Using analytics to improve your PPC performance is a topic that's been done to death. Then Google came and blew our minds by adding the stats straight into AdWords. Here's how to use these to actually make an easy improvement.
How To Get It
For years you've been able to go into Google Analytics and see extra stats at keyword level. It's a no brainer for running your campaigns.
But now Google have introduced the ability to include certain Google Analytics metrics directly into AdWords for extra convenience.
To link Google Analytics to AdWords:
- Make sure you're logged in with an address that is an admin for AdWords and Analytics. Access Google Analytics through the link in AdWords. Go to the Admin section.
- Click on the breadcrumb to ensure that you're at account (not web property or profile) level.
- Go to the "Data Sources" tab. Link the accounts here.
Now it’s time to customize your columns in AdWords:
- These new columns aren't enabled by default, so click the "Customize Columns" button and go to the Google Analytics section.
- Add the three new columns available. Do this for campaigns, ad groups, ads and keywords.
Be patient. This is a gradual rollout, so if these columns aren't available to you yet just wait a few days.
If the columns are available but no data is showing up, keep waiting. The data will be processed and populated soon.
Bounce rate is a fairly straightforward metric. It is the proportion of visitors who viewed only one page. These people left after viewing the landing page.
We don't know whether they spent ages reading the page in detail or just hit back as soon as they saw it. Whatever the reason, they never had a "page two".
On most sites a bounced visit is less engaged than a non-bounced visit. This isn't the case for every visit, but is a pretty strong trend.
Reducing your bounce rate is a two-pronged approach: make your page better, and find better traffic in the first place.
Including this metric in AdWords is about the latter. If a keyword has a high bounce rate even though it's going to the same page as a keyword with a low bounce rate, that's a sign that your keyword is on average bringing in fewer people who are super interested in your site.
Pages Per Visit
This is exactly how it sounds, but again you're looking at an average. In this case your average can get pushed up quite a bit.
A mean average means the total number of pages viewed, divided by the number of people. A mode average would tell you the most common number of pages per visit.
This is a mean average, so every visitor who looks at loads of pages will push this number upwards. If you pages per visit average is two, then a visitor who looks at 10 pages has as much impact on your average as eight bounced visits (obviously moving the average in opposite directions). That means it doesn't take many visitors to look at a lot of pages to significantly raise your average number, and some visitors really will look at every page on your site.
For example, say 1 percent of visitors view more than 20 pages on your site (the average is 2.7) but those 1 percent account for 13 percent of your total page views. That 1 percent of visitors has a big impact on your stats.
Visit duration has one major caveat: the final page of a visit isn't counted. For an e-commerce site this may not be a big deal, but for an informational site, blog, or any site where the most common user reads a long page then leaves, this can seriously mess with your numbers.
Why Use These At All?
Conversion is king right? Why can't we just make all our keyword decisions based on conversion rate and CPA targets? Because in any good long tail campaign there are a lot of keywords with very little data.
So long as the total cost of these is low, the occasional conversions they provide will help your campaign to be more cost efficient and allow you to spend more, achieve more volume, and all those good things.
How do we judge those low volume keywords? By using figures that we have reason to believe can act as a proxy for conversion likelihood.
If keyword A has 10 clicks in a month it may be unreasonable to expect a conversion, so we might have no data whatsoever to tell us how users searching for that keyword behave on our site. But if we can see their bounce rate we know more.
Bounce rate is notconversion rate, or even a causal factor in conversion rate. But the two can be related.
If a group of users is more likely to convert, it's not unreasonable to work on the assumption that same group of users might be less likely to bounce. Each bounce implies no conversion, so a high bounce rate is often going to be correlated with a lower conversion rate.
This helps bid management users (and vendors) a lot. This extra data isn’t as good as conversion data. But it has a value.
By computing the correlation between one of these metrics and your conversion rate you can get a better indication for low volume keywords of what sort of bid is going to be suitable.
Don't weight this data as highly as historical conversion data or other factors you're currently using. But whatever you do, don't ignore it.
What Does it Mean to Me?
Many of you won't be using bid management (there's no excuse for that since AdWords introduced their automation tools!). If you're not, then you have extra steps in your daily campaign management tasks.
Where you were previously looking at cost per keyword, comparing against traffic volumes and conversion rates, you're now going to need to add these metrics into the mix. I can't tell you which one is most important. The type of site you run and how users behave will determine that.
Bear in mind the limitations of each:
- If the final page is the one they'll spend most time on, steer clear of Visit Duration.
- If visitors can complete their actions all on the first page, don't use Bounce Rate.
- If a significant number of users will look at looooads of pages at a time (e.g., forums) don't use Pages per Visit.
Once you get these metrics in your account, it's time to go explore. Look for high volume keywords with unusually high or low values in these metrics, and have a good hard think about why.