The realm of paid search data can be overwhelming. Thousands

The realm of paid search data can be overwhelming. Thousands upon thousands of data points flow in every
second. Click-through-rates (CTR). Cost-per-clicks (CPC). Cost-per-acquisitions (CPA). Conversion rates. Google
provide this information, but what does it all mean? As our clients’ paid search team, it is our job to use our
knowledge and experience to interpret this data and use it to improve their accounts at scale.
If we look back to our economics textbooks, we recall that just because we increase production by a factor of
2 or 3, this does not mean we will see a return on that production increase by 2 or 3. There is a diminishing
effect. It costs more money to increase our output than what we are getting back in return for each extra unit
produced. You may be thinking, “Okay that’s great, but how the heck does this affect my paid search
account?!”
When we bid on a keyword, we are bidding on a position in Google’s search results. In the simplest terms, the
first position goes to the highest bidder, the second position to the second highest bidder and so forth. And
while we all know quality score plays a major role in determining bid rank, for the sake of this discussion let us
assume everyone has the same quality score. This is about output per dollar only.
We could blindly shoot for the top position because we assume the first position gives us the best return
(highest CTR, highest conversion rate, etc.). We sometimes forget that while it may be nice to see ourselves in
that top spot all the time, we are also paying more money for that spot and subsequent clicks. If we show our
ad in a lower position, we can save some money, right? Sure we can, but then overall performance will
decrease, right? Wrong! What if we could pay less, but get more? Let the law of diminishing returns be our
guide.
We started managing the AccuRadio AdWords account in June 2015 and within 3 months we took an account
that was already performing well to an even higher level of performance. We saw an increase in traffic and
conversions with a relatively low CPC but we thought we could still do better. We knew that the site
converted, but could we lower our bids to save some money without sacrificing the overall number of
conversions? There were nearly 25,000 keywords in the account. How do we know which bids to lower
without the risk of hurting the overall performance? The answer is to set up a controlled experiment.
Using the four highest spending non-branded campaigns, which accounted for approximately 27% of the
account’s total spend, we decided to test the effect of applying drastic bid decreases to all keywords in those
campaigns. We launched the first experiment in October, reducing all keyword bids by 50% in these top four
performing campaigns. This affected over 2500 keywords.
Once we began the experiment, we quickly saw average CPC decrease by a significant amount. Total clicks and
impression share decreased in the short term, but we were looking to see if that drop in volume was exceeded
by the gains made in CPA. After only a week of testing, and nearly 63K impressions, it appeared we were
winning. The amount of money we were saving outweighed the decrease in total traffic by a large margin. We
were most excited to discover an increased CTR on keywords that had moved from the first to the second
position in search results.
So we ran several more experiments. Each time we decreased max CPC bids from 25-50% on high volume
keywords, we saw CTRs increase, but upon reaching the 3rd and 4th positions, CTRs did start to trend down.
Amidst these tests, we even started to see higher conversion rates! The reason being a higher CTR is a great
indicator of a higher quality user. After all, you can’t have a conversion without the user being on your site and
a higher frequency of users arriving to the site should result in a higher frequency of conversions. Seeing the
higher conversion rates was therefore not unexpected.
Upon implementing the results from these series of experiments, we kept coming in under the client’s
monthly budget. This allowed us to increase the number of days we were running ads. We expanded into
additional markets, adding more campaigns and more keywords, and the account continued to improve not
only from an efficiency standpoint but was also able to show increases in total volume. As a result of all this, in
Q4 of 2015, total volume and efficiency numbers reached all-time highs by several fold. Upon completion of
January 2016, we were looking at YOY growth of 1800% in converted clicks, 1078% growth in total clicks, 82%
decrease in CPA, and 607% increase in CTR. Very impressive!
Traffic and paid search efficiency weren’t the only big winners as a result of these experiments. Since we had
access to how AccuRadio monetizes their traffic, we could calculate real year-over-year ROI as well. Return on
investment improved every month since our experiment began and we finished January 2016 looking at yearover-year ROI growth of 43 percent!
To this day, AccuRadio’s account performance continues to accelerate. We have been incredibly excited with
the success of AccuRadio’s paid search efforts, but what’s the moral of the story? We all know testing and
experimenting is a staple in paid search. You just don’t know how something is going to perform until you test
it. Use your data. There is a reason they call it “Big Data.”
Above all else, don’t be afraid to look and think outside the box, and don’t be afraid to be unconventional. You
really might come across a situation when less is really more.
2016 Logical Media Group, Inc.