Recently, I ran into an interesting change in my account. Interesting seems like a strange word to describe a change in performance, but, in this case, interesting was an accurate word.
The “Interesting” Change
For months, conversion rate had been steady at 8% with some natural variance that is always present in paid search. Then all of sudden conversion rate dropped by 3 percentage points.
You would think a drop of conversion rate by 3% would cause horrific shifts in cost-per-conversion or at least a decrease in spend. However, this was not the case as spend and conversion volume remained steady, even with the decrease in conversion rate. What happened?
If this sounds like the setup to a riddle, it kind of is. When there are no new campaigns, constant bid changes and the same website landing pages, where does one look first when conversion rate changes?
Going back to the basic formula for conversion rate (conversions/clicks), we know that something must have changed with the clicks if conversion volume has remained constant.
Cost-per-click, which had been averaging close to a $1.50 over the last 5 months, had suddenly dropped a dollar to average around $0.50 for the next 2 weeks.
The Solution To The Riddle
Cost-per-click is calculated by ad rank. Ad rank ties back to competitor ad rank, quality score, ad extensions and the max bid.
As a result, there are several items that can impact my cost-per-click:
- Change in competitors
- Change in quality score
- Change in ad extensions
- Change in bids
When calculating cost-per-click, you pay $.01 cent more than your competitor with the closest ad rank. If a competitor has stopped advertising and no longer competing for the same advertising space, cost-per-click will decrease. This can be viewed in the auction insights tab of Google AdWords and looking at competitive metrics, such as impression share and impression share lost to rank. However, this was not the case in my situation.
Quality score is determined by expected click-thru-rate, landing page relevance and ad relevance. There is no way to view historical quality score in the AdWords interface. While one could potentially look through old keyword reports and use VLOOKUP to determine any shifts in the quality score before the cost-per-click decrease, there is a simpler way to judge if it has changed. Ask the following questions:
- Has the website undergone any updates?
- Have there been any noticeable changes to click-thru-rate?
- Have new ads been uploaded recently?
Because I can answer a definitive no to all these questions, a change in quality score can be ruled out as the cause to the conversion rate decrease.
Here is the additional fourth option that can tie back to a change in maximum bids, placements on the Display Network. Continual bid changes is an obvious best practice of paid search. We bid up what works and down what isn’t. In the Display Network, increasing bids can translate to increasing the scope of websites that our ads can show.
The largest decrease in cost-per-click was from two Display campaigns and one Search campaign. For the Display campaign, new placements appeared where they had never been seen before. Average cost-per-click for those placements was $0.05. While $15 isn’t a large amount of spend in a week, $15 with a resulting 300 additional clicks that do not result in any conversions can really impact the overall conversion rate for the campaign and the account.
The Action Taken
In a situation like this, the underlying question is, “Should we be concerned with a decrease in conversion metrics when all the important metrics have remained steady?” We should be concerned when we are doing any type of testing that ties to conversion rate. This can be landing page tests or ad testing.
For me, I pulled a placement report and searched for any placement that had an average cost-per-click of $0.05, had spent over $15 in the last 6 months and had never had any conversions. These placements were added as exclusions to my campaign. While these placements weren’t adversely hurting the performance numbers that mattered, they weren’t helping to bring qualified traffic to the site. After excluding these low CPC placements, conversion rate returned to typical levels.