Using Web Analytics for PPC Optimization

Share with:


If you know anything at all about PPC, you know that campaign optimization is one of the most important tasks a PPC manager can perform. Managers spend most of their day tweaking bids, ad copy, campaign settings, networks, keywords, negative keywords, and many other data points within the PPC engines.

Campaign optimization is absolutely essential to PPC success. But if you’re spending all your time in the AdWords and Bing Ads interfaces, you’re missing a big part of the PPC optimization picture.

The Rest of the PPC Story

PPC metrics like CTR, conversion rate, cost per conversion, CPC, etc. are crucial elements that can’t be ignored. This data tells us what is happening with an account: how much traffic it’s generating, how many sales or leads it’s driving, and how much all of that cost. We get a great picture of what is happening.

The problem is, often we don’t know why.

That’s where web analytics come in. Web analytics tell us what PPC visitors did once they arrived at the website.

“Now wait a minute!” you might be thinking. “I’m using AdWords conversion tracking, so I can see conversions! Don’t those happen on the website?”

The answer is absolutely yes. And if you’re not tracking conversions via either the free PPC engine tracking scripts or a third-party tool, then shame on you.

But does that data tell you why someone converted? How many visits to the site did it take for that conversion to happen? What other pages did they view? Were they already a customer making a repeat purchase, or was this their first visit?

PPC conversion tracking can’t answer those questions. But web analytics can.

Key Analytics Measures for PPC

Even the most rudimentary web analytics measures can tell us something about our PPC campaigns. The following metrics can be found in any web analytics program. I’ll focus on Google Analytics, because it’s so ubiquitous – but you certainly don’t have to be using Google Analytics to get these numbers.

Bounce Rate

Bounce rate is the percentage of site visitors who visited one page and then left the site.

bounce rate

Usually, a high bounce rate is considered a bad thing. If you’re an ecommerce site, and 80 percent of your PPC visitors are bouncing, that’s not good – it means that only 20 percent of people are bothering to go beyond the landing page.

But what if you’re using PPC to generate leads, and you have a lead form right on the page? Visitors could conceivably fill out the form and convert right there, without going to another page. In this case, it’s a low bounce rate that’s bad – it means no one is converting!

As you look at bounce rate, think about your campaign goals, and what the numbers mean.

Average Time on Site

Average time on site measures how long visitors spend on your website, in minutes.

average time on site

It probably takes at least 4-5 minutes to complete an online order on an ecommerce site, so if you’re an ecommerce PPC advertiser, a longer time on site is good.

What about the one-page lead form, though? Best practices for online lead forms indicate that shorter forms are best. If it takes 4-5 minutes to fill out your form, you’re using the wrong form. In this case, shorter times on site are a good thing!

Number of Pages Visited

This metric is exactly what it sounds like: the average number of pages visited on your site.

pages per visit

Most ecommerce shopping carts are at least 4-5 pages. Add 1-2 pages for your landing page and any additional items the visitor might be interested in, and you’re looking at a good average of 5-7 pages per visit at a minimum.

I bet you know what I’m going to say here. For the one-page lead form, if your average number of pages visited is 5-7, you’ve probably lost the lead. An average of 1.5 is probably good in this case.

Are you seeing a pattern here? In order to accurately evaluate the meaning of web analytics metrics, it’s crucial to think about your PPC campaign goals. Good ecommerce metrics will be very different from good lead generation metrics.

2 Final Caveats

As with all aggregated data sets, web analytics represent averages. And as we all know, averages lie.

While spending a lot of time analyzing what one or two visitors to your site did probably isn’t efficient, it does pay to break your data out into segments. For now, just remember that averages may not tell the whole story.

On the flip side, watch out for outliers. Let’s say that your ecommerce campaigns have an average time on site of 7 minutes, but you have one campaign with an average time on site of 22 minutes. While on the surface that might sound good, it’s probably not – in all likelihood, it means your poor site visitors are trying in vain to find something and aren’t succeeding. So if your underperforming campaign has outlier metrics like this, it’s probably time to optimize your conversion path a bit.

Now go take a look at your bounce rate, average time on site, and number of pages visited. You might be surprised at what you learn!

Editor’s Note: This post originally appeared on Search Engine Watch on September 25, 2012.

Related Posts:

Comments

  1. Recipe for analysis paralysis.

    • Melissa Mackey says

      Analysis paralysis, really? I understand the concept: over-analyzing things to the point that nothing gets done. But analyzing what’s happening with PPC visitors once they arrive on your website is crucial to successful optimization.

  2. Rami El Chamaa says

    Good read; however, this is more of a Website Optimization analysis rather than a PPC optimization.

    • Melissa Mackey says

      Maybe you should go back and read the article again. I talk specifically about how to use analytics to improve your PPC performance.

  3. Excellent overview.
    I’m curious if there is a % for pages visited that would be considered normal in relation to size of the site.
    For my e-commerce site, if my site has 75,000 pages – I would presume that number of pages visited would increase.
    I realize it’s very difficult to gauge averages, but it would be extremely useful to gauge averages based on different volumes of sites. Bounce rate obviously would apply to any size site, but pages per visit and duration do seem to go hand in hand with site volume and effectiveness of UI.

    • Melissa Mackey says

      Hi Brett – good questions. Larger sites, especially ecommerce, would logically have more page views than a smaller lead-gen site. I’d start with looking at visits that resulted in sales – see how many pages they visited on average. Use that as a baseline and look at outliers – visits with far more or fewer pages than the average for conversions, and dig in from there. Thanks for reading!

Speak Your Mind

*

This site uses Akismet to reduce spam. Learn how your comment data is processed.