Case Studies

Rescuing a pay per click (PPC) campaign

SUMMARY: After working for three months on an existing pay per click campaign, Sparkdog reduced spending for the client's pay per click visitors and attracted traffic that spent 232.93% more time on the site with a lower bounce rate - 35.74% lower - than the traffic the client had been receiving. Along the way we fixed the conversion tracking features so that the client could, for the first time, know exactly which visitors became customers.

CASE STUDY: "We need someone to take care of our pay per click campaigns, and I wanted to see if you'd be interested." I was interested. During my first login to their Analytics, I saw this, covering the month of January:

Ouch. Line 2. A 72.30% bounce rate for paid traffic. And they were paying for nearly as much traffic as the site had from organic searches. Pay per click visitors can be less valuable than organic searchers but here they are a lot less valuable - they leave immediately 3 out of 4 visits (bounce rate) and the one who stays reads less than half as many web pages as a comparable organic searcher. So I clicked deeper into the reports to see what's going on with those 971 pay per click guests. How much was this costing my client?

  Impressions CTR Cost Conv Rate Cost/Conv Conversions
1. 1,428,576 0.10% $533.81 0.00% $0.00 0

Hmm. Not good. They have a lot of ad impressions, a very low click through rate, and they are spending money for those clicks at a rate of around $500 a month. I can also see, from other reports, that their spending will rise throughout the Spring each year. And no one is converting? Either their conversion tracking is broken or they are attracting the wrong people. What are they paying for, exactly? Here is a report of the top five PPC search sources.

  Keyword Visits Pages/Visit Avg Time on Site % New Visits Bounce Rate
1. (content targeting) 559 1.60 00:00:37 98.03% 85.51%
2. [Search term A] 161 1.80 00:01:00 98.14% 67.70%
3. [Search term B] 62 3.26 00:02:19 90.32% 58.06%
4. [Search term C] 21 2.81 00:01:14 100.00% 57.14%
5. [Search term D] 15 2.07 00:00:29 93.33% 60.00%

Well, Line 1, they're paying for a lot of traffic from Google's content network. I've edited out the other exact search terms to protect my client [My edits are in brackets]. So we have more trouble - 56% of their pay-per-click traffic is from Google's content network (read: not people searching using the regular search engine) and those visitors bounce 85.51% of the time. And other bad news: the other four terms they are paying to attract clicks from have high bounce rates and spend only modest time on the site. Of the four, search term B looks to be most valuable traffic, judged by the data in this table (Line 3). Search term A looks least valuable, even though it generated the second most visits (line 2)!

After talking with the owner, we settled on a sensible strategy: reduce spending and reduce traffic while trying to improve the quality of the remaining traffic (What good is traffic that you pay for but can't convert? Not much). First, I stopped showing ads to the content network in order to focus on search traffic. This is an easy step that everyone should take until they have a handle on search-only traffic. In this case, had we thrown away the January content network traffic, total site traffic would have dropped by 15.6%, but since those 15.6% of total visitors weren't even staying on the site, it's easier to stomach. Second, I began going through the ad campaigns, line by line, and making sure relevant ads were showing to each search term. After three months and many changes, here are the results. I'm comparing April time periods one year apart:

  Source/Medium Visits Pages/Visit Avg. Time on site % New visits Bounce Rate
1. google / organic
  Apr 1 - Apr 30 944 4.48 00:04:13 58.37% 34.96%
  [April, year later] 1,201 4.82 00:04:54 53.62% 30.39%
 
Change
 27.22% 7.46% 16.01% -8.13% -13.06%
2. google / cpc
  Apr 1 - Apr 30 1,516 2.06 00:00:46 95.91% 74.41%
  [April, year later] 778 3.57 00:02:35 89.85% 47.81%
 
Change
 -48.68% 73.12% 232.93% -6.32% -35.74%

Now that's more like it. Organic traffic is notably improved and PPC traffic is dramatically different. Though we have half as much PPC as we did a year before, the newer traffic is much more valuable - they stay on the site 232.93% longer, view 73.12% more pages, and our dreaded bounce rate has been reduced by 35.74%. Not shown in this table is the number of PPC ad impressions, which dropped by 90.19% during this time period. By analogy, imagine removing 90% of a store's inventory and still selling half as much stuff as when the store was full. This is a good result, and indicates that we mostly got rid of the "tire-kickers" and other traffic that we were paying for needlessly in the past. And those PPC keywords?

1. [Search term E] Visits Pages/Visit Avg Time on Site % New Visits Bounce Rate
  Apr 1 - Apr 30 3 4.67 00:00:15 100.00% 66.67%
  [April, year later] 77 3.68 00:02:43 77.92% 44.16%
 
Change
 2,466.67% -21.24% 960.93% -22.08% -33.77%
2. [Search term G]
  Apr 1 - Apr 30 17 2.00 00:00:47 94.12% 70.59%
  [April, year later] 54 5.17 00:02:08 94.44% 37.04%
 
Change
 217.65% 158.33% 175.50% 0.35% -47.53%
3. [Search term B]
  Apr 1- Apr 30 11 3.00 00:00:48 81.82% 63.64%
  [April, year later] 47 3.23 00:01:33 95.74% 55.32%
 
Change
 327.27% 7.80% 92.37% 17.02% -13.07%
4. [Search term C]
  Apr 1 - Apr 30 23 1.96 00:00:33 95.65% 56.52%
  [April, year later] 40 2.28 00:00:23 97.50% 60.00%
 
Change
 73.91% 16.28% -30.10% 1.93% 6.15%
5. [Search term F]
  Apr 1 - Apr 30 7 5.29 00:02:38 57.14% 28.57%
  [April, year later] 37 3.43 00:01:42 94.59% 64.86%
 
Change
 428.57% -35.06% -35.57% 65.54% 127.03%

This table shows the changes in traffic for the top five PPC search terms in April. These "main search terms" have changed from the ones in the January table. Gone are most of the terms that delivered weak traffic. The top three terms now pull strong traffic that stays on the site to learn about our client's business. One of those three, search term B, was also showing some strength back in January, but now it is delivering even better traffic (line 3).

This case study reflects the progress we made in one Google Adwords account for one client during a three month window of time. If you need help with similar problems, we'd love to help. Contact us today.