Understanding Customer Behavior through PPC data

You should always look for trends when you look at your web metrics data. Focusing on aggregate numbers just doesn't  make sense in any way.  For example knowing that I get 1000 unique visitors to my website in a month wont help me much.  But knowing where they come from, what time of the day, which days of the month (weekends or weekdays), what content they visit, how much they consume of it, is really meaningful.

Always focus on customer behavior rather than aggregate metrics. You can do some cool PPC analysis by applying that principle.

The setup is quite simple. Let's measure how long it takes for someone to convert from the first time they visit your website. The report is called Days to Purchase. Look at the figure below. You'll notice, top right, that it is segmented to show only Paid Traffic.

This data is for a travel website. It might strike you as odd that only 46 percent of the people make a purchase on the same day, simply because airline tickets and hotels and cruises all tend to get more expensive with each passing day. That's odd behavior for Visitors.

OK, let's find out whether we see the same behavior with traffic from other sources. Below figure shows the report for Direct Traffic.

Whoa! With Direct Traffic, 71 percent purchase on the same day. That is normal behavior for Visitors for a travel site. So, what is up with the PPC traffic? Why don't they behave like normal people? If the behavior from PPC traffic is so odd, then how can we treat them better or differently?

These are all great questions, and they're the reason why I recommend this type of analysis. Each traffic stream is unique!

In this case, the travel website took this data and reviewed the landing pages for its PPC campaigns. At the time of the previous report, it was geared toward converting Visitors quickly ("Buy now! Book now! Give us your money, now!").

The first action the marketers took was to soften the calls to action, because they realized a good chunk of the traffic does not want to buy right away. Then the marketers added a new feature, Save Your Itinerary. They realized Visitors would come back, so they might as well make it easy for them.

Finally, the marketers added another feature: "Email me if price goes up by x percent," where x was a number that the Visitors could input when saving an itinerary. This was a very clever move because the travel agency then had the contact information for the Visitors and could email them when the price went up by 10 percent or 20 percent or whatever number the customer input. This caused the customers to return and make a purchase sooner, and the customers were happier because they felt the site was watching out for them.

The net return for closely analyzing PPC customer behavior was that it brought forward Conversions to fewer days to purchase and tripled Conversion Rates. That's not too shabby for a simple segmented report, right?