in the cam­paigns vari­able and you have uti­lized SAINT clas­si­fi­ca­tions to roll up all paid search cam­paigns then you can answer this ques­tion in SiteCatalyst.

  1. Drag the Orders and Vis­its met­rics to the Paid Searches report.  Note: you can select any met­ric that you would con­sider a key suc­cess on your site.
  2. Cre­ate a cal­cu­lated met­ric for con­ver­sion rate.  In this exam­ple I am going to define it as “Orders / Visit”.  If you are unsure of how to cre­ate a cal­cu­lated met­ric in Dis­cover, you can ref­er­ence sec­tion 6.5 of the Dis­cover User Manual.

This report quickly illus­trates to me the value of my paid search ini­tia­tives in dri­ving a sig­nif­i­cant amount of traf­fic to my site (21% of total vis­its) and con­vert­ing those vis­i­tors.  In fact, I see now that my paid search ini­tia­tives are con­vert­ing vis­i­tors at higher rate than the site aver­age (13.19% vs. 10.35%).  After learn­ing of the stel­lar per­for­mance of paid search, com­pared to other chan­nels, you may ask…

Which of the key­words that I pur­chased are con­vert­ing the highest?

Not every key­word is cre­ated equal and we can get a bet­ter idea of which paid search key­words should receive some TLC, or if cer­tain key­words should just be dropped.

  1. Click on the green plus sign next to Paid to break down paid search by search keywords.
  2. Sort the Search Key­words by Order by click­ing on the arrow in the sub-column header.
  3. Of the top five key­words (by orders), “jj esquire” is the low­est per­former, con­vert­ing at 12%, com­pared to more generic key­words like “jeans”, which con­verts at 16%.  So now the next ques­tion I have is…

Why is key­word “jj equire” con­vert­ing at a lower rate?

There is a lot of data avail­able to try paint the pic­ture of why one key­word may be under­per­form­ing.  I will try to take you though a log­i­cal flow, but when doing your own analy­sis you may skip a step, if it is not rel­e­vant to your business.

  1. My first thought is maybe there is a dif­fer­ence in where vis­i­tors from the “jeans” key­word and the “jj esquire” key­word are land­ing on the site.  This can be uncov­ered by break­ing down both terms by Paths > Pages > Pages Entry.  Tip: If you have more than one row high­lighted when you click on the green plus but­ton to break­down then all the rows high­lighted will breakdown.
  2. In both instances, the major­ity of vis­i­tors (99% and 83%) are land­ing on the Home Page.
  3. Are more of the vis­i­tors from the “jj  esquire” key­word exit­ing imme­di­ately from the Home Page than vis­i­tors from the  “jeans” key­word?  To answer this ques­tion, drag over the met­ric Sin­gle Page Vis­its and cre­ate a cal­cu­lated met­ric for Bounce Rate (Sin­gle Page Visits/Entries) to add to the report.
  4. Again, we see sim­i­lar behav­ior from the Home Page for both key­words.  A pos­si­ble next step would be to A/B test changes to the Home Page bet­ter tai­lor the user’s expe­ri­ence.  We know that “jeans” con­verts at a higher rate, so I would con­sider using a ban­ner to high­light jeans to help drive vis­i­tors from generic brand terms like “jj esquire” deeper down the con­ver­sion funnel.

You may still be scratch­ing your head, try­ing to under­stand why there is a dif­fer­ence in con­ver­sion between these two key­words.  Per­haps ask­ing the fol­low­ing ques­tions will lead you to the answer you are yearn­ing for.

Since vis­i­tors are not imme­di­ately leav­ing the site at a sig­nif­i­cantly greater rate, are they not con­vert­ing at a higher rate because they can­not find what they are look­ing for?

  1. Cre­ate a Visit seg­ment where Paid Search = Paid and Search Key­word = jj esquire.  Note: the Paid Search and Key­words cri­te­ria need to be defined together in a Page View con­tainer so that we guar­an­tee that “jj esquire” was a paid search.  Then the Page View con­tainer must be nested within the Visit con­tainer so that we can cap­ture vis­i­tor path data in the segment.
  2. Open the Next Page Flow report (Paths > Pages > Next Page Flow) and drag your newly cre­ated seg­ment to the top of the report.
  3. A sig­nif­i­cant por­tion of vis­i­tors nav­i­gates to the Chil­dren & Toys sec­tion and then to Search.  With this greater insight into vis­i­tor behav­ior, you may want to test dif­fer­ent con­fig­u­ra­tions of the Home Page to high­light key areas of inter­est for vis­i­tors enter­ing the site from the paid search term “jj esquire”
  4. In con­trast, when vis­i­tors are com­ing from the paid search team “jeans”, a greater per­cent of vis­i­tors nav­i­gate directly to the Wom­ens sec­tion of the site.  This dif­fer­ence in behav­ior form the Home­page for the “jj esquire” and “jeans” terms fur­ther high­light the oppor­tu­nity to increase con­ver­sion by pro­vid­ing a more tar­geted expe­ri­ence, either through land­ing pages or tar­geted offers.

There are many dif­fer­ent paths for analy­sis to find oppor­tu­ni­ties to opti­mize your paid search and I have only cov­ered a few of them.  You will notice that with my sam­ple data set, the results of the analy­sis did not  yield that golden nugget, excite­ment induc­ing result, but  I hope that this sam­ple analy­sis will inspire you in your paid search opti­miza­tion efforts.  If you have a ques­tion that you would really like be to able to answer, but don’t know how to go about it, feel to post it in the com­ments and we will con­sider it for a follow-up post.


Ideally, you'd be deep-linking your keywords to the proper pages for each campaign, though you can test effectiveness of dropping visitors for certain terms on the home page, a custom landing page, or right into catalog results (in a retail scenario). One critical factor as well (again, for retail) is tying this data back to products. The relative performance of your products for paid/natural and new/repeat segments can tell you if you have an issue with the SKUs you're showing to new visitors you're acquiring. Demographic info (generally only available with Facebook paid ads, that I'm aware of) can be a dramatic help. I'd also recommend adding a search path here. Some segment (if you have enough traffic) of your visitors are going to perform an on-site search. Affinity between external and on-site keywords can be very actionable for SEM efforts. You might also want to look at low searches (you should have a site search event) and high conversion. That means you're dropping the right term in the right spot, so visitors don't need to search. "Don't make me think", right? Anyway, nice post, Laura.