Can relying on interface data, specifically lead numbers and CPL significantly hurt your PPC accounts’ performance? The following is a case study that sheds light on exactly this, and will show you why I have cut down on the time I spend looking at that data in the AdWords and adCenter interfaces.

The set up:

I recently was put on an account that has a history of up and down performance. This particular client has several sub-accounts we manage, and some of these accounts have conversion tracking, and others do not have conversion tracking. I was tasked with optimizing two of these sub-accounts. One of which had conversion tracking, and one that did not.

As most PPC managers would be, I was eager to jump into the account with tracking, and hesitant regarding the account without tracking.  But I jumped in with both feet with the goal of having the campaigns exceed goals within 30-days. My primary KPI (key performance indicator) was total lead volume.

I had a specific goal for each, and while CPL was important, the bottom line was hitting this lead number goal. Below are these goals (to protect our clients, all numbers used in this post are examples.)

PPC Case Study Goals

First Steps:

Both accounts received some structural reorganization. This included keywords being moved to more relevant, tightly themed ad groups, and adding a campaign to each with some generic terms that related to the service I was advertising but not as highly relevant as other keywords I already had in the campaign. I also conducted an ad review in each, and paused underperforming ads and added new ads that I hoped would increase my CTR’s.  Then…

For the account with conversion tracking…

I adjusted bids on keywords based on keyword level CPL and average position. I knew that an average position of around 2 worked really well with this account, so I adjusted my Max CPC in an attempt to get there. One thing I was cautious of was not hurting keywords that were performing really well for me. So if something had generated more than 4 conversions, I left it alone.

For the account without conversion tracking…

I adjusted bids based on an ideal CPC. This is a little trickier than what I did for the account with conversion tracking but here’s a breakdown of what I did.

I figured what my average conversion rate was over the past 30-days (the client sends raw lead data weekly, so I do know total leads, just not real time, and the data isn’t in the interface.) With that average conversion rate in hand I determined what my average CPC would need to be based on dividing my total lead goal by that average conversion rate. Below is a breakdown (the numbers are examples, not actual):

(For reference, the yellow fields were ones where I manually entered data.) You can pretty easily see that by doing this, I found that my average CPC was running $1.50 higher than my goal (goal was $7 and MTD was $8.50.) This means that unless I drastically increased my conversion rate, there was no way for me to hit my lead goal and stay within budget. And on this particular account, I have no access to the landing pages, so I did some bid changes on keywords that would make the needed 21% decrease to my average CPC.

Second Round of PPC Optimizations

For the account with conversion tracking…

I identified my top performing campaign. I adjusted the budget allocation to make sure it was receiving as many clicks as possible. I also found a few keywords that had bids below top of page estimates. Because I could tell these keywords performed well for me, I increased the bids.

I also increased the bids on every keyword that had converted at CPL goal for the last 30-days to try to get to an average position on those keywords of 1.5. Lastly, I added a few negative keywords I found through an SQR, and added about 20 new keywords that I found through research.

For the account without conversion tracking…

I identified the best campaigns and adjusted the budget allocation. I did a quick SQR and found a handful of negative keywords to ad. And I did about an hours worth of keyword research and found 20 keywords to add that I thought would help me hit my lead goal.

Then I compared my new MTD numbers (specifically CPC) to my goal numbers. I was closing the gap. So fearing I would make optimizations that would cut leads I left everything else untouched.

Third Round of PPC Optimizations

For the account with conversion tracking…

Spend was projected to be over my budget at this point and lead CPL was 20% high. I pulled back bids on keywords that had not converted in the past 90-days but that had spent at least $250. I also pulled back bids on keywords that had converted once in the past 90 days but had spent at least $500.

I also adjusted my budget allocations to restrict those campaigns with the lowest conversion rates. Then I looked at the ads I wrote about 45 days earlier and paused any underperformers based on CTR and conversions. I also adjusted some bids downwards on keywords that had performed okay over the past 90-days but that were above CPL goal.

For the account without conversion tracking…

Average CPC was now right in alignment with my goal CPC. Average conversion rate was still below the goal at 3.4%, and total projected leads were under lead goal by 10%. With 10 days left to make up this ground it was clear I needed to do something right away.

With lead data in hand, I left all keywords that had converted in the past 30-days alone. Then I reduced bids on the other keywords by 10%. I chose this 10% because I was about 2.5% below my conversion rate goal. So being 4x’s that difference I felt would make a significant different in average CPC to make up for it.

I took a quick look at the ads I wrote earlier in the month, and saw a few that had a CTR that was drastically lower than average, so I paused those. Then once again, not wanting to optimize blind, I left everything else alone.

PPC Lead Results over 30-days

When my 30-day run concluded, the results were in, and I had logged my projected hours, it was obvious that the account without conversion tracking outperformed the one that had it. And, I spent roughly 30% less time in the winning, non-conversion tracking account.

PPC Case Study Results

Wrap Up

The title of this post is a little misleading; I did use lead data to optimize the winning account. But it wasn’t through the interface and it was drastically less than the account with conversion tracking.

But the fact is, I was over-optimizing the account that had conversion tracking which lead to worse performance overall. It also made is o I didn’t really learn anything about the account over the 30-days. I didn’t learn the best CPC, because I kept changing bids, I didn’t learn much about ads, because I didn’t let them run long enough, and I didn’t learn the best average position to be in, because again, I changed bids too much.

For the account that didn’t have tracking I focused more on major changes I could make and let the incremental changes play themselves out. This resulted in a vast month-over-month improvement, and I learn a lot that I could apply to the following month to keep things rolling on the right direction.