Our Big Data Set: 1 Billion Phone Calls

big_data_mainEveryone is talking about Big Data these days and how companies can be more efficient, scale faster and unlock new revenue opportunities by analyzing the huge volume of data that has become available within most companies in recent years.

We’re certainly seeing that at Marchex, and so I thought I’d begin a series of posts around our “big data” set and what we’re researching at the Marchex Institute.

Fair Warning: we (obviously) won’t be posting our trade secrets, but we do hope to raise the level of discourse around mobile advertising, click-to-call and call analytics.  In addition, if you reach out to us at marchex-institute at marchex dot com with specific requests or areas you’d like us to explore we’ll provide you feedback and do our best to dedicate a future post.

So what is our Big Data Set?  More than 1 billion phone calls made from consumers to businesses in North America.  Our clients leverage our analytics platform or advertising network to generate new inbound sales calls, and so our mission is to use this data to inform how we can provide our clients with the best opportunity to improve their cost per acquisition.

This post will focus on something pretty simple – at what time of day (and day of week) we make phone calls to businesses.

In particular, we want to know:
1. What is the distribution of calls by time of day?
2. What is the distribution of calls by day of week?
3. Do “frequent callers” make calls at different times than “non-frequent callers”?
4. Do long calls happen at different times than short calls?

Why would we look at these questions?  The first two are straightforward: when considering when to buy advertisements, it is important to get a sense for call supply (when consumers want to call).

The last two might be less clear.  Frequent callers, those who phone a business several times per week, in some circumstances are less valuable than non-frequent callers.  Why?  People that call businesses day-in and day-out are frequently vendors, solicitors, job-seekers and robo-dialers.  It isn’t always the case that a frequent caller is a vendor, but it is rarely the case that a vendor will only make one outbound call to a business per week.

Long calls are also interesting, because they indicate a deeper and more involved conversation.  Duration is not a sufficient metric to determine quality, but super-short calls (which can be as many as half of all calls in some cases) are almost certainly not quality phone calls.

This first chart shows the time of all phone calls, and that Monday at 11EST is the peak hour for calls nationally.  Monday sees the most calls, followed by Tuesday, Wednesday and so on.

Slicing this data by frequent callers (those callers that phone a business more than 2 times per week, on average) have a different pattern.  Mid-week calls are the most frequent, followed by Monday and then Friday.  Peak activity happens a little later in the day.  If we believe that these calls are business-to-business phone calls, perhaps these callers see more success (through learned behavior) during non-peak times and have shifted behavior accordingly.

Finally, looking at this data by calls lasting longer than 90 seconds shows that long calls generally meet the same patterns as the entire set.

We’d love to hear your feedback, so please reach out at marchex-institute at marchex dot com

John Busby
VP, Marchex Institute

John Busby leads the Marchex Institute, a research and analytics team that publishes findings on mobile advertising and the growing digital call advertising industry. The Marchex Institute also provides custom research and consulting services for customers on their mobile and call-ready advertising campaigns. Previously, Mr. Busby served as Vice President, Product Engineering at Marchex. You can find him on Google+ and Twitter.

Posted in Data, Marchex Institute, Performance Advertising

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 1 other follower

Follow us on Twitter
%d bloggers like this: